The AI Demand Bubble
wheresyoured.at
wheresyoured.at
The obvious way to recoup spend is to grow, rugpull by cranking up costs 6x and reducing inference cost by half (highly achievable with improved silicon and technology), then simply fire a large percentage of software engineers. From that perspective the current behavior is a bit wicked but downright logical.
And having two whale customers is not really out of the ordinary for any software company... it's just the scale that is staggering.
The actual risk that the author does not even broach upon for investors... the thing that will actually torpedo this massive investment are the open source open weight chinese models that commoditize the entire endeavor. If you don't have a monopoly, you cannot rugpull and 6x the costs on the consumer.
Ironically, the actual thing that will likely kill OpenAI is ACTUAL OPEN AI.
So much this. I've never touched the Chinese models (no particular reason) but it is having an impact as the frontier companies are pushing the price down as a defensive measure. Is this pulling forward what would have happened eventually? No idea.
It is unclear how effective anyone beyond China and Mistral have been at developing cheaper, capable models. It is an expensive business. I'd be curious if anyone had any thoughts on that
I warn you that it is addictive.
I am using DS and MiMo on Pi.dev, and I don't see myself going back to OpenAI. They are criminally cheap in a way that I don't care to spend tokens. That leaves me space to experiment.
And in terms of capability, well... I use Sonnet and Opus at work (provided by my employer), and I see no difference in terms of what I can achieve. Well, besides Claude costing dozens of times more.
OpenAI and Anthropic investors yes, however open weight models are good for cloud providers. They can turn the two large customers into direct ai services that can be spread across many customers and reduce the cloud providers overhead on ai services.
It’s a grift, “AGI will save us” is the same as Musk’s “Mars colony”, it’s not supposed to ever happen, it’s supposed to be a goal post they ever move further
It sounds like an episode of Silicon Valley it really does I get it, you can laugh, it's all right, but it is what I actually believe is going to happen”
Whether or not that’s his current view, who knows, but everyone who’s invested in OpenAI since should have known that this was at least part of his mental model.
It's not really like that, at least the fundamentals, not necessarily what Altman / Musk say.
The significant point is when AI/robots can do what we do without us and improve themselves even if humans disappear. That'll be a new era on Earth.
The monopoly will likely then shift from the model to the compute, i.e. who has the GPUs to serve inference at scale from the open weight models. The cloud compute giants have basically bought everything that Nvidia, Broadcom etc. have to offer. Currently, the inference margins are shared between the cloud giants and OpenAI/Anthropic. But if training great models becomes easier for some reason, the cloud giants benefit. Then they'll have used the OpenAI/Anthropic revenue and spending commitments to grow their cloud business, and then can serve other models and make even more money.
Given that OpenAI and Anthropic are private, I don't think there is any risk to retail investors in this scenario. AI not turning out to be so useful, and OpenAI/Anthropic not being able to pay their bills is the correct failure scenario i think, as identified by the author.
System 2 (thoughtful slow planning) has already been cracked with VLA models. The seemingly easier system 1 (fast reactive) will be figured out in months... maybe with JEPA
It seems like user numbers are stagnating, and the ad play isn't working out so far. Where is the revenue growth coming from? I don't think API can be the answer, because API has absolutely no switching costs.
> The obvious way to recoup spend is to grow, rugpull by cranking up costs 6x
Presumably, that was the plan, but I don't see how that can possibly work with how quickly the Chinese models caught up.
Were that true, many embedded maps would have switched to OSM instead of Google Maps after Google massively increased costs. Adding your business to OSM is free, and API costs would fall pretty dramatically
Not in this specific article but Ed Zitron has been talking about open models quite a lot
I disagree that "the entire endeavor" is inherently doomed for Anthropic and OpenAI. There will always be a very top end of the market running humongous models (like the recently-teased OpenAI Astra and perhaps including future versions of the existing Claude Mythos) that's too large-scale to be successfully commoditized, and that's exactly where the ongoing investments in AI datacenter compute and model training are most likely to pay off at some point. Video generation is another emerging AI area that seems to require large-scale compute, though the value proposition is definitely iffy there.
You don't need Fields math PhDs to make you a web app or plan your vacation.
The remaining frontiers for high value ROI are ultra long horizon tasks, and lower cost faster inference. The former will rely on better architectures which will likely also get commoditized. The latter will probably be led by Chinese companies which already own most of the electronics manufacturing supply chain.
Video generation was already tried by OpenAI and was enormous bust. It busted so hard the current leading proprietary videogen model is Chinese, and the best open weight videogen model is also Chinese.
Does the Apple lawsuit over trade scret theft, including secrets concerning "metal-finishing finishing process", suggest that this "software company" may have plans to sell hardware
This is the thing: let's just reason it out. What if OpenAI and Anthropic both fail and end up in bankruptcy? What is the impact to the hyperscalers?
...they just start selling open weight model inference to businesses, or whatever other entity picks up the scraps from the collapsed labs.
It's abundantly evident that there's major demand for compute, and that it isn't going anywhere.
Zero profit means no recouping investment costs effectively bursting the bubble.
Technological progress can come at extreme pace but the lack of profits can still torch both the startup frontier labs and the hyperscalers.
We plebians can then pick up H100 gpus off Ebay to help us invent new superconductors and generate waifu videos in pur basement...
