$2b/month which is $24b/year. Not as much as I expected considering they were at $20b by end of 2025.[0] They only added $4b since?
Anthropic had $19b by end of February 2026 and they added $6b in February alone.[1] This means if they added another $6b in March, they're higher than OpenAI already.
However, I heard that OpenAI and Anthropic report revenue in a different way. OpenAI takes 20% of revenue from Azure sales and reports revenue on that 20%. Anthropic reports all revenue, including AWS's share.[2]
They aren't reporting anything yet. What we hearing is just from news media who get their leaks/info from investors who get some form of IR reports/ presentation.
Both will do public reporting only when they IPO[4] and have regulatory requirement to do so every quarter.
For private companies[1] reporting to investors there are no fixed rules really[3]
Even for public companies, there is fair amount of leeway on how GAAP[2]expects recognize revenue. The two ways you highlight is how you account for GMV- Gross Merchandise Value.
The operating margin becomes very less so multiples on absolute revenue gets impacted when you consider GMV as revenue.
For example if you consider GMV in revenue then AMZN only trades at ~3x ($2.25T/$~800B )to say MSFT($2.75T/$300B) and GOOG ($3.4T/$400B) who both trade at 9x their revenue.
While roughly similar in maturity, size, growth potential and even large overlap of directly competing businesses, there is huge (3x / 9x) difference because AMZN's number includes with GMV in retail that GOOG and MSFT do not have in same size in theirs.
---
[1] There are still a lot of rules reporting to IRS and other government entities, but that information we (and news media) get is from investors not leaks from government reporting - which would be typically be private and illegal to disclose to public.
[2] And the Big 4 who sign off on the audit for companies prefer to account for it.
[3] As long as it is not explicit fraud or cooking the books, i.e. they are transparent about their methods.
[4] Strictly this would be covered in the prospectus(S-1) few weeks before going public and that is first real look we get into the details.
Does the GAAP accounting matter if everyone passively buys shares due to the new fast entry rules, which corruptly will force us all to buy into these companies? The fundamentals and true value seem less relevant than ever:
For other readers, I want to add some context here. NASDAQ is pondering whether or not to change their NASDAQ 100 index membership rules for IPOs. Currently, there is a three month waiting rule for IPOs. They are proposing (not sure if passed/agree/completed yet) to remove this waiting rule for IPOs.
Real question: What is the real impact of this rule change? To me, it seems so minor. Three months is just a blip in time for any long term investor.
> which corruptly will force us all to buy into these companies
Why is this "corrupt"? That term makes no sense here.
Also, if you don't like the NASDAQ 100 rules, then you don't have to invest in securities that track it. You can trade the basket yourself minus the names that you don't like.
Finally, I would say that S&P 500 index is far more important than NASDAQ 100. To join the S&P 500 index, the name must be profitable for the most recent year. (four quarters). Recall that Uber IPO'd in 2019, but was not profitable until 2023. OpenAI probably will not be profitable when it goes public; thus, it will not join the S&P 500 immediately.
I think the bigger story is SpaceX. It will likely IPO very close to a 1T USD market cap (with a small float: ~10%). And, thanks to StarLink, I assume that SpaceX is now wildly profitable.
Diluting the index entry rules, only devalues the index utility. When it becomes a bigger problem, other indices with higher quality controls will out compete the current ones and be used by asset managers seeking safety.
More likely than not, most of us are already holding stock in these companies one way or another. All the Mag 7 hold a major chunk of OAI and Anthropic stock anyway, slower entry does not make it less risky for us.
Even if the big tech companies did not hold any stock, they are still the biggest vendors and their own order books is hugely impacted by the AI demand from these two ( and others in this space), either way we are all in this together.
> When it becomes a bigger problem, other indices with higher quality controls will out compete the current ones and be used by asset managers seeking safety
I personally find this is the correct solution, since indexes are over-inflated either way, this brings much needed sanity to the index. Your index is now worth much more or much less based on how you view the AI bubble and you are forced to understand and correct your forward looking investments accordingly.
Passive investments are good, but if taken too far as they clearly have been in the last decade they become a scam. Everyone is SIPing into it, and there is infinite liquidity. Until one big whale finally decides they are booking it, then all hell will break loose on the same damn day.
You can just choose not to play the accounting game, and only choose the ones that actually gaap viable as investment opportunities. For example mag7 - tesla are all relatively cheap when they dip.
Some times the best play is just not to play. If you think they are too risky, walk away. There are enough good oppotunities
> mag7 (minus) tesla are all relatively cheap when they dip
I asked ChatGPT for a list of Magnificent 7 stocks and their most recent price to earnings (PE) ratios.
Company Ticker P/E Ratio
Apple Inc. AAPL ~33
Microsoft Corporation MSFT ~25
Alphabet Inc. GOOGL ~29
Amazon.com Inc. AMZN ~30
NVIDIA Corporation NVDA ~38
Meta Platforms Inc. META ~28
Tesla Inc. TSLA ~378
In the last 50 years, I think the median PE ratio for S&P 500 index is about 15. Seven and below is considered rock bottom, and 30 and above is very high. These PE ratios look pretty damn high to me.
How much do these names need to "dip" for you to consider them cheap?
They aren't reporting anything yet. What we hearing is just from news media who get their leaks/info from investors who get some form of IR reports/ presentation.
The $24b figure is literally in OpenAI's announcement.
The $19b ARR and $6b added in Feb came directly from Anthropic CEO recently.
I am reminded of the "I declare bankruptcy" meme from the 2000's TV series Office.
When we say reporting it means there are statutory submissions with an auditor signing off, with legal liability. As the other reply referenced consequences for doing this incorrectly can be severe - Arthur Anderson is no more after all because of Enron.
A Press Release (of a private entity) does not have to satisfy this high bar.
Press release does mean no constraints, for public companies, disclosure of important information by officers and other insiders have strong controls. Even if its the just a rocket/poop emoji on a casual social media platform. Lawyers have to refile with the SEC in the expected format. Even private companies have restrictions on not claiming things fraudulently to investors, but these are accredited investors with lesser controls than retail.
Except it's not 100x revenues, and it's not 17% growth. I don't know where you got those numbers from?
The numbers OpenAI gave in the post would mean a 30x multiple pre-money. And the $20B -> $24B run-rate growth since the start of the year could plausibly mean anything from 110% to 200% annualized growth rate, depending on whether that happened over two or three months. The $24B is a lower bound as well, since they only gave use one significant digit for the monthly revenue.
You're right, I was thinking about 100x revenues and forgot to confirm the math. Updated to reflect your point. ChatGPT itself provided the 17% number (it's most recently available growth rate)...
let's not forget that these major LLMs are all the children of corporate hyper-piracy en masse, none of them are ethical even in origin unless you're talking about the pre-product company charter kind of ethics, like google .
Theyâre about even in general, but for me OpenAI is slightly or significantly ahead in the areas I care about the most. E.g. claude code is a backend slop cannon if you donât tell codex/gemini to review the outputs.
Oh, I read it as the number of subscribers would triple, but you're suggesting the price will?
That makes a little more sense, because the number of subscribers are so low that tripling won't really make much difference in terms of turning a profit.
And that is revenue only. In the past 15 or so years most US companies (and especially startups) always talk about revenue only. Wheras only profit should matter.
E.g. what good is 20 billion per year when "OpenAI is targeting roughly $600 billion in total compute spending through 2030". That is $150 billion per year?
The startup game is about building assets and then cashing out on them during exit.
Assets are harder to measure. Facebook used to say something silly like every user was worth $100. That sounded ridiculous for a completely free app but over a decade later, the company is worth more than that. Revenue is an easier way of measuring assets than profit.
Profit doesn't really matter. It gets taxed. But it's not about dodging taxes; it's because sitting on a pile of money is inefficient. They can hire people. They can buy hardware. They can give discounts to users with high CLTV. They can acquire instead of building. It's healthy to have profit close to $0, if not slightly negative. If revenues fall or costs increase, they can make up for the difference by just firing people or cutting unprofitable projects.
Also when they're raising money, it makes absolutely no sense to be profitable. If they were profitable, why would they raise money? Just use the profits.
Profit is money you couldnât figure out how to spend. During growth, you want positive operating margins with nominal profits. When the company/market matures, you want pure profits because shareholders like money. If you can find a way to invest those profits in new areas of growth, thatâs better.
> Profit is money you couldnât figure out how to spend.
Profit is the money showing your business is sustainable. Ever since the ZIRP era US companies keep haemorrhaging money at a rate that is physically impossible to recoup.
If OpenAI plans to lose 100+ billion dollars per year for half a decade, what profits are you talking about to offset the losses?
> When the company/market matures, you want pure profits because shareholders like money.
Ah yes. Shareholders like money. And not, you know, basic accounting like "we need money to actually pay salaries, pay for equipment and offices etc. without perpetually relying on seeming endless investor money".
In principle yes, but all metrics so far suggest they are losing money every user interaction. There is very little network effect with these tools so It's not like they can start cutting back on staff and feature deployment.
Everyone wants to treat OpenAI like a car wash business where they need to make a profit almost immediately. I donât know why people canât understand that the industry is in a rapid growth stage and investing the money is more important than making a profit now. The profits will come later.
It's not as much as you think. Google is spending $185b on data centers this year alone. Amazon is spending $200b this year. Total capex for big tech is ~$700b in 2026 and we're not including neo clouds, Chinese clouds, and other sovereign data centers.
Since everyone is trying to get compute from anywhere they can, including OpenAI going to Google, it's hard to tell what is used internally vs externally.
For example, it's entirely possible that Google's internal roadmap for Gemini sees it using $600b of compute through 2030 as well. In that case, OpenAI needs to match since compute is revenue.
Why are we treating OpenAI and Anthropic differently than say, Amazon or Uber? Both companies invested in growth for many years before making a profit. Most tech companies in the last 2-3 decades lost money for years before making a profit.
Why are we saying that OpenAI and Anthropic can't do the same?
How did Uber somewhat break even? They lost $34b before making a profit.
Uber was only on a path to monopoly in the US, not world wide. Itâs lost to local competitors in most countries. And it can get disrupted by self driving cars soon.
OpenAIâs SOTA LLM training smells like a natural monopoly or duopoly to me. The cost to train the smartest models keep increasing. Most competitors will bow out as they do not have the revenue to keep competing. You can already see this with a few labs looking for a niche instead of competing head on with Anthropic and OpenAI.
How do you distill when OpenAI and Anthropic inevitably move to tasks running in the cloud? IE. Go buy this extremely hard to get concert ticket for me.
Distilling might only be effective in the chat bot dominant era. We are about to move to an agents era.
Furthermore, Iâm guessing distilling will get harder and harder. Claude Code leak shows some primitive anti distilling methods already. Thereâs research showing that models know when itâs being benchmarked. Whoâs to say Anthropic and OpenAI arenât able to detect when their models are being distilled?
OpenAI have burned nearly 25 times what Uber did, it has more competitors, billions of dollars in obligations and no clear way to profitability.
The problem for OpenAI is that the cost of getting them where they are now has been to high and competitors can now establish themselves for much less money.
why should only profits matter? if i had a killer product today that i just need to sell tomorrow, wouldn't you still invest today knowing i'll probably only start to make money tomorrow (or perhaps next week)?
the expectation is that they'll eventually make money. they can't raise forever. only startups are not profitable for a few years. but most companies that have existed for a long while have been profitable
and since they're expected to make a LOT of money, everyone wants a piece of that future pie, pushing up the valuation and amount raised to admittedly somewhat delusional levels like here
It's well know everyone is making great money on inference. The cost is training.
Whether GPT-5 was profitable to run depends on which profit margin youâre talking about. If we subtract the cost of compute from revenue to calculate the gross margin (on an accounting basis),2 it seems to be about 30% â lower than the norm for software companies (where 60-80% is typical) but still higher than many industries.
(They go on to point out that there are other costs that might mean they didn't break even on other costs - although I suspect these costs should be partially amortized over the whole GPT 5.x series, not just 5.0)
"Most of what we're building out at this point is the inference [...] We're profitable on inference. If we didn't pay for training, we'd be a very profitable company"
"Thereâs a bright spot, however. OpenAI has gotten more efficient at serving paying users: Its compute marginâthe revenue left after subtracting the cost of running AI models for those customersâwas roughly 70% in October, an increase from about 52% at the end of last year and roughly 35% in January 2024."
> It's well know everyone is making great money on inference.
That is not, in fact, "well known", but based entirely on the announcements of the inference providers themselves who also get very cagey when asked to show their work and at least look like they're soliciting a constant firehose of investment money simply to keep the lights on. In particular there's a troubling tendency to call revenue "recurring" before it actually, you know, recurs.
> based entirely on the announcements of the inference providers themselves who also get very cagey when asked to show their work
I mean sure, it's self reported.
But the inference prices somewhere like Fireworks or TogetherAI charges is comparable to what Google/AWS/Azure charge for the same model an we know they aren't losing money - they have public accounts that show it, eg:
> If someone has a subscription then yes that is pretty normal.
Not if you've substantively changed rate limits 3 times in the last 5 months while still counting those forecast revenues. In most industries that's called rug-pulling.
It doesnât matter how you call it. A recurring subscription on the books is a recurring subscription. Yes you can cancel anytime (how generous of them), it also doesnât matter.
And why do you think twenty competitors can stay competitive for years to come?
Industries always consolidate and winners emerge. SOTA LLMs look like a natural monopoly or duopoly to me because the cost to train the next model keeps going up such that it won't make sense for 20 competitors to compete at the very high end.
TSMC is a perfect example of this. Fab costs double every 4 years (Rockâs Law). It's almost impossible to compete against TSMC because no one has the customer base to generate enough revenue to build the next generation of fabs - except those who are propped up by governments such as Intel and Rapidus. Samsung is basically the SK government.
I donât see how companies can catch OpenAI or Anthropic without the strong revenue growth.
Google has already surpassed them both in all areas except coding. People on HN only look at benchmarks, but Gemini's multimodal understanding, things like identifying what a plant is, normal user use cases (other than chatting), integration with other tools, is much better.
It's believable that Meta, ByteDance, etc. can catch up too. It is not certain that scaling will meaningfully increase performance indefinitely, and if it stops soon, they surely will. Furthermore, other market conditions (US political instability) can enable even more labs, like Mistral, to serve as compelling alternatives.
Uber, TSMC, etc. have strong moats in the form of physical goods and factories. LLMs have nothing even remotely comparable. The main moat is in knowledge, which is easy to transfer between labs. Do you think all the money that goes into training a model goes into the actual final training run? No, it is mostly experiments and failed ideas, which do not have to be repeated by future labs and offshoots.
The barrier to replicating TSMC isn't just cost, it's supply chain, geopolitics, and talent.
Only one company on Earth can make the UV lithography machines TSMC buys for their highest end fabs, and they're not selling to anyone else.
The PRC tried to brute force this supply chain backed by the full might of the Party's blank check, all red tape cut, literally the best possible duplication scenario, and they failed.
They will succeed eventually since they have proof itâs possible and their plans span decades. I expect them to have working EUV in 10 years. Whether itâll still be bleeding edge tech is a different question I dare not guess the answer to.
>Industries always consolidate and winners emerge.
no, most industries just sell boring generic products, a few industries favor monopolists. Semiconductors are one of them but LLMs are also as far removed from that business as is physically possible.
TSMC makes the most complicated machines humans have ever built, a LLM requires a few dozen nerds, a power plant, a few thousand lines of python and chips. That's why if you're Elon Musk you could buy all of the above and train yourself an LLM in a month.
LLMs are comically simple pieces of software, they're just big. But anyone with a billion dollars can have one, they're all going to be commoditized and free in due time, like search. Copying a lithography machine is difficult, copying software is easy. that's why Google burrowed itself into email, and browsers, and your phone's OS. Problem for openai is they don't have any of that, there's already half a dozen companies that, for 99% of people, do what they do.
Profit is money you can't find a use for to grow your business, so you give some of it to the government in the form of tax.
Also there is a big difference between operational expenses and capital expenses like building data centers.
I think OpenAI is being very aggressive on the growth vs conservative financial management spectrum but just saying "only profit should matter" is just wrong.
OpenAI is great at attracting people who say "yeah, sure, I'll give you capital at some point in the future" who then never actually give them the capital (or at least haven't yet).
Even a simple shop isn't profitable for months if it needs to buy stock up front, and run some ads to let people know about it. The money for that comes from the shop owners as an investment.
This is the same thing but on a slightly bigger scale, over a longer time frame.
If your shop is unprofitable for years with no chance to recoup any of the costs, you close it, as your investments run out, and investors and banks stop giving you money as you keep losing them.
US tech companies just continue operating because "revenue and growth".
> Today, we closed our latest funding round with $122 billion in committed capital at a post money valuation of $852 billion.
A couple things that stand out to me about this is the use of the phrase "committed capital", which only sounds like a promise that could break from various circumstances, and the valuation of their funding keeps changing so it sounds like a max rather than the valuation every investor invested at.
Probably a lot? It would be much more tax-advantageous to do it this way, $50B worth of credits != $50B worth of spend on Amazon's part, and they might meet in the middle about how much equity that translates to.
You're Amazon. You give OpenAI $50B cash investment, they then hand you back the $50B over time because they buy $50B worth of Amazon AWS services (they would use AWS or other equivalent compute anyway). OpenAI pays an additional $1-5B in sales taxes on top of their $50B compute purchase. Now let's say you have $25B opex for said compute. You then have $25B profits, you pay 21% corporate taxes on the profits, so you too owe the government about $5B. Government collects around $6-10B on this whole transaction.
Situation B:
You're Amazon. You let OpenAI use your services by handing them API credentials that unlock what would normally cost $50B worth of services, but no money changes hands. You have zero revenue from the transaction, write off the $25B opex as a tax loss on your other profits elsewhere in the company. You thus pay ~$5B less tax on your other income as a company, and OpenAI also doesn't have to pay sales tax because they didn't actually purchase anything.
Isnât sales tax only for consumers? Ie companies reverse charge sales tax or omit it entirely. Or what is this 2% - 10% sales tax youâre referring to?
It depends how you defined âconsumersâ. If you mean âthose who consume the good subject to the tax, rather than people who resell the goodâ, yes, ideally.
If you mean ânot businessesâ or âindividuals but not corporationsâ, then, no.
> Ie companies reverse charge sales tax or omit it entirely.
Generally, the theory of sales taxes is that people (including corporations) pay sales tax on things they consume as a final good rather than use as an intermediate good in production or simply resell. The exact way in which that is determined varies somewhat between jurisdictions with sales taxes, but generally (assuming paper is subject to sales tax in a jurisdiction), if you are buying paper to print books that you sell, you don't pay sales tax, if you are buying paper to print internal documents that you use in running the business, you do pay sales tax.
Thatâs typical. Large funding rounds usually arenât delivered as one single giant lump sum into the bank account. The capital is committed in stages that can depend on hitting milestones or goals.
This is done even in smaller startup funding rounds some times.
Fair, I think a lot of what I've been perceiving is the gymnastics in how funding and valuation and deals get reported. There ends up being a ton of asterisks that makes the headline news deviate quite significantly from reality, e.g. https://arstechnica.com/information-technology/2026/02/five-...
This announcement completes the betrayal of their founding principles.
"Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return."
- Not advancing digital intelligence
- While locking people into a superapp
- Because they are further constrained to generating financial returns
> The broad consumer reach of ChatGPT creates a powerful distribution channel into the workplace
They mention this line in different forms a couple of times in the article. Itâs clear theyâre pretty rattled about Anthropicâs momentum in enterprise, I wonder how confident they really are in this rationale.
It's an interesting strategy, I see a pretty big risk from them leaning into it like this. We already have a vibe in my circles of the old "gmail vs yahoo" type thing where if you saw someone had a yahoo mail address you assumed they were technologically illiterate. Similarly it's mildly embarrassing already to say you used ChatGPT for something. It's not unrecoverable, but it's a pretty steep slippery slope they probably don't want to be anywhere near if they care about enterprise.
Did you miss the cancel/unsubscribe gpt boycott? It was only about a month ago. Many people I know cancelled/unsubscribed. To be fair though, most people I have talked with needed almost no encouragement to move to anthropic or google (better products, easy to switch etc). Consumer sentiment can change quickly.
I don't know if it's guaranteed to work, but the strategy is real. I know Notion won vs. competitors in the space because they focused on consumer first, and consumers then brought Notion into their workplaces.
I am amused that you think IT is going to respond to an unmanaged LLM tool that operates outside of the LLM policies all serious enterprises have set up by now and say 'wow, that is cool and maybe we should buy in to this!'
What is going to happen is that the emplyee who tries to sneak OpenAI into our org is going to have two meetings set up by the end of the day, one with IT to ensure the whatever tool they installed is burned out with fire and one with HR to ensure they know the company policy and acknowledge that another fuck-up like this is a firing offense.
Kind of makes me wonder how 'accelerated' the timeline of publishing this article was based upon the Claude Code leak today. Considering everyone has gotten a sneak peek at what Anthropic is working on OpenAI might be a little worried. This could also just be coincidence, but this piece really does read like self-encouraging fluff.
No, they didn't raise $122B as the HN title implies. A big chunk of that $122B is a "maybe" that depends on various things that need to happen in the future.
Oh, man... I can't wait to see where this is going. Might not be pretty after all.
I've wondered how many announced fundraising rounds were like this. It's in everyone's interest (VCs and entrepreneurs) if the message to the outside world is "this company is amazing so they've raised a boatload of cash". But VCs might not want to give it all up front, or unconditionally.
It makes it hard to say what the valuation of a company is. If the milestones are unlikely to be hit, then it's anyone's guess.
This is a common structure. It's confusing to people who don't know finance or startups when they first see it.
Even VCs don't get all of their fund money delivered into their bank account when they raise a funding round. It's inefficient and undesirable for everyone involved to have to move all of the money up-front, at once.
If you talk to anyone in startup funding or finance they'll be familiar with the term "capital call" which describes how committed capital obligations are delivered at a later date than the initial deal: https://en.wikipedia.org/wiki/Capital_call
I've been involved in many startups, and this type of fundraising is not common, or at least it wasn't common before a few years or so ago
The whole concept of talking about "runway" is basically calculating how much cash in the bank, that is actually in your bank account, will last. And this arrangement is different, as there are contingencies. In the past, VCs would just give you money in a particular series, and then if your business did well, they'd eventually give you more money in a later series. But it wasn't like they announced it all up front in, say, a Series A, but a big chunk of the money would only be delivered if you met milestones.
$100B isn't a startup. And if there's a $100B deal, you better believe the cash is there. Case in point - Netflix/Paramount wanting to buy WB. Or the $44B that Musk had to raise to buy out Twitter shareholders.
Both your examples are purchases. Musk had to raise actual capital to buy Twitter because the people getting the money were taking it and walking away.
Funding doesn't work like that. Investors are giving you money as part of a longer-term deal where they stick around.
This was already common in tech for Series C+ fifteen years ago when I raised a round. Once youâre talking tens or hundreds of millions, almost everyone wants milestones and tranches instead of giving all the money up front.
Both of CFS B rounds were cash, in recent years, and each in the range of "low billions". Sure another 2 orders of magnitude is another story, but so is selling hope. I'd say the latter is the thing that is unique here.
No, it's not common for the startup itself to make capital calls. The phrase (and your link) refers to capital calls made by VC firms to their limited partners. Same thing in PE.
I think more people are aware that VCs raise commitments for a fund that they can pull in via capital calls than are aware that startup funding from VCs come with hurdles to clear.
This is perhaps because the most common round to raise is a small/early one, and these tend not to have hurdles. Founders that only ever raised these rounds wouldn't necessarily know what happens in later/bigger rounds.
Also, I wonder if capital calls come with hurdles as well? That is, can an LP refuse to put in more money if the VC's recent investments have not done well? I would think not, since it typically takes many years to determine whether investments were good or not.
Gotta hit that high IRR as a fund manager and the clock starts when the cash comes in so capital calls are appreciated by fund managers. Unless they are emerging managers (the startup equivalent in finance) and their LPâs are less than institutional and ghost them when the capital call hits.
IRR is so trivial to manipulate - it'd be wonderful if more investors began demanding actual metrics on capital performance. If you're parking cash with an investment firm you want to know about how much of a return you can expect when it is withdrawn, and while history is a guide and not a guarantee, there are much better ways to inform that expected return than IRR. "My million got a return of 2% during a year when your reported IRR was 10% - where's the other 8%!" is a common cry from those who haven't just rolled over their investment, unaware of how little it has functionally appreciated.
There's an accelerator here in Taiwan with a model I truly don't understand: 100k usd for 10%. 10%!! You've just valued the company at only 1 million! And taken a HUGE chunk of equity, not much left on the table!
Maybe it makes sustainable sense but in the world of venture capital it seems the most profitable thing to do is lie through a Cheshire grin, every day.
It just makes comparing funding rounds hard to understand, since money in the bank is money in the bank, and a lot of the "committed capital if you reach a milestone" is capital that would be easy to get if you reached that milestone, if it is sufficiently advanced, and has enough outs, etc., that you may as well have just raised another round in the future.
Note that even that "money in the bank" of traditional venture firm is not really money in the bank. VC, PE, and hedge fund managers usually don't have all the cash for the fund sitting in the bank at all times. Rather, their agreement with the LPs that fund the fund is structured as a series of capital calls: it gives the fund the right to demand that their LPs deposit cash in their bank accounts within 10-30 days, which can then be used to fund the investments that the VC firm makes. The capital calls are backed by legal documents enforceable in court, with pretty stiff penalties for failing to meet a capital call.
Such a funding structure here isn't all that different: the funding agreement gives OpenAI the right to call on their backers to make certain cash deposits, contingent upon milestones being met. Deep down inside, "money in the bank" doesn't actually exist, it's just mutual agreements backed by force of law.
Thatâs logically inconsistent. If the company was performing poorly enough that they couldnât meet their funding milestones from a previous round, theyâre not going to have an easy time raising the same money in a future round.
The milestones arenât a hard-stop that forbids the previous funding round participants from providing the money if they still choose. Itâs just an out.
What I am saying is that if you do meet the milestones from your previous round, you're going to have an easy time fundraising anyway, so funding contingent on milestones isn't that different than just saying "well, if we need more money we can do another round"
Fundraising rounds are difficult, laborious, and distracting. It would be extremely different to try to multiply the number of rounds by 3-5X. There's nothing easy about that.
You're also ignoring that the market changes frequently. If you only raised as much money as you needed for the next 4-6 months with plans to re-raise all the time, you'd have to constantly be sizing your growth plans up or down based on how the market felt about startup investing that month.
Imagine the company having to either do speed hiring or large layoffs every few months to adjust to the size of the fundraising round they were able to get this time around.
Nothing about what you're suggesting would be easier, or easy at all
Funding Round A: VC âAâ invests 200M (100M immediately and another 100M if sales grow 10% or whatever)
At 6 months the company will either get the other 100M automatically (meaning they grew sales 10%) or they donât (meaning they grew less than 10%).
Assuming itâs the later they can then do another round during which they try to get the other 100M. In all likelihood VC âAâ wonât be interested (or interested at a lesser amount). They could go ask VC âBâ for an investment but it will likely be less than 100M as well because they didnât grow as much as âthe marketâ anticipated.
Nothing complicated at all.
Iâll give you $1 dollar for your banana today and another dollar in a week when it has ripened. If itâs rotten when I come to get my banana I wonât give you the other dollar. You have your original $1 and you can still try to sell your rotten banana to another HN reader but you probably wonât get another dollar this time.If instead you have a ripe banana Iâm sure you could easily find a buyer.
The funds are committed under the terms of the deal (share price, things like board seats, and other details). There are legal obligations to provide it.
This is a common structure for large investments. It would be really inefficient for all of these investors and companies to have to have the money sitting in cash to do a deal and then transfer it into the company's bank where it sits and earns interest for years until they can deploy it.
Even VC firms who raise funds work this way. The capital is "committed" but investors don't wire all of the money over right away so it can sit in the VC firm's bank accounts, waiting. The VCs do what's called a "capital call" through which they're legally bound to provide the money they committed when requested, under the terms of the deal.
It's splashier this way, and is meant to shape the narrative, make other companies fear their warchest, and make hiring easier. Of course, those who are in-the-know won't be fooled, but the perception of the general public will be set in stone by the PR framing.
Their ~$50 million total Alibaba investment turned into ~$70 billion. As of two years ago they were still liquidating out of it.
January 26, 2024 - "Japanese investment holding firm SoftBank Group Corp has largely cleared its ownership in e-commerce giant Alibaba Group Holding, concluding one of the most successful deals in China's internet industry and a holding that spanned about 23 years."
"SoftBank, which invested US$20 million into Alibaba when it was still a start-up in 2000, said in a corporate filing on Thursday that it was set to book a gain of 1.26 trillion yen (US$8.5 billion) - about 425 times the value of its initial outlay - for the Tokyo-based firm's 2024 financial year after divesting its [remaining] shares via subsidiary Skybridge."
"As of two years ago they were still liquidating out of it"
I get that people are scared of investing in China. But if I still made single stock investments, I would seriously consider BABA, it seems well positioned.
Tell that to Trump and his glorious way of bombing Iran. Nothing against the idea itself, the Mullahs all but asked for it to happen.
But the execution? That was a level of dogshit I haven't seen in the time I was alive lol. Even Russia was better prepared with their invasion of Ukraine.
Both Trump and Netanyahu had a somewhat solid perspective on not getting utterly wasted in the next elections. Instead they go on one of the most ill-prepared wars in modern history, with results that may seriously upend the global economy if not lead us to WW3 outright.
It's also like... >$50 billion in compute credits and discounted hardware between Amazon/Microsoft/NVidia. Which is all inside baseball since they simultaneously juice their financials with OpenAI's cloud compute bill
With NASDAQ and NYSE looking to reduce the timelines for new public companies to be included into indices (âfast entryâ rule), I have a feeling that OpenAI and SpaceX and Anthropic are mostly looking to dump their inflated shares into the publicâs retirement accounts by force.
Michael Burry called out this structural manipulation play recently:
Retirement accounts already own funds and those in turn are often tied to the underlying index. If the time to being included in an index is reduced, they end up being automatically bought sooner. And that keeps their price from collapsing artificially.
The title is incorrect. The $122B includes previous promises. They raised an additional $12B of promises:
"The round totaled $122 billion of committed capital, up from the $110 billion figure that the company announced in February. SoftBank co-led the round alongside other investors, including Andreessen Horowitz and D. E. Shaw Ventures, OpenAI said."
This IPO, if anyone underwrites it, is going to fleece retail so hard. Better make it a SPAC with the help of Chamath and Cantor & Fitzgerald.
They were trying to keep the facade up until they were allowed to become a public benefit corporation. At least that's the way it seemed to me. Now they are fully mask off.
I'm old enough to remember when companies worth $1 billion were called "unicorns." Now we have a company raising 122 times that? Valued at nearly 1000 times that...?
At least they're throwing consumers a bone via the ARK deal. It's crazy how little AI exposure is available to anyone who isn't already wealthy and/or connected.
I think this is reality-distortion field rivaling that of Jobs', and a crisis of faith. Nobody apparently believes that capital is worth investing into anything but AI.
> Nobody apparently believes that capital is worth investing into anything but AI.
This is the main reason we see this insane investment into AI imo. If you imagine having lots of money, where should you invest that currently?
Housing market: Seems very overvalued (at least in germany). Also with the current uncertainty and inflation its hard to make an investment that pays back over 20-30 years. So building is also difficult.
Stocks are very volatile currently. Not only since Iran. To me it seems since the financial crisis 2008 investors don't enjoy stocks as before.
Gold: Only if you are paranoid about collapse of society. It doesn't make sense to invest into s.th. without interest rates.
Crypto: Same as gold, but better if you like gamling. I would assume most people who are very rich don't gamble with most of their fortune.
Looking around, and especially forward, it would be military tech, e.g. [1], and its supply chain, e.g. [2] :-\ Valuations are not as crazy, but I bet there'll going to be a lot of demand in the coming decade, unfortunately.
Chip production, too, of course, but it's overflowing with money already, apparently. It's growing though, because there are real actual shortages of stuff like RAM and SSDs, there's money to be made immediately if you can. Chinese RAM manufacturers are building out like crazy.
Would you be fine with the ethical implications of funding the industry to fight WWII? Would you consider funding Ukrainian military unethical? Or Taiwanese?
This is, sadly, not theoretical, and I'm afraid we'll soon see more of such choices, not fewer.
Anduril is the only company in this sector in the US that has any promise and they aren't even public. Most of us are not going to get our hands on this.
Traditional defense sector looks more like Jeep, or Kodak...
Anduril has yet to deliver anything of consequence. I hope they shake up the industry but to say they are the next hot thing and write off the primes at this stage is premature.
"The Roaring Twenties roared loudest and longest on the New York Stock Exchange. Share prices rose to unprecedented heights. The Dow Jones Industrial Average increased six-fold from sixty-three in August 1921 to 381 in September 1929. After prices peaked, economist Irving Fisher proclaimed, "stock prices have reached 'what looks like a permanently high plateau.'"
You can argue that current market multiples are higher than 1929 [1] - and they're certainly high - but this also ignores the mechanism that drove that crash, focusing only on the symptoms. We simply aren't doing the kind of consumer margin buying that drove the '29 crash. It isn't even close. Average schlubs were leveraged to the stratosphere to buy shares of boring industrial stocks.
> The US stock market has nearly tripled since then. Literally the best period of stock growth in history.
The only thing I meant to point out was that a very high stock price by itself is no guarantee that there isn't a crisis around the corner. We plugged a lot of holes after 2008 and then reversed a lot of those fixes, I hear retail investors talking about their stocks at birthday parties again. Deja vu... of course this time it will be different. Or not. Let's just say that with the proverbial bull in the earthenware goods store on the loose if we only end up with another financial crisis that might actually not be so bad.
Do you know the actual lessons of that crash? Because we don't allow retail investors to go 10:1 on leverage anymore. There are a lot more lessons and none of them apply to this situation (even Glass-Steagall). This is much closer to the dot com crash in 2001 in how it looks, just a lot more concentrated and probably a bit bigger. If all you got is "number go up too much" then you probably shouldn't be investing your own money.
The good news is that its almost all rich folks money on the line here and a small amount of dumb money. That's very different than, 2008 where it was mostly the indexes that got hit and that's more middle class/upper middle class concentrated.
It's the result of too much echo chambered bullshit floating around daily about how capable LLMs really are. It's literally crypto/blockchain all over again. It's one big lie that a lot of people have bought into which causes it to self-perpetuate, like religion.
That's the tao of hyper-financialization. It must keep growing irrational exuberance big and up forever like stonks or it bursts like DotCom and tulip mania. It's funny money that cannot be liquidated for real value for more than a tiny fraction of the imaginary trillions being thrown around. Similarly, Nvidia $4T mkt cap makes absolutely no sense when it has but a few incestuous customers-parters-investors throwing around tens of billions each per year devoid of fundamentals like essential service offerings that turn a profit. Those handful of whale customers will make their own chips or cease buying large qtys at any time.
Even a billion dollars is crazy money. If you have a company with a subscription service that costs $100 yearly, you have ~2m customers, with a 50% profit margin. Your company makes ~100m every year in profit. Imo that's what is actually worth a billion dollars, maybe even a bit less.
Also, the valuation for such a debt laden company should be viewed with great skepticism. I'm afraid a lot of mutual funds will end up holding the bags.
> At least they're throwing consumers a bone via the ARK deal. It's crazy how little AI exposure is available to anyone who isn't already wealthy and/or connected.
It is deliberate. Period.
It's always been known that you make money in the private markets and pre-IPO companies and retail is the final exit for insiders and early investors.
Retail is not allowed to be early into these companies (Because that would ruin the point of being an insider) and this "exposure" has to be at the near top.
I can't help but think building an "everything" app is so.. both unbelievably ambitious, and a folly. I am not personally convinced that people want all the things that this super app purports to do.
I am from a generation that still sits behind a desktop computer when making "big purchases." I can't even buy a flight on my phone. I am so much less likely to want to have an AI agent do that for me.
Then the idea that daily consumption of these products will drive people to use them more at work... I have a very different life outside of work. My use of AI outside of work is exceedingly different to what I use it for at work.
I sometimes feel wildly out of touch. But sometimes I view this as the VR moment. To me there are some things that I think may always be preferable to do outside of that ecosystem. And for me, a lot of tasks that 'agents' enable are small enough or important enough that I want to do them myself.
I don't think I'll ever be comfortable allowing an agent to call me a taxi, or order food on my behalf. Because the convenience of asking for food isn't worth the chance it'll mess up, and opening an app and looking at a menu is simpler.
I also think we're coming to a moment where we can start identifying the markers of AI generated content on sight. And I think there's a growing animosity to it. I might be comfortable asking AI something, but when I am looking for or searching for other content, seeing AI content markers make me angry at this point.
To finish, I do just sort of straight up hate the idea that we're comparing this moment to the invention of electricity. It's on the face of it absurd.
Is it a browser or like a browser? I've never actually used it but from what I understand WeChat's mini programs are like web apps but not something you can open up in a typical browser.
Alternatively, you could say browsers are the original super app.
I think the core issue isnât what underlying technology is used, but rather the service providers. They package their services into mobile apps or WeChat mini programs, and restrict functionality on browsers. For many ordinary people, this provides convenience, but for those who care more about privacy, itâs quite problematic.
WeChat in China covers almost every aspect of life. Even someone like me, who doesnât want to use it often, canât avoid it. Some restaurantsâ online ordering systems only support scanning via WeChatâthat is, WeChat mini programs. People can pay utility bills, call taxis, shop, and make financial investments all within WeChat. Alipay offers similar functions as well.
WeChat is also one of the largest content platforms in China, similar to Medium. Countless creators set up subscription accounts on WeChat and gain more users through readersâ sharing and reposting.At the same time, government information is often released through the WeChat platform.
>> To finish, I do just sort of straight up hate the idea that we're comparing this moment to the invention of electricity. It's on the face of it absurd.
Do you feel that any technology is comparable in itâs impact?
Most of modern medicine, by which I mean each discovery and invention in their own right, stand alongside electricity. Particularly vaccines.
AI isnât there yet. You could turn off AI tomorrow and thereâd be a shock but people would quickly switch back. You could not do the same for electricity, medicine, combustion engines (or steam engines/turbines), computers, the internet, modern building materials, etc. You try to swap back off any of those and the modern world (literally and figuratively) collapses. Turn off AI, and thereâd be a financial collapse but afterwards everything would return relatively easily to an earlier way of doing things (ye know, the way from just 4 years ago, and which is still 99% of how people do things :) )
Sure, but compare this to "turn[ing] off" combustion engines a mere four years after commercial adoption rather than 162 years later (now). Back then, going back to horses wouldn't have been as big of a deal as it would be now.
I think the Internet is the more apt analogy. But even with electricity, you could have taken it away within the first couple decades of its popularity and society would have shrugged it off. Once they got used to that telegraph thing, not so much.
Yeah, I agree, but AI isnât there yet. Itâs too early to call it one way or the other. Thereâs plenty else thatâs as important as electricity in my view, and maybe AI will join those ranks in 15 years or so when itâs gone through the hype loop and when the economy has recovered from the now-basically-inevitable AI- and war-fueled turmoil of the next decade.
That's primarily a function of the time for adoption, though, not the utility of the technology. In 20 years, people would not be able to so easily say that they could turn off AI with no impact.
That..what..no. The question was whether there are any comparable to electricity, of which I have put forth a number of examples. And also offered my opinion that it is too early to judge whether AI will be as significant or not.
There are loads of technologies that, despite being decades old, do not qualify. So, no, itâs not âprimarily a function of timeâ. It absolutely is about the utility. We can only be in a position to judge utility when sufficient time has passed, and AI ainât had enough time yet to prove its utility. Given enough time, it might prove as useful as electricity, or it might just sit alongside computer operating systems - never quite making it onto anyoneâs âthis changed the worldâ list, even if it has as much utility as an OS.
I think you lack imagination. This is going to be the future because it is legitimately a step up from the previous ways of doing things. I can do things that were way more difficult before.
It doesn't have to be AI all the way - no one's asking AI to book things on its own and make the payments on their own. What does work is, make AI do the research and you verify and you do the payment. Human in the loop.
To me this is clearly the future - AI has access to all the data sources and can translate your intent by accessing these tools in a loop and use intelligence to automate things.
Maybe there's a scenario where that is useful. But again, I don't know why I'd want an AI to do this research for me. I hop on Skyscanner. I type my location, and where I'd like to go. It presents me with a list of options, and I can then use the filters to find times that work best for me.
I see a flight that isn't in my time frame, but is actually like 400 euros cheaper. And I decide in that moment that waking up at 5am is worth the savings.
I'd have not typed that into a prompt. I made that decision at the moment I saw the possibility. I didn't even know that it was an option prior to that moment.
Then I go look at hotels. I have a list of requirements, but I see that one of the hotels that I just glanced at has a really nice long pool, and the amenities look nicer from the images. I change my mind at that exact moment, I can walk 15 minutes more to the beach.
Now it should be even clearer why this is important for food.
You had me going until you said this is even clearer why it is important for food. Food is cheaper and has less impact on your life. I'm much more tolerant of a mistake or suboptimal experience with food.
Think of personal assistants that rich people have. They learn your habits and take care of business, like making your travel plans. The promise is to give you that for much less.
I get the appetite for frontier models. But why not just invest in Google. Do they really expect the return profile of OpenAI to be vastly different than Googleâs Gemini?
it's not really investing though. Amazon will provide 50B worth of compute and Nvidia will provide 30B worth of chips etc. Google doesnt need any of those.
The only one that is really investing is SoftBank who is pushing for a faster IPO so they hope to make a profit on that and again Google does not offer that opportunity
I imagine they get a bigger slice of the pie with OpenAI than they do with Google, an extremely mature company whoâs had investors buying in for 30 years.
Alphabetâs market cap is $3.5 trillion, compared to OpenAIâs $850 billion reported here.
Google has little need for more money, so the price will be much higher. OpenAI being a separate entity also means Google's competitors (Microsoft, Amazon) can invest there without looking silly.
This leak and looking into source code gave me an impulse to try OpenCode with codex models. I am very impressed with how well it works, and the UI is beautiful.
Google makes money like that in their sleep from mostly just advertising.
OpenAI charges about 20$/month to tens of millions of users right now. 240/year. About 12 billion per year. That's a market that could grown to billions of people with a long tail probably not paying a lot but being served ads; and a fat high end paying a lot more than that for an ad free/premium experience. They should be able to reach billions of users.
It's why Google is matching investments in AI like this entirely from the profit from ads.
Going to space isn't that expensive. SpaceX bootstrapped with a lot less than that.
This all smells fishy. They didnât âraiseâ $122B. Raise means someone put funds in your bank account and said send us the next quarterly report to tell us how our investment is doing.
They have pieces from paper of folks saying they may put up funds or goods and services in that amount. But itâs important to remember that:
1. While they are âraisingâ commitments others are backing out of deals (see Disney, various data center things). Big deals announced to major fanfare are falling through.
2. They slashed capital expenditure for the future after previously boasting about all the commitments. This is turning into bonkers math of X + Y - X + Z + W - 1/2 of Y = ? On trying to keep track of whatâs actually âraised / realâ vs what was PR puffery that folks ran away from later.
3. Circular financing still seems to be going on. Big difference of hereâs cash, have fun and various âcommitmentsâ and balance sheet games that seem to still be going on.
Net net this all still looks very scary and iffy at best.
If OpenAI goes down their investors will lose any chance at getting their money back. They need to keep pretending things are going great for as long as possible.
Are we truly arguing semantics on HN which is a news aggregator for startups and everyone truly knows what a "raise" is and it is obviously not funds in your bank account? I don't disagree with the rest of your comment and the core thesis is valid that OpenAI is very much doing circular financing.
Edit: A raise comes with stipulations on what you can use the money for. I don't know if I was being too mean about responding to a parent but before you comment just google what a raise has..
It's a different definition of the word for one thing, and anyway, unless their compensation is prepaid this would only suggest that "raise" doesn't mean liquid money in an account, because an employee's raise is a promise to pay an amount over the remainder of the year with the stipulation the employee continues at the job.
I hate to read this line when academics and graduate students who work in basic and hard sciences have their funding cut. The grand funding that pays minimum wage to grad students is a burden for this society, yet for a company that took all the valuable data from sources that never got credit, raises billions of dollars. Open says the name, but closed it is by operation. Sorry for this rant, but the priorities of this world suck.
Or all of the people that they didnât ask, let alone compensate, that made all of the stuff they munged up for training data, so they could sell cheap knockoffs in the same markets.
Fortunately, OpenAI's naming didn't cause a wave of greedy and predatory new companies, like OpenHealth, OpenEnergy, OpenCompute. The non-sexy name of their product using the originating algorithm (GPT) is punishing them. It could be anything else more attractive that would bring more customers. Because, for me, a good, inspiratory, engaging name is half the success.
What if all the people currently using these "AI" services are the entire market for those services? I'm pretty sure everyone that wants to use LLMs is already doing so and already paying for the service.
That would mean the only way to increase growth would be to charge more per token and to get the existing people to use more tokens. Both of which seem to be only what mature companies do when trying to squeeze the cash cow for all it's worth.
It also explains why they're trying to stuff AI into everything, to keep the numbers up, and to get everyone to try and pay them money.
When I show people personal projects Iâve vibe-coded with Claude Code, they often seem impressed and envious. They come up with ideas for things that they would like to do, too. But they have full-time jobs outside of IT, and when I mention they might need to use the terminal to do what Iâm doing now their eyes glaze over.
A couple of such people, after they learned about Claude Cowork, signed up for Anthropic subscriptions and are now using it in their jobs. But overall my impression is that there is still huge potential demand from regular people who use computers for agentic systems with less barrier to entry, and that many will be willing to pay for such systems when more mature and user-friendly ones arrive.
Most of humanity hasn't figured out they need to adapt yet. It's a bit like email and the internet in the mid nineties. People had heard about it but hadn't really embraced it yet. Five years later most people with white collar jobs had email addresses. Fifteen years later, billions of internet capable smartphones were in circulation.
The AI revolution is following a similar adoption curve. Right now many of the tools are only really usable if you are a developer or at least not too shy making AI agents use developer tools on your behalf. That's not going to stay like that for very long. It's going to be a messy transition that will likely take much longer than some people seem to think. But eventually most people doing knowledge work will be leaning heavily on all sorts of AI agents to do their thing. And quite a few will have to learn new skills as most of the stuff they still do manually today just goes away as a thing that you do manually.
Like the mid nineties, these are amazing times for people with a slight head start over everybody else. Which is why there is such an investment frenzy around AI right now. Lots of possibilities where lots of money might be made. And lots of things that won't work out. And lots of people really not seeing the forest for the trees as well. And generally behaving like headless chickens. But the internet in the end proved to be not a fad and it didn't all go back to normal after the hype died and the .com bubble burst.
IMHO, the bubble around AI is not so much the technology but things like data center and energy pricing. The cost of data center production is long term a fraction of current cost (dominated by GPUs costing tens of thousands of dollars). Likewise cheap and plentiful energy to power them is going to eventually cost a lot less. Short term scarcity eats up a lot of billions right now. But you'd be mistaken to confuse that for long term structural cost. Cost is going to come down and that will drive adoption. And that's before you consider edge compute on commodity phones and laptops. There will be billions of devices running small AI agents. Add robotics to the mix and it's a whole new world.
In short, companies like OpenAI and Anthropic are valued so high because all of that is happening right now. Yes, it's a bit of a bubble. But stuff will definitely happen.
On the other hand, the productivity gains from AI automation are so large that you are forced to use it to compete in the workplace, even if you strongly dislike the terminal, you will dislike homelessness more.
Give it a year or two, and apple or any other hardware will have unified memory OR AMD will have a good offering to run all that stuff locally. It won't be as good as Claude, but it'll do for 90% of the things. It will be expensive as first, just like the first mainframes, then give it another 5 years or so, and it'll be affordable.
Think about all the "people" AI services can displace in due time. There's a fuckload of pencil pushers / knowledge workers with 100k student loans whose lifetime contribution can probably be measured in a few hundred dollars in tokens. And TBH normalizing AI crutch for kids is going to make large % of future cohorts even more replaceable. Skill atrophy among youth is declining hard, but AI is basically crippling future workforce quality to make their displacement even easier. There's even less reason to hire entry level in 4 years not just because models get better but human capita is going to be so much worse.
The market hasnât been built out yet. Thereâs that post from a couple days ago where someone frontloaded the entire UX of an operating system onto an LLM, so you just tell the hardware what you want to do and it does it. https://news.ycombinator.com/item?id=47557165
The growth is there but itâs going to be a marathon, not a sprint. I donât know why everyoneâs in such a goddamn hurry all the time
feels like an insult to readers to try to pretend that their revenue per month is comparable to google or apples growth when the funding is absurdly different, not to mention inflation itself.
I am very much onboard with AI within my workflow. I just don't really see a future where openai/anthropic are the absolute front runners for devs though. Maybe OpenAI does just have the better vision by targeting the general public instead, and just competing to become the next google before google can just stay google?
What is their next step to ensure local models never overtake them? If i could use opus 4.6 as a local model isntead and wrap it in someone else's cli tool, i 100% do it today. are the future model's gonna be so far beyond in capability that this sounds foolish? the top models are more than enough to keep up with my own features before i can think of more... so how do they stretch further than that?
A side note i keep thinking about, how impossible is a world where open source base models are collectively trained similar to a proof of work style pool, and then smaller companies simply spin off their own finishing touches or whatever based on that base model? am i thinking of thinks too simplistically? is this not a possibility?
Anthropic is definitely gaining ground over OpenAI in the business world. Cowork is the absolute hotness right now, and even prompted MSFT to drop their own variant yesterday
Codex and Gemini CLI seem 1-2 months behind Claude Code. They will catch up. This race will eventually be won by whoever can come up with the cheapest compute.
I agree that that's what it would take, but compute would need to get very cheap for it to be feasible to keep models running locally. That's an awful lot of memory to have just sitting with the model running in it.
True. I was thinking more of power users. Do you think Opus level capabilities will run on your average laptop in a year? I think that's pretty far away if ever.
You can demonstrate "running" the latest open Kimi or GLM model on a top-of-the-line laptop at very low throughput (Kimi at 2 tok/s, which is slow when you account for thinking time) today, courtesy of Flash-MoE with SSD weights offload. That's not Opus-like, it's not an "average" laptop and it's not really usable for non-niche purposes due to the low throughput. But it's impressive in a way, and it does give a nice idea of what might be feasible down the line.
> how impossible is a world where open source base models are collectively trained similar to a proof of work style pool
Current multi-GPU training setups assume much higher bandwidth (and lower latency) between the GPUs than you can get with an internet connection. Even cross-datacenter training isn't really practical.
LLM training isn't embarrassingly parallel, not like crypto mining is for example. It's not like you can just add more nodes to the mix and magically get speedups. You can get a lot out of parallelism, certainly, but it's not as straightforward and requires work to fully utilize.
It's hard to train models in the open. All the big players are using lots of "dodgy" training data. Like books, video, code, destinations. If you did that in the open, the lawyers would shut you down.
Though I think these companies are wildly overvalued, I don't see LLMs as a service going away in the future. The value in OpenAI is that it provides extra compute, data access, etc. My money is on local AI becoming more of a thing, while services like OpenAI still exist for local AIs to consult with. If a local model can somehow know that it's out of it's depth on a question/prompt, it can ask an OpenAI model if it's available, but otherwise still work locally if OpenAI fails to respond or goes out of business. To me that makes a lot more sense than the future being either-or.
> What is their next step to ensure local models never overtake them?
As someone who experiments with local models a lot, I donât see this as a threat. Running LLMs on big server hardware will always be faster and higher quality than what we can fit on our laptops.
Even in the future when there are open weight models that
I can run on my laptop that match todayâs Opus, I would still be using a hosted variant for most work because it will be faster, higher quality, and not make my laptop or GPU turn into a furnace every time I run a query.
If your laptop overheats when you push your GPU, you can buy purpose-built "gaming" laptops that are at least nominally intended to sustain those workloads with much better cooling. Of course, running your inference on a homelab platform deployed for that purpose, without the thermal constraints of a laptop, is also possible.
I didn't say it overheats. It gets hot and the fans blow, neither of which are enjoyable.
MacBook Pro laptops are preferred over "gaming" laptops for LLM use because they have large unified memory with high bandwidth. No gaming laptop can give you as much high-bandwidth LLM memory as a MacBook Pro or an AMD Strix Halo integrated system. The discrete gaming GPUs are optimized for gaming with relatively smaller VRAM.
Does anyone know if this, like the spacex/xAI stock â will list on nasdaq? And will it be part of every market cap weighted index fund like VOO(S&P) etc or just for nasdax-100 etfs?
The exchanges are bending head over heels to accommodate these IPOs[1] and make our retirement index funds the exit-liquidity strategy to the thievery of pump and dump actors that buy it low and then sell high? As i understand the way thievery works is:
1. List at many multiples of market valuation on an exchange. So if you company is just 10 billion$ nasdaq and theives collude and say "can make it 100 billion..".
2. Lots of institutional investors and rich billionaires get stock options.
3. All market weighted index funds â aka all *your* low expense ratio ETF money â have to re-balance and buy them, raising their value: the exit-liquidity event
4. Rich A**** get richer by making an profit by selling higher.
isn't it weird that there is no attribution to a human here? i mean, eventually, they have to dropkick sama and install GPT itself as king, right? EOQ seems as good a time as any
the only lesson the common man should take from these valuations: start to protest against AI conpanies being included in the sp500!
unless you're a private investors in these preIPO, the whole plan is to get big enough, get forced entry into indexes, and leave early with everyone else holding the bag.
What will happen first? The Singularity arrives, and hyper-intelligent AI causes such rapid technological change that the world becomes unrecognizable overnight?
Or OpenAI pays off it's investors? Lol.
I am not sure if I believe in the Singularity or not. But it's kind of the best story ever to support the game of musical chairs that is Silicon Valley investing.
There is a lot of talk about the AI bubble. I think there are comparisons to the late 90's/early 00's here with early stars rising quickly but, ultimately, falling. Since essentially everything touches the internet now it is clear that the 'internet bubble' was more of a shakeup of companies than a real over-hype of the internet. That, I think, is at play right now too with the 'AI bubble'. AI isn't going away but some of the early stars may not make it.
So, the real question here is: Is OpenAI Netscape, or are they Google?
Almost a trillion for a company that hasnât proved it can reach any form of profitability, all on the promise of an elusive messianic concept of machine superintelligence through probabilistic algorithms.
Basically like praying cancer away.
Peak magical thinking, peak America.
I'm seeing diminishing returns, though in fairness we have no idea yet how to integrate properly with existing good practices and principles. I suspect improvement is going to come mainly from improved took usage rather than more impressive models.
I feel that too, every technology has its limits.
I use AI daily. But I canât see the âintelligenceâ.
All I see is fine tuning and bigger datasets.
Yesterday I asked claude to fix the color issues of graph. It failed miserably.
Opus 4.6 wasnât able to figure out why the text was grey. It made something up, instead of realizing the problem was simple, oklch wrapped inside a hsl color. hsl(oklch(âŚ))
I easily figured this out by just looking at the css and adding some logs to js.
This is not intelligence. This is a tool thatâs smart. Not sentient. AGI wonât be achieved by scaling alone.
Anthropic doesn't have anything else other than the Claude models.
But notice that no-one, not a single mention of Deepseek tells me that they are preparing to scare everyone again. Which is why Dario continues to scare-monger on local models.
Sometimes you do not need hundreds of billions of dollars for inference when it can be done locally with efficient software; and Google proved that. But where is the money in that? So continues the flawed belief in infinitely buying GPUs to scale which Nvidia needs you to do.
Only a matter of time for local models to reach Opus level. We are 1 or at most 2 years behind that and Anthropic knows that.
You have a GPU already (at least an iGPU and an NPU on most newer platforms) as part of your computer, might as well get some use out of it with local inference. And trying to do inference on a larger model with an undersized GPU will have you idling a lot less than 99% - but that still makes a lot of sense for most casual users who will only rarely need a genuine "Pro" class answer from AI. Doing that locally is way less hassle than paying for a subscription or messing with API spend.
No mention of "AGI" this time. Since we all knew it was a scam. But this is the most damning of them all:
> The OpenAI flywheel is simple. More compute drives more intelligent models. More intelligent models drive better products. Better products drive faster adoption, more revenue and more cashflow.
FTX had a "flywheel". It fell off. Being saddled with hundreds of billions of debt makes this situation ten times worse.
"Commited capital" - is this the same commited as the $500B for the Stargate project ?
Not gonna lie, I hate those announcements lately. It's full bullshit mode, worse than the Dot-com bubble. Numbers don't make any sense, any more, and yet journalist don't ask any real questions...
They have to focus on the distant future (where they are frankly unlikely to exist) because they are falling further and further behind in the immediate future.
Their latest desperate bid for relevance is a plugin for Claude Code that uses Codex as a second opinion. Please clap.
This a big exaggeration. Codex is probably one of the top two LLM programming tools, along with Claude Code. GPT-5.4 models are strong, unlike the initial GPT-5 ones, which were comparatively bad, and can hold up against Opus 4.6. In my experience, they are better at analytical work.
I cannot really see how they are "far behind," or how some plugin for Claude Code is a "last desperate bid." The tools are close enough to each other that I regularly use Codex one month and Claude Code the next without much disruption, just to try out any new models or features that might be available.
I do not have much visibility into the non-code applications, so maybe it is stickier there.
If/when the AI bubble pops and takes OpenAI down with it, I would not expect Anthropic to come out unscathed either.
> The OpenAI flywheel is simple. More compute drives more intelligent models. More intelligent models drive better products. Better products drive faster adoption, more revenue and more cashflow. That gives us the ability to reinvest and deliver intelligence more efficiently to consumers, enterprises, and builders around the world.
-x-
In short, the musical chairs are still playing... Keep on walkin' round, y'all, till the music stops.
Fortunately, I was afraid that this was another bubble. /s
I wonder what the bet is here, long term that valuation is going to have to go up even further for this investment to make sense so they're clearly betting that at IPO time they'll be able to convincingly demonstrate AGI or something extremely close to it. That's a pretty risky bet, and meanwhile, whatever they come up with will be a commodity within a year. And that's besides OpenAI no longer being seen as the dominant player or the player with the best edge.
What what? Are you surprised it's that low, that high, that they can tell what their revenue is, that they report it on a monthly rather than annual basis, or something totally different?
It's going to be pretty hard to get a good answer to whatever you're having difficulties understanding if you can't be bothered to write more than a word.
> Within a year of launching ChatGPT, we reached $1B in revenue. By the end of 2024 we were generating $1B per quarter. We are now generating $2B in revenue per month.
They raised $122B.
122 / 12*2 = 5 years to get your money back (I simplify, I know revenue <> profit)
They are so big that almost no one can afford to acquire them. It is similar as someone would like to acquire MSFT or AAPL.
$2b/month which is $24b/year. Not as much as I expected considering they were at $20b by end of 2025.[0] They only added $4b since?
Anthropic had $19b by end of February 2026 and they added $6b in February alone.[1] This means if they added another $6b in March, they're higher than OpenAI already.
However, I heard that OpenAI and Anthropic report revenue in a different way. OpenAI takes 20% of revenue from Azure sales and reports revenue on that 20%. Anthropic reports all revenue, including AWS's share.[2]
[0]https://www.reuters.com/business/openai-cfo-says-annualized-...
[1]https://finance.yahoo.com/news/anthropic-arr-surges-19-billi...
[2]https://x.com/EthanChoi7/status/2036638459868385394
They aren't reporting anything yet. What we hearing is just from news media who get their leaks/info from investors who get some form of IR reports/ presentation.
Both will do public reporting only when they IPO[4] and have regulatory requirement to do so every quarter. For private companies[1] reporting to investors there are no fixed rules really[3]
Even for public companies, there is fair amount of leeway on how GAAP[2]expects recognize revenue. The two ways you highlight is how you account for GMV- Gross Merchandise Value.
The operating margin becomes very less so multiples on absolute revenue gets impacted when you consider GMV as revenue.
For example if you consider GMV in revenue then AMZN only trades at ~3x ($2.25T/$~800B )to say MSFT($2.75T/$300B) and GOOG ($3.4T/$400B) who both trade at 9x their revenue.
While roughly similar in maturity, size, growth potential and even large overlap of directly competing businesses, there is huge (3x / 9x) difference because AMZN's number includes with GMV in retail that GOOG and MSFT do not have in same size in theirs.
---
[1] There are still a lot of rules reporting to IRS and other government entities, but that information we (and news media) get is from investors not leaks from government reporting - which would be typically be private and illegal to disclose to public.
[2] And the Big 4 who sign off on the audit for companies prefer to account for it.
[3] As long as it is not explicit fraud or cooking the books, i.e. they are transparent about their methods.
[4] Strictly this would be covered in the prospectus(S-1) few weeks before going public and that is first real look we get into the details.
Does the GAAP accounting matter if everyone passively buys shares due to the new fast entry rules, which corruptly will force us all to buy into these companies? The fundamentals and true value seem less relevant than ever:
https://www.benzinga.com/markets/tech/26/03/51248353/michael...
For other readers, I want to add some context here. NASDAQ is pondering whether or not to change their NASDAQ 100 index membership rules for IPOs. Currently, there is a three month waiting rule for IPOs. They are proposing (not sure if passed/agree/completed yet) to remove this waiting rule for IPOs.
Real question: What is the real impact of this rule change? To me, it seems so minor. Three months is just a blip in time for any long term investor.
Why is this "corrupt"? That term makes no sense here.Also, if you don't like the NASDAQ 100 rules, then you don't have to invest in securities that track it. You can trade the basket yourself minus the names that you don't like.
Finally, I would say that S&P 500 index is far more important than NASDAQ 100. To join the S&P 500 index, the name must be profitable for the most recent year. (four quarters). Recall that Uber IPO'd in 2019, but was not profitable until 2023. OpenAI probably will not be profitable when it goes public; thus, it will not join the S&P 500 immediately.
I think the bigger story is SpaceX. It will likely IPO very close to a 1T USD market cap (with a small float: ~10%). And, thanks to StarLink, I assume that SpaceX is now wildly profitable.
what would force you? I guess if you are a greedy bastard you might feel that way...
Diluting the index entry rules, only devalues the index utility. When it becomes a bigger problem, other indices with higher quality controls will out compete the current ones and be used by asset managers seeking safety.
More likely than not, most of us are already holding stock in these companies one way or another. All the Mag 7 hold a major chunk of OAI and Anthropic stock anyway, slower entry does not make it less risky for us.
Even if the big tech companies did not hold any stock, they are still the biggest vendors and their own order books is hugely impacted by the AI demand from these two ( and others in this space), either way we are all in this together.
> When it becomes a bigger problem, other indices with higher quality controls will out compete the current ones and be used by asset managers seeking safety
Doubt it.
The world does not allow perfect competition.
lol imagine someone believing in the invisible hand of the free market in 2026
In the short term there are distortions and inefficiencies. It may feel like free market is done .
However in the long term, economics usually finds the most efficient way.
Maintaining inefficient structures like tariffs or monopolies becomes more and more expensive and eventually untenable and disruptions will occur.
In the long term we are all dead. (Keynes)
Really feels like 1928
I personally find this is the correct solution, since indexes are over-inflated either way, this brings much needed sanity to the index. Your index is now worth much more or much less based on how you view the AI bubble and you are forced to understand and correct your forward looking investments accordingly.
Passive investments are good, but if taken too far as they clearly have been in the last decade they become a scam. Everyone is SIPing into it, and there is infinite liquidity. Until one big whale finally decides they are booking it, then all hell will break loose on the same damn day.
Yes gaap absolutely matters.
You can just choose not to play the accounting game, and only choose the ones that actually gaap viable as investment opportunities. For example mag7 - tesla are all relatively cheap when they dip.
Some times the best play is just not to play. If you think they are too risky, walk away. There are enough good oppotunities
How much do these names need to "dip" for you to consider them cheap?
The $19b ARR and $6b added in Feb came directly from Anthropic CEO recently.
Until theyâre using consistent methods of reporting those figures, theyâre not comparable. Same as any other company pre vs post IPO
Was referring to this:
Announcing isn't reporting. Am I the only one old enough to remember Enron?
I am reminded of the "I declare bankruptcy" meme from the 2000's TV series Office.
When we say reporting it means there are statutory submissions with an auditor signing off, with legal liability. As the other reply referenced consequences for doing this incorrectly can be severe - Arthur Anderson is no more after all because of Enron.
A Press Release (of a private entity) does not have to satisfy this high bar.
Press release does mean no constraints, for public companies, disclosure of important information by officers and other insiders have strong controls. Even if its the just a rocket/poop emoji on a casual social media platform. Lawyers have to refile with the SEC in the expected format. Even private companies have restrictions on not claiming things fraudulently to investors, but these are accredited investors with lesser controls than retail.
30x revenues at 17% revenue growth is... aggressive.
Except it's not 100x revenues, and it's not 17% growth. I don't know where you got those numbers from?
The numbers OpenAI gave in the post would mean a 30x multiple pre-money. And the $20B -> $24B run-rate growth since the start of the year could plausibly mean anything from 110% to 200% annualized growth rate, depending on whether that happened over two or three months. The $24B is a lower bound as well, since they only gave use one significant digit for the monthly revenue.
You're right, I was thinking about 100x revenues and forgot to confirm the math. Updated to reflect your point. ChatGPT itself provided the 17% number (it's most recently available growth rate)...
OpenAI is a few years behind Anthropic, and it's unlikely they'll catch up at this point.
I'm following this very closly and i'm stunned. Any infos on why you think they are behind antrophic in years?
I do see less quality from reasoning at chatgpt compared to Gemini but otherwise i'm not seeing a year or years gap.
Where exactly are they behind?
everywhere, but most important in ethics
your ethics.
let's not forget that these major LLMs are all the children of corporate hyper-piracy en masse, none of them are ethical even in origin unless you're talking about the pre-product company charter kind of ethics, like google .
Theyâre about even in general, but for me OpenAI is slightly or significantly ahead in the areas I care about the most. E.g. claude code is a backend slop cannon if you donât tell codex/gemini to review the outputs.
Still a huge amount of revenue for any company. Those $20/month fees are going to triple in a couple yrs. But the VCs expect much much more.
Friends of mine working in AI companies are saying weâll be lucky if they only triple. More like 10-20x long term, especially for enterprise
Oh, I read it as the number of subscribers would triple, but you're suggesting the price will?
That makes a little more sense, because the number of subscribers are so low that tripling won't really make much difference in terms of turning a profit.
And that is revenue only. In the past 15 or so years most US companies (and especially startups) always talk about revenue only. Wheras only profit should matter.
E.g. what good is 20 billion per year when "OpenAI is targeting roughly $600 billion in total compute spending through 2030". That is $150 billion per year?
The startup game is about building assets and then cashing out on them during exit.
Assets are harder to measure. Facebook used to say something silly like every user was worth $100. That sounded ridiculous for a completely free app but over a decade later, the company is worth more than that. Revenue is an easier way of measuring assets than profit.
Profit doesn't really matter. It gets taxed. But it's not about dodging taxes; it's because sitting on a pile of money is inefficient. They can hire people. They can buy hardware. They can give discounts to users with high CLTV. They can acquire instead of building. It's healthy to have profit close to $0, if not slightly negative. If revenues fall or costs increase, they can make up for the difference by just firing people or cutting unprofitable projects.
Also when they're raising money, it makes absolutely no sense to be profitable. If they were profitable, why would they raise money? Just use the profits.
> Wheras only profit should matter
Profit is money you couldnât figure out how to spend. During growth, you want positive operating margins with nominal profits. When the company/market matures, you want pure profits because shareholders like money. If you can find a way to invest those profits in new areas of growth, thatâs better.
> Profit is money you couldnât figure out how to spend.
Profit is the money showing your business is sustainable. Ever since the ZIRP era US companies keep haemorrhaging money at a rate that is physically impossible to recoup.
If OpenAI plans to lose 100+ billion dollars per year for half a decade, what profits are you talking about to offset the losses?
> When the company/market matures, you want pure profits because shareholders like money.
Ah yes. Shareholders like money. And not, you know, basic accounting like "we need money to actually pay salaries, pay for equipment and offices etc. without perpetually relying on seeming endless investor money".
> what profits are you talking about to offset the losses?
You donât need profit to offset the losses.
You can simply reduce spending / expenses.
In principle yes, but all metrics so far suggest they are losing money every user interaction. There is very little network effect with these tools so It's not like they can start cutting back on staff and feature deployment.
lol thatâs a line so incredibly naive it hurts.
One does not âsimplyâ reduce spending.
> One does not âsimplyâ reduce spending.
Why does stock price go up after mass layoffs?
Not sure why youâre downvoted.
Everyone wants to treat OpenAI like a car wash business where they need to make a profit almost immediately. I donât know why people canât understand that the industry is in a rapid growth stage and investing the money is more important than making a profit now. The profits will come later.
"Profits will come later" https://news.ycombinator.com/item?id=47597480
It's not as much as you think. Google is spending $185b on data centers this year alone. Amazon is spending $200b this year. Total capex for big tech is ~$700b in 2026 and we're not including neo clouds, Chinese clouds, and other sovereign data centers.
Since everyone is trying to get compute from anywhere they can, including OpenAI going to Google, it's hard to tell what is used internally vs externally.
For example, it's entirely possible that Google's internal roadmap for Gemini sees it using $600b of compute through 2030 as well. In that case, OpenAI needs to match since compute is revenue.
But if Gemini doesn't end up using the compute because of whatever reason, Google has other ways to monetize that compute. OpenAI doesn't?
So the same money spent by OpenAI and Google doesn't carry nearly the same amount of risk?
Give me a billion and I'll have 500M of revenue in no time by selling dollars at 50 cents.
Why are we treating OpenAI and Anthropic differently than say, Amazon or Uber? Both companies invested in growth for many years before making a profit. Most tech companies in the last 2-3 decades lost money for years before making a profit.
Why are we saying that OpenAI and Anthropic can't do the same?
Amazon had a clear business model. They had positive gross margin from, if not day 1, then pretty close to it.
I remain skeptical of Uber.
Sure, maybe OpenAI and Anthropic will make it work. It's not impossible. But it's far from guaranteed.
>Most tech companies in the last 2-3 decades lost money for years
Yes
>before making a profit.
No
Two reasons. They somewhat broke even, and kept getting investment. The potential for quasi monopoly was obvious.
Openai can't claim either.
How did Uber somewhat break even? They lost $34b before making a profit.
Uber was only on a path to monopoly in the US, not world wide. Itâs lost to local competitors in most countries. And it can get disrupted by self driving cars soon.
OpenAIâs SOTA LLM training smells like a natural monopoly or duopoly to me. The cost to train the smartest models keep increasing. Most competitors will bow out as they do not have the revenue to keep competing. You can already see this with a few labs looking for a niche instead of competing head on with Anthropic and OpenAI.
The cost of copying SOTA models though is super cheap and doesnât take super long.
How do you distill when OpenAI and Anthropic inevitably move to tasks running in the cloud? IE. Go buy this extremely hard to get concert ticket for me.
Distilling might only be effective in the chat bot dominant era. We are about to move to an agents era.
Furthermore, Iâm guessing distilling will get harder and harder. Claude Code leak shows some primitive anti distilling methods already. Thereâs research showing that models know when itâs being benchmarked. Whoâs to say Anthropic and OpenAI arenât able to detect when their models are being distilled?
Worse, Google can afford to outspend them in this game and basically run them both out of money.
It's not even remotely comparable. Uber burnt some $30B over a decade or so.
It seems like it is comparable based on what you just said.
OpenAI have burned nearly 25 times what Uber did, it has more competitors, billions of dollars in obligations and no clear way to profitability.
The problem for OpenAI is that the cost of getting them where they are now has been to high and competitors can now establish themselves for much less money.
> Why are we treating OpenAI and Anthropic differently than say, Amazon or Uber?
The dame Uber that lost close to 30 billion dollars over 10 years to subsidize its price dumping?
No, no we are not treating OpenAI differently than Uber
why should only profits matter? if i had a killer product today that i just need to sell tomorrow, wouldn't you still invest today knowing i'll probably only start to make money tomorrow (or perhaps next week)?
the expectation is that they'll eventually make money. they can't raise forever. only startups are not profitable for a few years. but most companies that have existed for a long while have been profitable
and since they're expected to make a LOT of money, everyone wants a piece of that future pie, pushing up the valuation and amount raised to admittedly somewhat delusional levels like here
> why should only profits matter?
In this case because it's not clear that anybody has actually figured out how to sell inference for more than it costs
It's well know everyone is making great money on inference. The cost is training.
Whether GPT-5 was profitable to run depends on which profit margin youâre talking about. If we subtract the cost of compute from revenue to calculate the gross margin (on an accounting basis),2 it seems to be about 30% â lower than the norm for software companies (where 60-80% is typical) but still higher than many industries.
(They go on to point out that there are other costs that might mean they didn't break even on other costs - although I suspect these costs should be partially amortized over the whole GPT 5.x series, not just 5.0)
https://epochai.substack.com/p/can-ai-companies-become-profi...
https://martinalderson.com/posts/are-openai-and-anthropic-re... (with math working backwards from GPU capacity)
"Most of what we're building out at this point is the inference [...] We're profitable on inference. If we didn't pay for training, we'd be a very profitable company"
https://simonwillison.net/2025/Aug/17/sam-altman/
"Thereâs a bright spot, however. OpenAI has gotten more efficient at serving paying users: Its compute marginâthe revenue left after subtracting the cost of running AI models for those customersâwas roughly 70% in October, an increase from about 52% at the end of last year and roughly 35% in January 2024."
https://archive.is/OqIny#selection-1279.0-1279.305 (Note this is after having to pay higher spot rates for compute because of higher than expected demand)
> It's well know everyone is making great money on inference.
That is not, in fact, "well known", but based entirely on the announcements of the inference providers themselves who also get very cagey when asked to show their work and at least look like they're soliciting a constant firehose of investment money simply to keep the lights on. In particular there's a troubling tendency to call revenue "recurring" before it actually, you know, recurs.
> based entirely on the announcements of the inference providers themselves who also get very cagey when asked to show their work
I mean sure, it's self reported.
But the inference prices somewhere like Fireworks or TogetherAI charges is comparable to what Google/AWS/Azure charge for the same model an we know they aren't losing money - they have public accounts that show it, eg:
https://au.finance.yahoo.com/news/wall-street-resets-amazon-...
Fireworksâ gross marginâgross profit as a percentage of revenueâis roughly 50%, according to the same person
https://archive.is/Y26lA#selection-1249.65-1249.173
> In particular there's a troubling tendency to call revenue "recurring" before it actually, you know, recurs.
If someone has a subscription then yes that is pretty normal.
> If someone has a subscription then yes that is pretty normal.
Not if you've substantively changed rate limits 3 times in the last 5 months while still counting those forecast revenues. In most industries that's called rug-pulling.
It doesnât matter how you call it. A recurring subscription on the books is a recurring subscription. Yes you can cancel anytime (how generous of them), it also doesnât matter.
not if your product is selling two dollars for one dollar and as soon as you'll start to charge more I'll switch to one of your twenty competitors
profit isn't a function of having a killer product, it's a function of having no competition
And why do you think twenty competitors can stay competitive for years to come?
Industries always consolidate and winners emerge. SOTA LLMs look like a natural monopoly or duopoly to me because the cost to train the next model keeps going up such that it won't make sense for 20 competitors to compete at the very high end.
TSMC is a perfect example of this. Fab costs double every 4 years (Rockâs Law). It's almost impossible to compete against TSMC because no one has the customer base to generate enough revenue to build the next generation of fabs - except those who are propped up by governments such as Intel and Rapidus. Samsung is basically the SK government.
I donât see how companies can catch OpenAI or Anthropic without the strong revenue growth.
Google has already surpassed them both in all areas except coding. People on HN only look at benchmarks, but Gemini's multimodal understanding, things like identifying what a plant is, normal user use cases (other than chatting), integration with other tools, is much better.
It's believable that Meta, ByteDance, etc. can catch up too. It is not certain that scaling will meaningfully increase performance indefinitely, and if it stops soon, they surely will. Furthermore, other market conditions (US political instability) can enable even more labs, like Mistral, to serve as compelling alternatives.
Uber, TSMC, etc. have strong moats in the form of physical goods and factories. LLMs have nothing even remotely comparable. The main moat is in knowledge, which is easy to transfer between labs. Do you think all the money that goes into training a model goes into the actual final training run? No, it is mostly experiments and failed ideas, which do not have to be repeated by future labs and offshoots.
> It is not certain that scaling will meaningfully increase performance indefinitely
It's certain that it won't. We've already hit diminishing returns.
Google has completely caught OpenAI. Anthropic has a better coding model, but I'm sure Google is working on that too.
> Anthropic has a better coding model
Iâll be polite and call this statement âa very debatableâ one.
The barrier to replicating TSMC isn't just cost, it's supply chain, geopolitics, and talent.
Only one company on Earth can make the UV lithography machines TSMC buys for their highest end fabs, and they're not selling to anyone else.
The PRC tried to brute force this supply chain backed by the full might of the Party's blank check, all red tape cut, literally the best possible duplication scenario, and they failed.
The PRC didn't fail, they haven't finished succeeding yet.
They will succeed eventually since they have proof itâs possible and their plans span decades. I expect them to have working EUV in 10 years. Whether itâll still be bleeding edge tech is a different question I dare not guess the answer to.
>Industries always consolidate and winners emerge.
no, most industries just sell boring generic products, a few industries favor monopolists. Semiconductors are one of them but LLMs are also as far removed from that business as is physically possible.
TSMC makes the most complicated machines humans have ever built, a LLM requires a few dozen nerds, a power plant, a few thousand lines of python and chips. That's why if you're Elon Musk you could buy all of the above and train yourself an LLM in a month.
LLMs are comically simple pieces of software, they're just big. But anyone with a billion dollars can have one, they're all going to be commoditized and free in due time, like search. Copying a lithography machine is difficult, copying software is easy. that's why Google burrowed itself into email, and browsers, and your phone's OS. Problem for openai is they don't have any of that, there's already half a dozen companies that, for 99% of people, do what they do.
no competition is a bit extreme. Limited competition yes due to competitive advantages.
What is the point - exactly - of profit?
Profit is money you can't find a use for to grow your business, so you give some of it to the government in the form of tax.
Also there is a big difference between operational expenses and capital expenses like building data centers.
I think OpenAI is being very aggressive on the growth vs conservative financial management spectrum but just saying "only profit should matter" is just wrong.
> What is the point - exactly - of profit?
It's what attracts capital investment, which businesses need
OpenAI seems to do reasonably well at attracting capital investment without profits.
As did Amazon, Google, Meta etc etc.
OpenAI is great at attracting people who say "yeah, sure, I'll give you capital at some point in the future" who then never actually give them the capital (or at least haven't yet).
They seem to be spending lot of cash too...
If I remember correctly, Facebook took 10 years raising money before going ipo.
Could be wrong though.
What's the point - exactly - of a company being sustainable?
Being profitable isn't the same as sustainable.
Even a simple shop isn't profitable for months if it needs to buy stock up front, and run some ads to let people know about it. The money for that comes from the shop owners as an investment.
This is the same thing but on a slightly bigger scale, over a longer time frame.
If your shop is unprofitable for years with no chance to recoup any of the costs, you close it, as your investments run out, and investors and banks stop giving you money as you keep losing them.
US tech companies just continue operating because "revenue and growth".
> Today, we closed our latest funding round with $122 billion in committed capital at a post money valuation of $852 billion.
A couple things that stand out to me about this is the use of the phrase "committed capital", which only sounds like a promise that could break from various circumstances, and the valuation of their funding keeps changing so it sounds like a max rather than the valuation every investor invested at.
I do wonder how much of Amazon's $50B share (per last press release) is in AWS credits rather than money in the bank.
To then claim that Trainium is âsellingâ and not a dud? Iâd bet a lot.
Probably a lot? It would be much more tax-advantageous to do it this way, $50B worth of credits != $50B worth of spend on Amazon's part, and they might meet in the middle about how much equity that translates to.
I can see a lot of advantages for Amazon, but I don't see why it would be tax-advantageous.
Situation A:
You're Amazon. You give OpenAI $50B cash investment, they then hand you back the $50B over time because they buy $50B worth of Amazon AWS services (they would use AWS or other equivalent compute anyway). OpenAI pays an additional $1-5B in sales taxes on top of their $50B compute purchase. Now let's say you have $25B opex for said compute. You then have $25B profits, you pay 21% corporate taxes on the profits, so you too owe the government about $5B. Government collects around $6-10B on this whole transaction.
Situation B:
You're Amazon. You let OpenAI use your services by handing them API credentials that unlock what would normally cost $50B worth of services, but no money changes hands. You have zero revenue from the transaction, write off the $25B opex as a tax loss on your other profits elsewhere in the company. You thus pay ~$5B less tax on your other income as a company, and OpenAI also doesn't have to pay sales tax because they didn't actually purchase anything.
You have to report barter transactions as income. And Amazon already pays 0% corporate income tax.
Isnât sales tax only for consumers? Ie companies reverse charge sales tax or omit it entirely. Or what is this 2% - 10% sales tax youâre referring to?
> Isnât sales tax only for consumers?
It depends how you defined âconsumersâ. If you mean âthose who consume the good subject to the tax, rather than people who resell the goodâ, yes, ideally.
If you mean ânot businessesâ or âindividuals but not corporationsâ, then, no.
> Ie companies reverse charge sales tax or omit it entirely.
Generally, the theory of sales taxes is that people (including corporations) pay sales tax on things they consume as a final good rather than use as an intermediate good in production or simply resell. The exact way in which that is determined varies somewhat between jurisdictions with sales taxes, but generally (assuming paper is subject to sales tax in a jurisdiction), if you are buying paper to print books that you sell, you don't pay sales tax, if you are buying paper to print internal documents that you use in running the business, you do pay sales tax.
good catch! committed capital is not same as we raised.
It makes sense for such a huge amount to be "committed", not sitting idle in a bank somewhere.
Thatâs typical. Large funding rounds usually arenât delivered as one single giant lump sum into the bank account. The capital is committed in stages that can depend on hitting milestones or goals.
This is done even in smaller startup funding rounds some times.
Fair, I think a lot of what I've been perceiving is the gymnastics in how funding and valuation and deals get reported. There ends up being a ton of asterisks that makes the headline news deviate quite significantly from reality, e.g. https://arstechnica.com/information-technology/2026/02/five-...
that's why they have to open through banks and other less valuable more sliced share system.
This announcement completes the betrayal of their founding principles.
"Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return."
Is there a way out of this though? The seed is already planted and everything else is just a PR.
Founding principles = initial marketing strategy
> The broad consumer reach of ChatGPT creates a powerful distribution channel into the workplace
They mention this line in different forms a couple of times in the article. Itâs clear theyâre pretty rattled about Anthropicâs momentum in enterprise, I wonder how confident they really are in this rationale.
It's an interesting strategy, I see a pretty big risk from them leaning into it like this. We already have a vibe in my circles of the old "gmail vs yahoo" type thing where if you saw someone had a yahoo mail address you assumed they were technologically illiterate. Similarly it's mildly embarrassing already to say you used ChatGPT for something. It's not unrecoverable, but it's a pretty steep slippery slope they probably don't want to be anywhere near if they care about enterprise.
After the DoD moves? It is not just the technologically illiterate, it is part of the US culture war. OpenAI is the MAGA brand, like tesla.
What an absurd leap. OpenAI is the maga brand because they simp for government contracts like any massive company would?
Did you miss the cancel/unsubscribe gpt boycott? It was only about a month ago. Many people I know cancelled/unsubscribed. To be fair though, most people I have talked with needed almost no encouragement to move to anthropic or google (better products, easy to switch etc). Consumer sentiment can change quickly.
Many people in your bubble cancelled.
I don't know if it's guaranteed to work, but the strategy is real. I know Notion won vs. competitors in the space because they focused on consumer first, and consumers then brought Notion into their workplaces.
I am amused that you think IT is going to respond to an unmanaged LLM tool that operates outside of the LLM policies all serious enterprises have set up by now and say 'wow, that is cool and maybe we should buy in to this!'
What is going to happen is that the emplyee who tries to sneak OpenAI into our org is going to have two meetings set up by the end of the day, one with IT to ensure the whatever tool they installed is burned out with fire and one with HR to ensure they know the company policy and acknowledge that another fuck-up like this is a firing offense.
Kind of makes me wonder how 'accelerated' the timeline of publishing this article was based upon the Claude Code leak today. Considering everyone has gotten a sneak peek at what Anthropic is working on OpenAI might be a little worried. This could also just be coincidence, but this piece really does read like self-encouraging fluff.
The timing of this coming out today is cause today is the last day of the month/quarter and has nothing to do with Claude.
Ah, yeah that makes way more sense, I always forget about financial quarter timings.
0. These things take an enormous amount of time and coordinating to release.
No, they didn't raise $122B as the HN title implies. A big chunk of that $122B is a "maybe" that depends on various things that need to happen in the future.
Oh, man... I can't wait to see where this is going. Might not be pretty after all.
I've wondered how many announced fundraising rounds were like this. It's in everyone's interest (VCs and entrepreneurs) if the message to the outside world is "this company is amazing so they've raised a boatload of cash". But VCs might not want to give it all up front, or unconditionally.
It makes it hard to say what the valuation of a company is. If the milestones are unlikely to be hit, then it's anyone's guess.
This is a common structure. It's confusing to people who don't know finance or startups when they first see it.
Even VCs don't get all of their fund money delivered into their bank account when they raise a funding round. It's inefficient and undesirable for everyone involved to have to move all of the money up-front, at once.
If you talk to anyone in startup funding or finance they'll be familiar with the term "capital call" which describes how committed capital obligations are delivered at a later date than the initial deal: https://en.wikipedia.org/wiki/Capital_call
I've been involved in many startups, and this type of fundraising is not common, or at least it wasn't common before a few years or so ago
The whole concept of talking about "runway" is basically calculating how much cash in the bank, that is actually in your bank account, will last. And this arrangement is different, as there are contingencies. In the past, VCs would just give you money in a particular series, and then if your business did well, they'd eventually give you more money in a later series. But it wasn't like they announced it all up front in, say, a Series A, but a big chunk of the money would only be delivered if you met milestones.
Sure for like $5-10MM, but no one is landing $100B cash in a startup's bank account
$100B isn't a startup. And if there's a $100B deal, you better believe the cash is there. Case in point - Netflix/Paramount wanting to buy WB. Or the $44B that Musk had to raise to buy out Twitter shareholders.
Both your examples are purchases. Musk had to raise actual capital to buy Twitter because the people getting the money were taking it and walking away.
Funding doesn't work like that. Investors are giving you money as part of a longer-term deal where they stick around.
This was already common in tech for Series C+ fifteen years ago when I raised a round. Once youâre talking tens or hundreds of millions, almost everyone wants milestones and tranches instead of giving all the money up front.
Same. I know $100m+ range arrives in the bank account. Donât know more than that. But for that sum, I know it routinely just arrives.
Obviously this is 1000x as large so I make no claims to knowing that sum. But itâs routine for startup funding to arrive in bank account.
Both of CFS B rounds were cash, in recent years, and each in the range of "low billions". Sure another 2 orders of magnitude is another story, but so is selling hope. I'd say the latter is the thing that is unique here.
No, it's not common for the startup itself to make capital calls. The phrase (and your link) refers to capital calls made by VC firms to their limited partners. Same thing in PE.
> Amazon agreed to invest up to $50 billion in the startup
> Nvidia invested $30 billion
> Microsoft, one of OpenAIâs longtime partners, also participated
There is a lot of non-cash, never-will-be cash, investment here. Credits for compute.
I think more people are aware that VCs raise commitments for a fund that they can pull in via capital calls than are aware that startup funding from VCs come with hurdles to clear.
This is perhaps because the most common round to raise is a small/early one, and these tend not to have hurdles. Founders that only ever raised these rounds wouldn't necessarily know what happens in later/bigger rounds.
Also, I wonder if capital calls come with hurdles as well? That is, can an LP refuse to put in more money if the VC's recent investments have not done well? I would think not, since it typically takes many years to determine whether investments were good or not.
Gotta hit that high IRR as a fund manager and the clock starts when the cash comes in so capital calls are appreciated by fund managers. Unless they are emerging managers (the startup equivalent in finance) and their LPâs are less than institutional and ghost them when the capital call hits.
IRR is so trivial to manipulate - it'd be wonderful if more investors began demanding actual metrics on capital performance. If you're parking cash with an investment firm you want to know about how much of a return you can expect when it is withdrawn, and while history is a guide and not a guarantee, there are much better ways to inform that expected return than IRR. "My million got a return of 2% during a year when your reported IRR was 10% - where's the other 8%!" is a common cry from those who haven't just rolled over their investment, unaware of how little it has functionally appreciated.
I think both things are true at once:
1) For a $100bn round, you won't get a single transfer of $100bn into your checking account, this is normal
2) Sam Altman is a liar and people (correctly) don't really believe him when he starts throwing numbers around
Itâs confusing because itâs meant to be confusing. Bigger numbers are more impressive.
There's an accelerator here in Taiwan with a model I truly don't understand: 100k usd for 10%. 10%!! You've just valued the company at only 1 million! And taken a HUGE chunk of equity, not much left on the table!
Maybe it makes sustainable sense but in the world of venture capital it seems the most profitable thing to do is lie through a Cheshire grin, every day.
This is very standard, Ycombinator, which hosts this board, does the same: https://www.ycombinator.com/about
Not exactly the same...
> YC invests $500,000 in every company on standard terms. Our $500K investment is made on 2 separate safes:
> We invest $125,000 on a post-money safe in return for 7% of your company (the "$125k safe")
> We invest $375,000 on an uncapped safe with a Most Favored Nation ("MFN") provision (the "MFN safe")
So YC values the company at $1.8 million. I don't think that's so different.
Having large funding rounds contingent on meeting milestones is common. Always has been.
It just makes comparing funding rounds hard to understand, since money in the bank is money in the bank, and a lot of the "committed capital if you reach a milestone" is capital that would be easy to get if you reached that milestone, if it is sufficiently advanced, and has enough outs, etc., that you may as well have just raised another round in the future.
Note that even that "money in the bank" of traditional venture firm is not really money in the bank. VC, PE, and hedge fund managers usually don't have all the cash for the fund sitting in the bank at all times. Rather, their agreement with the LPs that fund the fund is structured as a series of capital calls: it gives the fund the right to demand that their LPs deposit cash in their bank accounts within 10-30 days, which can then be used to fund the investments that the VC firm makes. The capital calls are backed by legal documents enforceable in court, with pretty stiff penalties for failing to meet a capital call.
Such a funding structure here isn't all that different: the funding agreement gives OpenAI the right to call on their backers to make certain cash deposits, contingent upon milestones being met. Deep down inside, "money in the bank" doesn't actually exist, it's just mutual agreements backed by force of law.
When a startup raises money without contingencies, typically they do get a large amount of money in the bank all at once.
If investments are not tranched then the money is not delivered in tranches, yes.
The first rule of tautology club is...
Thatâs logically inconsistent. If the company was performing poorly enough that they couldnât meet their funding milestones from a previous round, theyâre not going to have an easy time raising the same money in a future round.
The milestones arenât a hard-stop that forbids the previous funding round participants from providing the money if they still choose. Itâs just an out.
sure they can. that's the whole point of the "pivot"
What I am saying is that if you do meet the milestones from your previous round, you're going to have an easy time fundraising anyway, so funding contingent on milestones isn't that different than just saying "well, if we need more money we can do another round"
Fundraising rounds are difficult, laborious, and distracting. It would be extremely different to try to multiply the number of rounds by 3-5X. There's nothing easy about that.
You're also ignoring that the market changes frequently. If you only raised as much money as you needed for the next 4-6 months with plans to re-raise all the time, you'd have to constantly be sizing your growth plans up or down based on how the market felt about startup investing that month.
Imagine the company having to either do speed hiring or large layoffs every few months to adjust to the size of the fundraising round they were able to get this time around.
Nothing about what you're suggesting would be easier, or easy at all
It seems pretty simple:
Funding Round A: VC âAâ invests 200M (100M immediately and another 100M if sales grow 10% or whatever)
At 6 months the company will either get the other 100M automatically (meaning they grew sales 10%) or they donât (meaning they grew less than 10%).
Assuming itâs the later they can then do another round during which they try to get the other 100M. In all likelihood VC âAâ wonât be interested (or interested at a lesser amount). They could go ask VC âBâ for an investment but it will likely be less than 100M as well because they didnât grow as much as âthe marketâ anticipated.
Nothing complicated at all.
Iâll give you $1 dollar for your banana today and another dollar in a week when it has ripened. If itâs rotten when I come to get my banana I wonât give you the other dollar. You have your original $1 and you can still try to sell your rotten banana to another HN reader but you probably wonât get another dollar this time.If instead you have a ripe banana Iâm sure you could easily find a buyer.
Why not announce the funding after the milestones have been met?
The funds are committed under the terms of the deal (share price, things like board seats, and other details). There are legal obligations to provide it.
This is a common structure for large investments. It would be really inefficient for all of these investors and companies to have to have the money sitting in cash to do a deal and then transfer it into the company's bank where it sits and earns interest for years until they can deploy it.
Even VC firms who raise funds work this way. The capital is "committed" but investors don't wire all of the money over right away so it can sit in the VC firm's bank accounts, waiting. The VCs do what's called a "capital call" through which they're legally bound to provide the money they committed when requested, under the terms of the deal.
It's splashier this way, and is meant to shape the narrative, make other companies fear their warchest, and make hiring easier. Of course, those who are in-the-know won't be fooled, but the perception of the general public will be set in stone by the PR framing.
Most people wonât look at the structure, theyâll just internalize that OpenAI is massively funded.
That effect kicks in well before the money actually does.
The assumption that's conveniently left out is that the milestones are realistic
One of the stipulations is that OpenAI achieves "AGI"... Need I say more?
Also a lot of this "money" is in cloud compute and credits not cash so...
that being said, how can Softbank keep throwing around all these astronomical numbers after so many bad investments? Leftover iPhone money?
Most people know Softbank as the company who lost billions on WeWork and not the company who made several more billions on the ARM IPO.
with these swings, I'm not sure how Son-san keeps himself from getting an ulcer
At some point it must just be Monopoly money.
Sir, itâs called investing
Yeah I checked their overall return. It's average. Not better than index funds IIRC.
They borrowed $40B from JP Morgan. They literally did not have the money otherwise.
also they need to pay back that in one year, so if OpenAI don't IPO this year they are screwed
Was curious about the source here. Seems widely reported and I just missed it. This a unpaywalled source I found
https://techcrunch.com/2026/03/27/why-softbanks-new-40b-loan...
they are Hwanging it
Their ~$50 million total Alibaba investment turned into ~$70 billion. As of two years ago they were still liquidating out of it.
January 26, 2024 - "Japanese investment holding firm SoftBank Group Corp has largely cleared its ownership in e-commerce giant Alibaba Group Holding, concluding one of the most successful deals in China's internet industry and a holding that spanned about 23 years."
"SoftBank, which invested US$20 million into Alibaba when it was still a start-up in 2000, said in a corporate filing on Thursday that it was set to book a gain of 1.26 trillion yen (US$8.5 billion) - about 425 times the value of its initial outlay - for the Tokyo-based firm's 2024 financial year after divesting its [remaining] shares via subsidiary Skybridge."
https://finance.yahoo.com/news/japans-softbank-concludes-run...
"As of two years ago they were still liquidating out of it"
I get that people are scared of investing in China. But if I still made single stock investments, I would seriously consider BABA, it seems well positioned.
Saudi oil money
Which might not be a thing anymore soon the way things are goingâŚ
Other way around - the Saudis are making bank.
well, until the Iranians blow up their refineries.
Iranians are also making bank. Why kick a hornets nest when you're winning?
> Why kick a hornets nest when you're winning?
Tell that to Trump and his glorious way of bombing Iran. Nothing against the idea itself, the Mullahs all but asked for it to happen.
But the execution? That was a level of dogshit I haven't seen in the time I was alive lol. Even Russia was better prepared with their invasion of Ukraine.
Both Trump and Netanyahu had a somewhat solid perspective on not getting utterly wasted in the next elections. Instead they go on one of the most ill-prepared wars in modern history, with results that may seriously upend the global economy if not lead us to WW3 outright.
cant stop winning!
Don't let reality get in the way of vibes
It's also like... >$50 billion in compute credits and discounted hardware between Amazon/Microsoft/NVidia. Which is all inside baseball since they simultaneously juice their financials with OpenAI's cloud compute bill
With NASDAQ and NYSE looking to reduce the timelines for new public companies to be included into indices (âfast entryâ rule), I have a feeling that OpenAI and SpaceX and Anthropic are mostly looking to dump their inflated shares into the publicâs retirement accounts by force.
Michael Burry called out this structural manipulation play recently:
https://www.benzinga.com/markets/tech/26/03/51248353/michael...
Retirement accounts aren't required to buy a stick just because it's listed.
... probably will though
Retirement accounts already own funds and those in turn are often tied to the underlying index. If the time to being included in an index is reduced, they end up being automatically bought sooner. And that keeps their price from collapsing artificially.
Probably again cancelled or significantly reduced in the near future when all the current investors inevitably cash out.
Ok, let's switch to the HTML doc title above.
> No, they didn't raise $122B as the HN title implies.
What is this about? The title says "OpenAI closes funding round at an $852B valuation". It does not mention $122B.
Edit: the linked page's title is different and indeed states $122B.
Seems like all of OpenAI's "deals" are announcement fodder with no real contract, primed to quietly fall through later.
"Here comes another bubble..."
The title is incorrect. The $122B includes previous promises. They raised an additional $12B of promises:
"The round totaled $122 billion of committed capital, up from the $110 billion figure that the company announced in February. SoftBank co-led the round alongside other investors, including Andreessen Horowitz and D. E. Shaw Ventures, OpenAI said."
This IPO, if anyone underwrites it, is going to fleece retail so hard. Better make it a SPAC with the help of Chamath and Cantor & Fitzgerald.
> This is not just product simplification. It is a distribution and deployment strategy.
iykyk
Are you suggesting this was written by AI?
You are absolutely right.
It's a very frequently used structure by LLMs especially ones writing for LinkedIn.
"It's not X, it's Y."
A linguistic presentation commonly referred to as constrastive negation.
Humans may use commas here, but LLMs always use a full stop, always.
Was looking for a precise term for that. Thank you!
Also AI-Linkedin-Bullshit likes to use "just" additionally and it's mostly along the lines of Y being something much more impactful then X.
Itâs not just a suggestion. Itâs a demonstration.
And you are so bold for identifying it. Youâre not just passively commenting. Youâre creating something bigger, larger than yourself.
It's the demonstration layer
agentic demonstration layer
Why would they not use their own AI?
Funny how quickly they have become like every other tech company. There is basically no hint of OpenAI the non-profit anymore.
Edit: Why did this go from their press release to a news story?
Sir, I'm sorry to be the one to tell you this. But you've been in a coma since 2022 after a severe car accident.
It's now the year 2026. That dead horse has already been beaten.
They were trying to keep the facade up until they were allowed to become a public benefit corporation. At least that's the way it seemed to me. Now they are fully mask off.
I'm old enough to remember when companies worth $1 billion were called "unicorns." Now we have a company raising 122 times that? Valued at nearly 1000 times that...?
At least they're throwing consumers a bone via the ARK deal. It's crazy how little AI exposure is available to anyone who isn't already wealthy and/or connected.
I think this is reality-distortion field rivaling that of Jobs', and a crisis of faith. Nobody apparently believes that capital is worth investing into anything but AI.
> Nobody apparently believes that capital is worth investing into anything but AI.
This is the main reason we see this insane investment into AI imo. If you imagine having lots of money, where should you invest that currently?
Housing market: Seems very overvalued (at least in germany). Also with the current uncertainty and inflation its hard to make an investment that pays back over 20-30 years. So building is also difficult.
Stocks are very volatile currently. Not only since Iran. To me it seems since the financial crisis 2008 investors don't enjoy stocks as before.
Gold: Only if you are paranoid about collapse of society. It doesn't make sense to invest into s.th. without interest rates.
Crypto: Same as gold, but better if you like gamling. I would assume most people who are very rich don't gamble with most of their fortune.
Looking around, and especially forward, it would be military tech, e.g. [1], and its supply chain, e.g. [2] :-\ Valuations are not as crazy, but I bet there'll going to be a lot of demand in the coming decade, unfortunately.
Chip production, too, of course, but it's overflowing with money already, apparently. It's growing though, because there are real actual shortages of stuff like RAM and SSDs, there's money to be made immediately if you can. Chinese RAM manufacturers are building out like crazy.
[1]: https://www.ultimamarkets.com/academy/anduril-stock-price-ho...
[2]: https://www.marketscreener.com/quote/stock/RHEINMETALL-AG-43...
> Looking around, and especially forward, it would be military tech, e.g. [1], and its supply chain, e.g. [2]
Only viable if youâre okay with the ethical implications of funding war.
Would you be fine with the ethical implications of funding the industry to fight WWII? Would you consider funding Ukrainian military unethical? Or Taiwanese?
This is, sadly, not theoretical, and I'm afraid we'll soon see more of such choices, not fewer.
> it would be military tech
Anduril is the only company in this sector in the US that has any promise and they aren't even public. Most of us are not going to get our hands on this.
Traditional defense sector looks more like Jeep, or Kodak...
Anduril has yet to deliver anything of consequence. I hope they shake up the industry but to say they are the next hot thing and write off the primes at this stage is premature.
Invest in the Ukrainian drone producers which proved themselves on the literal battlefield! Some of the Gulf states already did.
The demand for more Patriot missiles is large, now that much of the stockpiles have suddenly been spent. Raytheon should do fine just based on that.
this admin will probably source patriots made in china
> Stocks are very volatile currently. Not only since Iran. To me it seems since the financial crisis 2008 investors don't enjoy stocks as before.
These returns do not qualify as âenjoying stocksâ?
https://investor.vanguard.com/investment-products/etfs/profi...
The returns are higher than before 2008, the previous 15 years are unprecedented.
https://www.macrotrends.net/2526/sp-500-historical-annual-re...
> To me it seems since the financial crisis 2008 investors don't enjoy stocks as before
Maybe in Europe. The US stock market has nearly tripled since then. Literally the best period of stock growth in history.
"The Roaring Twenties roared loudest and longest on the New York Stock Exchange. Share prices rose to unprecedented heights. The Dow Jones Industrial Average increased six-fold from sixty-three in August 1921 to 381 in September 1929. After prices peaked, economist Irving Fisher proclaimed, "stock prices have reached 'what looks like a permanently high plateau.'"
https://www.federalreservehistory.org/essays/stock-market-cr...
You can argue that current market multiples are higher than 1929 [1] - and they're certainly high - but this also ignores the mechanism that drove that crash, focusing only on the symptoms. We simply aren't doing the kind of consumer margin buying that drove the '29 crash. It isn't even close. Average schlubs were leveraged to the stratosphere to buy shares of boring industrial stocks.
[1] https://www.multpl.com/shiller-pe
> The US stock market has nearly tripled since then. Literally the best period of stock growth in history.
The only thing I meant to point out was that a very high stock price by itself is no guarantee that there isn't a crisis around the corner. We plugged a lot of holes after 2008 and then reversed a lot of those fixes, I hear retail investors talking about their stocks at birthday parties again. Deja vu... of course this time it will be different. Or not. Let's just say that with the proverbial bull in the earthenware goods store on the loose if we only end up with another financial crisis that might actually not be so bad.
Ok second best :-) I wasn't alive in the 1920s though
True, but it is close enough in time that we should heed the lessons learned lest we repeat the experience.
Do you know the actual lessons of that crash? Because we don't allow retail investors to go 10:1 on leverage anymore. There are a lot more lessons and none of them apply to this situation (even Glass-Steagall). This is much closer to the dot com crash in 2001 in how it looks, just a lot more concentrated and probably a bit bigger. If all you got is "number go up too much" then you probably shouldn't be investing your own money.
The good news is that its almost all rich folks money on the line here and a small amount of dumb money. That's very different than, 2008 where it was mostly the indexes that got hit and that's more middle class/upper middle class concentrated.
you gotta have some of all the above actually.
OpenAI is making $24b a year. It's a 32x revenue multiple. High, but not insane. Spinning this as a story of overinvestment doesn't make sense.
Are you conflicting price to earnings to price to revenue?
I wonder what is not getting invested in bc AI has been crowding out everything else since 22.
It has to be brutal out there for everybody else, if all the money is going to AI.
But they're really cagey about actually handing money over to them today
It's the result of too much echo chambered bullshit floating around daily about how capable LLMs really are. It's literally crypto/blockchain all over again. It's one big lie that a lot of people have bought into which causes it to self-perpetuate, like religion.
That's the tao of hyper-financialization. It must keep growing irrational exuberance big and up forever like stonks or it bursts like DotCom and tulip mania. It's funny money that cannot be liquidated for real value for more than a tiny fraction of the imaginary trillions being thrown around. Similarly, Nvidia $4T mkt cap makes absolutely no sense when it has but a few incestuous customers-parters-investors throwing around tens of billions each per year devoid of fundamentals like essential service offerings that turn a profit. Those handful of whale customers will make their own chips or cease buying large qtys at any time.
https://www.ark-funds.com/funds/arkvx
The fund is invested in most of the hot tech companies.
ARK was all the rage around early pandemic time when wallstreetbets was in the news a lot. Most people probably know it from then.
An ARK ETF is a smell to me. Besides, based on their holdings, i would never invest. 18% of the fund is SpaceX
Even a billion dollars is crazy money. If you have a company with a subscription service that costs $100 yearly, you have ~2m customers, with a 50% profit margin. Your company makes ~100m every year in profit. Imo that's what is actually worth a billion dollars, maybe even a bit less.
That company is probably worth about $8b, FYI. Obviously that's an estimated average but a P/E ratio of 80 give you that valuation.
I would not call an effective 2.9% expense ratio "throwing a bone".
Also, the valuation for such a debt laden company should be viewed with great skepticism. I'm afraid a lot of mutual funds will end up holding the bags.
It's not that far off from the standard 2% mgmt fee and 20% of excess performance?
VCX (Fundrise) has way more exposure than ARKVX
It's also trading at a huge premium. Probably worth a read if you're considering it: https://www.morningstar.com/funds/fundrise-innovation-is-not...
The money is worth much much less than it was before, we live in times of global hyper inflation.
> At least they're throwing consumers a bone via the ARK deal. It's crazy how little AI exposure is available to anyone who isn't already wealthy and/or connected.
It is deliberate. Period.
It's always been known that you make money in the private markets and pre-IPO companies and retail is the final exit for insiders and early investors.
Retail is not allowed to be early into these companies (Because that would ruin the point of being an insider) and this "exposure" has to be at the near top.
There are ways now for retail to get in to these companies including, check out hiive or equityzen...just beware of massive dilution.
Who are "these" companies? Did retail get into Google, Facebook, Amazon, Tesla, etc before the top?
Also, aren't AI businesses losing a lot of money each year? Pretty sure there is some risk involved that is not good for retail.
This has to be just an extension of their previous raise, right? This was a month ago: https://openai.com/index/scaling-ai-for-everyone/
Doesn't look like it, that previous round was entirely Softbank + Nvidia + Amazon, while this one is VC + private investors.
The valuation seems odd though, you'd expect $840B post-money from that earlier round?
Maybe? Previous valuation is $730B + $122B raised in this announcement = $852B valuation in this announcement (no actual increase in valuation)
Previous was $730B pre money. This one is $852B post money. So yeah it's the same one. Good catch.
yup and begging for retailers money.
I can't help but think building an "everything" app is so.. both unbelievably ambitious, and a folly. I am not personally convinced that people want all the things that this super app purports to do.
I am from a generation that still sits behind a desktop computer when making "big purchases." I can't even buy a flight on my phone. I am so much less likely to want to have an AI agent do that for me.
Then the idea that daily consumption of these products will drive people to use them more at work... I have a very different life outside of work. My use of AI outside of work is exceedingly different to what I use it for at work.
I sometimes feel wildly out of touch. But sometimes I view this as the VR moment. To me there are some things that I think may always be preferable to do outside of that ecosystem. And for me, a lot of tasks that 'agents' enable are small enough or important enough that I want to do them myself.
I don't think I'll ever be comfortable allowing an agent to call me a taxi, or order food on my behalf. Because the convenience of asking for food isn't worth the chance it'll mess up, and opening an app and looking at a menu is simpler.
I also think we're coming to a moment where we can start identifying the markers of AI generated content on sight. And I think there's a growing animosity to it. I might be comfortable asking AI something, but when I am looking for or searching for other content, seeing AI content markers make me angry at this point.
To finish, I do just sort of straight up hate the idea that we're comparing this moment to the invention of electricity. It's on the face of it absurd.
I've worked for 3 different startups where the CEO at one point gave us the talk of "we're building a super app".
Admittedly openAI is in a better position to do it, but not by much.
Everyone wants to be WeChat in china. No user wants that from them.
WeChat is not a super app. It's a browser. Tencent WeChat is the equivalent of Google Chrome.
Is it a browser or like a browser? I've never actually used it but from what I understand WeChat's mini programs are like web apps but not something you can open up in a typical browser.
Alternatively, you could say browsers are the original super app.
its a super app.
I think the core issue isnât what underlying technology is used, but rather the service providers. They package their services into mobile apps or WeChat mini programs, and restrict functionality on browsers. For many ordinary people, this provides convenience, but for those who care more about privacy, itâs quite problematic.
WeChat in China covers almost every aspect of life. Even someone like me, who doesnât want to use it often, canât avoid it. Some restaurantsâ online ordering systems only support scanning via WeChatâthat is, WeChat mini programs. People can pay utility bills, call taxis, shop, and make financial investments all within WeChat. Alipay offers similar functions as well.
WeChat is also one of the largest content platforms in China, similar to Medium. Countless creators set up subscription accounts on WeChat and gain more users through readersâ sharing and reposting.At the same time, government information is often released through the WeChat platform.
cries in Musk
>> To finish, I do just sort of straight up hate the idea that we're comparing this moment to the invention of electricity. It's on the face of it absurd.
Do you feel that any technology is comparable in itâs impact?
Most of modern medicine, by which I mean each discovery and invention in their own right, stand alongside electricity. Particularly vaccines.
AI isnât there yet. You could turn off AI tomorrow and thereâd be a shock but people would quickly switch back. You could not do the same for electricity, medicine, combustion engines (or steam engines/turbines), computers, the internet, modern building materials, etc. You try to swap back off any of those and the modern world (literally and figuratively) collapses. Turn off AI, and thereâd be a financial collapse but afterwards everything would return relatively easily to an earlier way of doing things (ye know, the way from just 4 years ago, and which is still 99% of how people do things :) )
Sure, but compare this to "turn[ing] off" combustion engines a mere four years after commercial adoption rather than 162 years later (now). Back then, going back to horses wouldn't have been as big of a deal as it would be now.
I think the Internet is the more apt analogy. But even with electricity, you could have taken it away within the first couple decades of its popularity and society would have shrugged it off. Once they got used to that telegraph thing, not so much.
Yeah, I agree, but AI isnât there yet. Itâs too early to call it one way or the other. Thereâs plenty else thatâs as important as electricity in my view, and maybe AI will join those ranks in 15 years or so when itâs gone through the hype loop and when the economy has recovered from the now-basically-inevitable AI- and war-fueled turmoil of the next decade.
That's primarily a function of the time for adoption, though, not the utility of the technology. In 20 years, people would not be able to so easily say that they could turn off AI with no impact.
That..what..no. The question was whether there are any comparable to electricity, of which I have put forth a number of examples. And also offered my opinion that it is too early to judge whether AI will be as significant or not.
There are loads of technologies that, despite being decades old, do not qualify. So, no, itâs not âprimarily a function of timeâ. It absolutely is about the utility. We can only be in a position to judge utility when sufficient time has passed, and AI ainât had enough time yet to prove its utility. Given enough time, it might prove as useful as electricity, or it might just sit alongside computer operating systems - never quite making it onto anyoneâs âthis changed the worldâ list, even if it has as much utility as an OS.
I think you lack imagination. This is going to be the future because it is legitimately a step up from the previous ways of doing things. I can do things that were way more difficult before.
It doesn't have to be AI all the way - no one's asking AI to book things on its own and make the payments on their own. What does work is, make AI do the research and you verify and you do the payment. Human in the loop.
To me this is clearly the future - AI has access to all the data sources and can translate your intent by accessing these tools in a loop and use intelligence to automate things.
Maybe there's a scenario where that is useful. But again, I don't know why I'd want an AI to do this research for me. I hop on Skyscanner. I type my location, and where I'd like to go. It presents me with a list of options, and I can then use the filters to find times that work best for me.
I see a flight that isn't in my time frame, but is actually like 400 euros cheaper. And I decide in that moment that waking up at 5am is worth the savings.
I'd have not typed that into a prompt. I made that decision at the moment I saw the possibility. I didn't even know that it was an option prior to that moment.
Then I go look at hotels. I have a list of requirements, but I see that one of the hotels that I just glanced at has a really nice long pool, and the amenities look nicer from the images. I change my mind at that exact moment, I can walk 15 minutes more to the beach.
Now it should be even clearer why this is important for food.
You had me going until you said this is even clearer why it is important for food. Food is cheaper and has less impact on your life. I'm much more tolerant of a mistake or suboptimal experience with food.
Yeah, sure, but you also have many more options. More intolerances. And opening a menu to look at some food takes a minute.
Think of personal assistants that rich people have. They learn your habits and take care of business, like making your travel plans. The promise is to give you that for much less.
I get the appetite for frontier models. But why not just invest in Google. Do they really expect the return profile of OpenAI to be vastly different than Googleâs Gemini?
it's not really investing though. Amazon will provide 50B worth of compute and Nvidia will provide 30B worth of chips etc. Google doesnt need any of those.
The only one that is really investing is SoftBank who is pushing for a faster IPO so they hope to make a profit on that and again Google does not offer that opportunity
I imagine they get a bigger slice of the pie with OpenAI than they do with Google, an extremely mature company whoâs had investors buying in for 30 years.
Alphabetâs market cap is $3.5 trillion, compared to OpenAIâs $850 billion reported here.
Google has little need for more money, so the price will be much higher. OpenAI being a separate entity also means Google's competitors (Microsoft, Amazon) can invest there without looking silly.
This leak and looking into source code gave me an impulse to try OpenCode with codex models. I am very impressed with how well it works, and the UI is beautiful.
Amazing. Reporting two mercurial numbers as if they were cornerstones.
Ugh, does this mean I should say goodbye to what little RAM and other hardware remains available on the consumer market?
With $122B what are you going to build next? Spaceships?
Rather than advancing the state of the art, they'll use it to slow down competition by starving them of resources. In the style of a monopoly.
Google makes money like that in their sleep from mostly just advertising.
OpenAI charges about 20$/month to tens of millions of users right now. 240/year. About 12 billion per year. That's a market that could grown to billions of people with a long tail probably not paying a lot but being served ads; and a fat high end paying a lot more than that for an ad free/premium experience. They should be able to reach billions of users.
It's why Google is matching investments in AI like this entirely from the profit from ads.
Going to space isn't that expensive. SpaceX bootstrapped with a lot less than that.
All the high performance RAM, frontier node wafer capacity, flash and disk drives on Earth. Also, all the gas turbines.
If they buy all the jet engines too (to turn them into gas turbines) they can be carbon neutral. More gas burned, but less jet fuel. It's a win-win.
Universal paperclips. And here I am asking ChatGPT which type of onion to use dor coq au vin.
Nah, you'll get a few days worth of inference.
This all smells fishy. They didnât âraiseâ $122B. Raise means someone put funds in your bank account and said send us the next quarterly report to tell us how our investment is doing.
They have pieces from paper of folks saying they may put up funds or goods and services in that amount. But itâs important to remember that:
1. While they are âraisingâ commitments others are backing out of deals (see Disney, various data center things). Big deals announced to major fanfare are falling through.
2. They slashed capital expenditure for the future after previously boasting about all the commitments. This is turning into bonkers math of X + Y - X + Z + W - 1/2 of Y = ? On trying to keep track of whatâs actually âraised / realâ vs what was PR puffery that folks ran away from later.
3. Circular financing still seems to be going on. Big difference of hereâs cash, have fun and various âcommitmentsâ and balance sheet games that seem to still be going on.
Net net this all still looks very scary and iffy at best.
If OpenAI goes down their investors will lose any chance at getting their money back. They need to keep pretending things are going great for as long as possible.
Are we truly arguing semantics on HN which is a news aggregator for startups and everyone truly knows what a "raise" is and it is obviously not funds in your bank account? I don't disagree with the rest of your comment and the core thesis is valid that OpenAI is very much doing circular financing.
Edit: A raise comes with stipulations on what you can use the money for. I don't know if I was being too mean about responding to a parent but before you comment just google what a raise has..
Other than funds in a bank account I do not know what it would mean.
Stipulations on what you can use the funds for.
*hypothetical future funds
So when my coworker tells me he got a "raise", they're not talking about money that will end up in their bank account?
It's a different definition of the word for one thing, and anyway, unless their compensation is prepaid this would only suggest that "raise" doesn't mean liquid money in an account, because an employee's raise is a promise to pay an amount over the remainder of the year with the stipulation the employee continues at the job.
Right because, what does even "buy" mean?
https://thedeepdive.ca/openai-locked-up-40-of-global-ram-wit...
I hate to read this line when academics and graduate students who work in basic and hard sciences have their funding cut. The grand funding that pays minimum wage to grad students is a burden for this society, yet for a company that took all the valuable data from sources that never got credit, raises billions of dollars. Open says the name, but closed it is by operation. Sorry for this rant, but the priorities of this world suck.
Weâre all feeling it. Unfortunately our country is run by a dude with dementia and a carnival of Fox News hosts.
But I suspect thatâll u-turn hard when the economy implodes
Or all of the people that they didnât ask, let alone compensate, that made all of the stuff they munged up for training data, so they could sell cheap knockoffs in the same markets.
Fortunately, OpenAI's naming didn't cause a wave of greedy and predatory new companies, like OpenHealth, OpenEnergy, OpenCompute. The non-sexy name of their product using the originating algorithm (GPT) is punishing them. It could be anything else more attractive that would bring more customers. Because, for me, a good, inspiratory, engaging name is half the success.
https://www.opencompute.org
Are there any Polymarket / Kalshi bets on the over-under for the price? I wonder when the music will stop.
https://polymarket.com/event/openai-ipo-closing-market-cap-a...
I don't gamble but if I did, I'd bet a lot on $1.6t.
Microsoft will buy OpenAI for $500B, rebrand it as Microsoft AI.
This is where they are headed. But the real price will be a lot less.
What if all the people currently using these "AI" services are the entire market for those services? I'm pretty sure everyone that wants to use LLMs is already doing so and already paying for the service.
That would mean the only way to increase growth would be to charge more per token and to get the existing people to use more tokens. Both of which seem to be only what mature companies do when trying to squeeze the cash cow for all it's worth.
It also explains why they're trying to stuff AI into everything, to keep the numbers up, and to get everyone to try and pay them money.
When I show people personal projects Iâve vibe-coded with Claude Code, they often seem impressed and envious. They come up with ideas for things that they would like to do, too. But they have full-time jobs outside of IT, and when I mention they might need to use the terminal to do what Iâm doing now their eyes glaze over.
A couple of such people, after they learned about Claude Cowork, signed up for Anthropic subscriptions and are now using it in their jobs. But overall my impression is that there is still huge potential demand from regular people who use computers for agentic systems with less barrier to entry, and that many will be willing to pay for such systems when more mature and user-friendly ones arrive.
Most of humanity hasn't figured out they need to adapt yet. It's a bit like email and the internet in the mid nineties. People had heard about it but hadn't really embraced it yet. Five years later most people with white collar jobs had email addresses. Fifteen years later, billions of internet capable smartphones were in circulation.
The AI revolution is following a similar adoption curve. Right now many of the tools are only really usable if you are a developer or at least not too shy making AI agents use developer tools on your behalf. That's not going to stay like that for very long. It's going to be a messy transition that will likely take much longer than some people seem to think. But eventually most people doing knowledge work will be leaning heavily on all sorts of AI agents to do their thing. And quite a few will have to learn new skills as most of the stuff they still do manually today just goes away as a thing that you do manually.
Like the mid nineties, these are amazing times for people with a slight head start over everybody else. Which is why there is such an investment frenzy around AI right now. Lots of possibilities where lots of money might be made. And lots of things that won't work out. And lots of people really not seeing the forest for the trees as well. And generally behaving like headless chickens. But the internet in the end proved to be not a fad and it didn't all go back to normal after the hype died and the .com bubble burst.
IMHO, the bubble around AI is not so much the technology but things like data center and energy pricing. The cost of data center production is long term a fraction of current cost (dominated by GPUs costing tens of thousands of dollars). Likewise cheap and plentiful energy to power them is going to eventually cost a lot less. Short term scarcity eats up a lot of billions right now. But you'd be mistaken to confuse that for long term structural cost. Cost is going to come down and that will drive adoption. And that's before you consider edge compute on commodity phones and laptops. There will be billions of devices running small AI agents. Add robotics to the mix and it's a whole new world.
In short, companies like OpenAI and Anthropic are valued so high because all of that is happening right now. Yes, it's a bit of a bubble. But stuff will definitely happen.
On the other hand, the productivity gains from AI automation are so large that you are forced to use it to compete in the workplace, even if you strongly dislike the terminal, you will dislike homelessness more.
I think there is a massive market in waiting for something that uses a way better UI/UX. Iâd say chat is only great for developers/technical people.
Give it a year or two, and apple or any other hardware will have unified memory OR AMD will have a good offering to run all that stuff locally. It won't be as good as Claude, but it'll do for 90% of the things. It will be expensive as first, just like the first mainframes, then give it another 5 years or so, and it'll be affordable.
I'd argue this has already happened. The highest end mac pro is $10k and down from there. We might be returning to the appliance business model.
Its just a matter of Productizing the software to plug and play light office admin work.
Think about all the "people" AI services can displace in due time. There's a fuckload of pencil pushers / knowledge workers with 100k student loans whose lifetime contribution can probably be measured in a few hundred dollars in tokens. And TBH normalizing AI crutch for kids is going to make large % of future cohorts even more replaceable. Skill atrophy among youth is declining hard, but AI is basically crippling future workforce quality to make their displacement even easier. There's even less reason to hire entry level in 4 years not just because models get better but human capita is going to be so much worse.
> only way to increase growth would be to charge more per token
Or spend less per token (ie. increase profit margin without raising prices).
The market hasnât been built out yet. Thereâs that post from a couple days ago where someone frontloaded the entire UX of an operating system onto an LLM, so you just tell the hardware what you want to do and it does it. https://news.ycombinator.com/item?id=47557165
The growth is there but itâs going to be a marathon, not a sprint. I donât know why everyoneâs in such a goddamn hurry all the time
The only thing that's really accelerating is how fast you get rate-limited on ChatGPT.
The big AI firms are all heavily compute-constrained, so that shouldn't be much of a surprise.
feels like an insult to readers to try to pretend that their revenue per month is comparable to google or apples growth when the funding is absurdly different, not to mention inflation itself.
I am very much onboard with AI within my workflow. I just don't really see a future where openai/anthropic are the absolute front runners for devs though. Maybe OpenAI does just have the better vision by targeting the general public instead, and just competing to become the next google before google can just stay google?
What is their next step to ensure local models never overtake them? If i could use opus 4.6 as a local model isntead and wrap it in someone else's cli tool, i 100% do it today. are the future model's gonna be so far beyond in capability that this sounds foolish? the top models are more than enough to keep up with my own features before i can think of more... so how do they stretch further than that?
A side note i keep thinking about, how impossible is a world where open source base models are collectively trained similar to a proof of work style pool, and then smaller companies simply spin off their own finishing touches or whatever based on that base model? am i thinking of thinks too simplistically? is this not a possibility?
Anthropic is definitely gaining ground over OpenAI in the business world. Cowork is the absolute hotness right now, and even prompted MSFT to drop their own variant yesterday
Ask anybody you know that works in Big Tech. They're all pushing hard for Claude Code adoption.
Codex and Gemini CLI seem 1-2 months behind Claude Code. They will catch up. This race will eventually be won by whoever can come up with the cheapest compute.
And that's a dangerous game because the cheaper compute gets, the more likely consumers are to self-host rather than pay a subscription.
Apple could figure out a way to neatly package it into their ecosystem.
Not really. Most people won't self host.
The general public will self-host it's built in to your next phone or laptop straight out of the box or maybe from the App Store.
I agree that that's what it would take, but compute would need to get very cheap for it to be feasible to keep models running locally. That's an awful lot of memory to have just sitting with the model running in it.
True. I was thinking more of power users. Do you think Opus level capabilities will run on your average laptop in a year? I think that's pretty far away if ever.
You can demonstrate "running" the latest open Kimi or GLM model on a top-of-the-line laptop at very low throughput (Kimi at 2 tok/s, which is slow when you account for thinking time) today, courtesy of Flash-MoE with SSD weights offload. That's not Opus-like, it's not an "average" laptop and it's not really usable for non-niche purposes due to the low throughput. But it's impressive in a way, and it does give a nice idea of what might be feasible down the line.
> how impossible is a world where open source base models are collectively trained similar to a proof of work style pool
Current multi-GPU training setups assume much higher bandwidth (and lower latency) between the GPUs than you can get with an internet connection. Even cross-datacenter training isn't really practical.
LLM training isn't embarrassingly parallel, not like crypto mining is for example. It's not like you can just add more nodes to the mix and magically get speedups. You can get a lot out of parallelism, certainly, but it's not as straightforward and requires work to fully utilize.
It's hard to train models in the open. All the big players are using lots of "dodgy" training data. Like books, video, code, destinations. If you did that in the open, the lawyers would shut you down.
Though I think these companies are wildly overvalued, I don't see LLMs as a service going away in the future. The value in OpenAI is that it provides extra compute, data access, etc. My money is on local AI becoming more of a thing, while services like OpenAI still exist for local AIs to consult with. If a local model can somehow know that it's out of it's depth on a question/prompt, it can ask an OpenAI model if it's available, but otherwise still work locally if OpenAI fails to respond or goes out of business. To me that makes a lot more sense than the future being either-or.
Models not being able to reliably know if they are out of their depth is a foundational limitation of the currently generation of models, though.
Best they can do is to somewhat reliably react to objective signals that they've failed at something (like test failures).
> What is their next step to ensure local models never overtake them?
As someone who experiments with local models a lot, I donât see this as a threat. Running LLMs on big server hardware will always be faster and higher quality than what we can fit on our laptops.
Even in the future when there are open weight models that I can run on my laptop that match todayâs Opus, I would still be using a hosted variant for most work because it will be faster, higher quality, and not make my laptop or GPU turn into a furnace every time I run a query.
If your laptop overheats when you push your GPU, you can buy purpose-built "gaming" laptops that are at least nominally intended to sustain those workloads with much better cooling. Of course, running your inference on a homelab platform deployed for that purpose, without the thermal constraints of a laptop, is also possible.
I didn't say it overheats. It gets hot and the fans blow, neither of which are enjoyable.
MacBook Pro laptops are preferred over "gaming" laptops for LLM use because they have large unified memory with high bandwidth. No gaming laptop can give you as much high-bandwidth LLM memory as a MacBook Pro or an AMD Strix Halo integrated system. The discrete gaming GPUs are optimized for gaming with relatively smaller VRAM.
You can host a website on any rackmount server for pennies compared to AWS. But people still use AWS.
The market for local models is always gonna be a small niche, primarily for the paranoid.
"The market for local models is always gonna be a small niche, primarily for the paranoid."
Have you ever heard of industrial espionage? Pr privacy regulations? Or military applications?
(Also the US military runs claude as a local model)
>"But people still use AWS"
I do not, I self host. My current client is also got rid from AWS packing up nice savings as a result
Standby for tiered information.
Have a question on health symptoms? You must subscribe to openAI health
Have a question on math? You must subscribe to openAI math
The only way they can manage this type of monetization is a lot of compute to process outputs.
Each passing day the lie that markets efficiently allocate capital is further exposed
I thought they needed $7 trillion or they'd be unable to keep training new models?
Intuitively, it just doesnât feel like OpenAI has a trillion dollar moat.
Alphabet is worth $3.5 Trillion. Whatâs their moat?
Looking forward to the movie about this absolute scammer Sam Altman when this is all said and done.
Does anyone know if this, like the spacex/xAI stock â will list on nasdaq? And will it be part of every market cap weighted index fund like VOO(S&P) etc or just for nasdax-100 etfs?
The exchanges are bending head over heels to accommodate these IPOs[1] and make our retirement index funds the exit-liquidity strategy to the thievery of pump and dump actors that buy it low and then sell high? As i understand the way thievery works is:
1. List at many multiples of market valuation on an exchange. So if you company is just 10 billion$ nasdaq and theives collude and say "can make it 100 billion..".
2. Lots of institutional investors and rich billionaires get stock options.
3. All market weighted index funds â aka all *your* low expense ratio ETF money â have to re-balance and buy them, raising their value: the exit-liquidity event
4. Rich A**** get richer by making an profit by selling higher.
[1] https://www.economist.com/leaders/2026/03/31/index-providers...
It does seem likely. Iirc SpaceX wants to waive the year long holding period before their stock enters the indexes.
Which to me seems like a very naked attempt at getting 401k to bag hold
They are pushing hard to be part of the nasdaq 100. You can shift your retirement funds from the nasdaq to the Dow or s&p
With $122 billion at your disposal, what do you plan to build nextâspaceships?
isn't it weird that there is no attribution to a human here? i mean, eventually, they have to dropkick sama and install GPT itself as king, right? EOQ seems as good a time as any
the only lesson the common man should take from these valuations: start to protest against AI conpanies being included in the sp500!
unless you're a private investors in these preIPO, the whole plan is to get big enough, get forced entry into indexes, and leave early with everyone else holding the bag.
I do not understand how these companies can have such high valuations when minimax m2.7 is going to be open weight.
Pardon my ignorance, but how can a company that spends almost $2 for every $1 it brings in be worth anything?
What will happen first? The Singularity arrives, and hyper-intelligent AI causes such rapid technological change that the world becomes unrecognizable overnight?
Or OpenAI pays off it's investors? Lol.
I am not sure if I believe in the Singularity or not. But it's kind of the best story ever to support the game of musical chairs that is Silicon Valley investing.
Are we going to see a trillion dollar IPO?
Would be nice to have this seperated from, i-owe-you, vs actual funding that's liquid!
Feels like the number is more about FOMO than fundamentals at this point.
There is a lot of talk about the AI bubble. I think there are comparisons to the late 90's/early 00's here with early stars rising quickly but, ultimately, falling. Since essentially everything touches the internet now it is clear that the 'internet bubble' was more of a shakeup of companies than a real over-hype of the internet. That, I think, is at play right now too with the 'AI bubble'. AI isn't going away but some of the early stars may not make it.
So, the real question here is: Is OpenAI Netscape, or are they Google?
What is their moat?
It seems like its just consumer name recognition at the moment
I donât care if OpenAI is Netscape or Google.
What I want to know is are we in 1994 of the AI bubble or 1999?
Because even after the dotcom bubble popped, it was still much better than it was in 1994.
I think the cycle is way faster than it used to be so even if we are in 1994, the pop is probably pretty close.
Given that Nvidia's forward earnings in 2026 is 19, which is lower than the average for S&P500, what do you think a "pop" will do to Nvidia stock?
Wallstreet is already pricing Nvidia to have no growth above the average S&P500 company.
More importantly, do you think that a "pop" means supply of compute is suddenly more than demand?
What the heck are you talking about? Nvidiaâs value of growth assets are already priced in.
Youâve written so many posts with glaring flaws. Are you trying to come across as smart? Lmao
$852B and we still don't have an IAM layer for AI agents.
Almost a trillion for a company that hasnât proved it can reach any form of profitability, all on the promise of an elusive messianic concept of machine superintelligence through probabilistic algorithms. Basically like praying cancer away. Peak magical thinking, peak America.
[flagged]
Please don't post generated text to HN. We ban accounts that keep doing this.
https://news.ycombinator.com/newsguidelines.html#generated
https://news.ycombinator.com/item?id=47340079
Don't know why so many down votes for something that's clearly just another opinion
Probably because it reads a little like an LLM and also has an emdash
Iâve seen a lot of anti AI people use ChatGPT to write why AI bubble is about to pop.
Ironic, isnât it?
well we do need at least two powerful AI companies, so they can cross checkout each other when I use them.
does Softbank run of out money i.e announce these deals but never follow up ?
This is so bad. OpenAI has already singlehandedly destroyed the global flash memory market with their war chest, this will just give them more power.
Google is totally going to run this company out of money aren't they?
This is going to be bad
They are trying very hard to convince themselves that it's going to work, when we see all the models plateauing... it's clearly hitting the ceiling
I don't know anyway using these models everyday who think they are hitting a ceiling.
If anything there's a plateau between each model release.
I'm seeing diminishing returns, though in fairness we have no idea yet how to integrate properly with existing good practices and principles. I suspect improvement is going to come mainly from improved took usage rather than more impressive models.
I feel that too, every technology has its limits. I use AI daily. But I canât see the âintelligenceâ. All I see is fine tuning and bigger datasets.
Yesterday I asked claude to fix the color issues of graph. It failed miserably. Opus 4.6 wasnât able to figure out why the text was grey. It made something up, instead of realizing the problem was simple, oklch wrapped inside a hsl color. hsl(oklch(âŚ)) I easily figured this out by just looking at the css and adding some logs to js.
This is not intelligence. This is a tool thatâs smart. Not sentient. AGI wonât be achieved by scaling alone.
Wow. I doubt Anthropic can raise that. Are they more efficient, can they do with less?
Given how all of Big Tech (except Google obviously) is going all in on Claude Code, I wouldn't be surprised if Anthropic becomes profitable first.
Anthropic doesn't have anything else other than the Claude models.
But notice that no-one, not a single mention of Deepseek tells me that they are preparing to scare everyone again. Which is why Dario continues to scare-monger on local models.
Sometimes you do not need hundreds of billions of dollars for inference when it can be done locally with efficient software; and Google proved that. But where is the money in that? So continues the flawed belief in infinitely buying GPUs to scale which Nvidia needs you to do.
Only a matter of time for local models to reach Opus level. We are 1 or at most 2 years behind that and Anthropic knows that.
> Only a matter of time for local models to reach Opus level. We are 1 or at most 2 years behind that and Anthropic knows that.
Can confirm. Kimi K2.5 is pretty intelligent and most of the time there's no difference between Opus and Kimi.
Local models just make no economic sense since the GPU will idle 99% of the time.
You have a GPU already (at least an iGPU and an NPU on most newer platforms) as part of your computer, might as well get some use out of it with local inference. And trying to do inference on a larger model with an undersized GPU will have you idling a lot less than 99% - but that still makes a lot of sense for most casual users who will only rarely need a genuine "Pro" class answer from AI. Doing that locally is way less hassle than paying for a subscription or messing with API spend.
False on a team thatâs distributed
No mention of "AGI" this time. Since we all knew it was a scam. But this is the most damning of them all:
> The OpenAI flywheel is simple. More compute drives more intelligent models. More intelligent models drive better products. Better products drive faster adoption, more revenue and more cashflow.
FTX had a "flywheel". It fell off. Being saddled with hundreds of billions of debt makes this situation ten times worse.
"Commited capital" - is this the same commited as the $500B for the Stargate project ?
Not gonna lie, I hate those announcements lately. It's full bullshit mode, worse than the Dot-com bubble. Numbers don't make any sense, any more, and yet journalist don't ask any real questions...
thats a really large number
God help us if all of our retirement index funds are forced to buy this bankruptcy bomb.
People won't be able to afford to leave their beds because food prices will be too high to justify the energy cost of standing up.
Every joule of human energy is energy that could have been better spent to produce AI slop for other AI agents to consume.
Nah this raise (commitment) is... blah.
Last announcement I reckon pre-IPO and the inevitable collapse.
Inflation is a hell of a drug.
the hype machine knows no bounds. it should be illegal to publish such farcical claims when they are intended to manipulate markets...
Money has lost all meaning in tech. 122 Billion raise! This is some kind of dream.
didn't they just raise last month?
I think that runway has run out /s
Thsi is the most blatant pump & dump scam I've ever seen. It's going to crash like a meteor.
Good luck competing with Google which has "unlimited" budget.
They have to focus on the distant future (where they are frankly unlikely to exist) because they are falling further and further behind in the immediate future.
Their latest desperate bid for relevance is a plugin for Claude Code that uses Codex as a second opinion. Please clap.
This a big exaggeration. Codex is probably one of the top two LLM programming tools, along with Claude Code. GPT-5.4 models are strong, unlike the initial GPT-5 ones, which were comparatively bad, and can hold up against Opus 4.6. In my experience, they are better at analytical work.
I cannot really see how they are "far behind," or how some plugin for Claude Code is a "last desperate bid." The tools are close enough to each other that I regularly use Codex one month and Claude Code the next without much disruption, just to try out any new models or features that might be available.
I do not have much visibility into the non-code applications, so maybe it is stickier there.
If/when the AI bubble pops and takes OpenAI down with it, I would not expect Anthropic to come out unscathed either.
"Despite unprecedented capital investment in our R&D, our core product isn't getting meaningfully better so now we're building an app."
Doesn't really strike me as the kind of statement that comes out of a company that can sustain a ~$1T market cap...
> The OpenAI flywheel is simple. More compute drives more intelligent models. More intelligent models drive better products. Better products drive faster adoption, more revenue and more cashflow. That gives us the ability to reinvest and deliver intelligence more efficiently to consumers, enterprises, and builders around the world.
-x-
In short, the musical chairs are still playing... Keep on walkin' round, y'all, till the music stops.
/s
Fortunately, I was afraid that this was another bubble. /s
I wonder what the bet is here, long term that valuation is going to have to go up even further for this investment to make sense so they're clearly betting that at IPO time they'll be able to convincingly demonstrate AGI or something extremely close to it. That's a pretty risky bet, and meanwhile, whatever they come up with will be a commodity within a year. And that's besides OpenAI no longer being seen as the dominant player or the player with the best edge.
"This is not just product simplification. It is a distribution and deployment strategy."
I am so sick of AI writing.
The snowclone is annoying, but comparisons are sometimes necessary. The problem here is the actual content is sloppy.
corporate speech existed long before AI
Mmm, itâs almost like OpenAI built statistical models using pre-existing corporate speech as the target data... ;)
Get used to it. All PR statements nowadays get AI treatment before going public.
> We are now generating $2B in revenue per month
What??
They got a very sweet deal from the Pentagon, it seems.
Word on the street is that Anthropic is roughly at half that. Hard to know what they include and not, and what their real, non-subsidized costs are.
What what? Are you surprised it's that low, that high, that they can tell what their revenue is, that they report it on a monthly rather than annual basis, or something totally different?
It's going to be pretty hard to get a good answer to whatever you're having difficulties understanding if you can't be bothered to write more than a word.
> Within a year of launching ChatGPT, we reached $1B in revenue. By the end of 2024 we were generating $1B per quarter. We are now generating $2B in revenue per month.
They raised $122B.
122 / 12*2 = 5 years to get your money back (I simplify, I know revenue <> profit)
They are so big that almost no one can afford to acquire them. It is similar as someone would like to acquire MSFT or AAPL.
WCGW?
Correction: no one can afford to acquire them.