"fine" and "highly profitable" are doing a whole lot of heavy lifting here, actually masking the quantitative reality that these "high profits" are nowhere near the height required to pay back their debts and remain profitable enough to prevent a sharp drop of their stock prices - that kind of development is the opposite of "fine" as far as investors, ordinary people and the wider economy are conserned.
> Profit margin determines whether OpenAI and Anthropic survive. The hyperscalers would be fine.
That's not how the market works, if OAI or Anthro default, they'll bring down the whole market, it'll be 2008 on steroids. Assuming the fire can be contained to only two trees misses the reality of how dense the forest is and how hot and strong are the winds blowing towards it.
How do you know it's not "as bad"? Whenever things got that bad a crash followed, a lot of people lost their shirts and savings. For them it's that bad and then some. Why mislead these people?
> The obvious way to recoup spend is to grow, rugpull by cranking up costs 6x and reducing inference cost by half... then simply fire software engineers.
OK, I didn't know that brutal/shmutal was the only way to do business these days. If I could only get with the program, I would understand how rosy-smelly the situation is and not at all "as bad as the author makes it out to be".
> The thing that will actually torpedo this massive investment are the open source open weight chinese models that commoditize the entire endeavor.
Oh, wait, I thought it wasn't that bad? Which way is it? Brutal/shmutal failed to work? But, but but, you said, you promised... "not as bad".
Besides, Chinese competition isn't the only way to commoditize AI, new tech developments are the single most significant risk in tech - a well known fact.
> If you don't have a monopoly, you cannot rugpull and 6x the costs on the consumer.
Anyone who invests while convinced that only a monopoly would save his investments has got his brain screwed on backwards,
Whats the common trait between Warren Buffet and Peter Thiel? They only invest in companies that are monopolies or on their way to being monopolies. As Thiel (who very much has his head screwed on backwards) says "Competition is for losers"
What does Rockerfeller, Carnegie, Gates, and Musk have in common? They own monopolies.
So yes! The only way these companies are going to make their investments back is by being monopolies.
Past Performance Is No Guarantee of Future Results
> "Competition is for losers"
That's how we got where we are, losing the global competition game, high polarization, inflation, debt, and wars. Communist China managed to surge ahead mainly due to their purposeful development of a highly competitive industry and market. Oh, the irony.
> The only way these companies are going to make their investments back is by being monopolies.
Well, they aren't and they won't be, the game has changed. Seems like when you say "monopoly" you mean "on the US market" but that's not enough to sustain anything resembling a good life here. You may be thinking of an isolated and self-sufficient national economy but that's a mirage. The world is still global and only a major extinction event, like an all out world war, can change that - do you want to go there?
Buffet, Thiel, Gates, Musk, etc have nice bunkers to hide in... you don't.
He has been predicting a crash for how many years now? And while I can totally see Anthropic and OpenAI going through some things on the way to post-IPO FMV, those things do not include AI going away. It truly doesn't matter whether closed source Frontier lab models are spewing tokens or large foreign open weight models are doing it, the token factories will be just fine, and that's really all I care about.
The question to me is why the media favors influencers like this over practitioners.
And it's not like there aren't more balanced takes out there, here's just one...
https://overweightskepticism.substack.com/p/ais-cash-cushion...
I'm sure there weren't a ton of bankers predicting the mortgage collapse of 2008 but I, a young programmer of mortgage software could see something was weird (but didn't realize that it wasn't the norm).
People were getting multi-million dollar loans with no documentation which had affordable introductory payment and then ballooned to many times the payment.
I had no idea that wasn't normal but I was surprised when I learned that it was possible. I just lacked context and understanding of how the mortgage/banking industry really worked. Had I known as much as I do now after having been adulting for a while and become much more familiar with how banking really works, I could have seen that there was a massive pool of risk that was relying on the value of the housing market to not just keep increasing but increase at an incredible rate. All that it took was for the market to slow down just a little and then all those balloon payments would start defaulting because they couldn't be refinance again.
I couldn't have predicted everything that happened or who exactly would be holding the bag but, with just a couple more pieces of knowledge, I could have easily seen it was unsustainable.
IMO that's anti-AI Psychosis, the evil twin of believing GPT-4o was sentient. And a really bad case of it is labeling anyone who disagrees with you in any way as having AI Psychosis for doing so. And your first line sounds exactly like that to me.
IMO coding agents alone have made AI viable. Healthcare applications have done the same, but coding agents will print money as the cost of tokens drops, and it is dropping. In the meantime, max plans are obviously subsidized, but as long as most of their holders don't token max, they are the netflix of AI until that changes.
https://pricepertoken.com/trends
We needed OpenAI and Anthropic to get us there, but we're there now, and they need to adapt or they will be reduced to glorified neoclouds in the long run. I also predict publicly traded companies that are AI-first with huge PE ratios will go through some things. What I will not do is even try to pin a date on that. The market can stay irrational longer than any of us can stay solvent. But I wouldn't worry so much about companies with high gross margin and PE ratios of 40 or less. Time in the market beats timing the market and all that.
TBF my realtor was telling me about the "Got a pulse? Here's your mortgage!" issue starting in late 2004 after a bizarre conversation with his favorite loan agent after she had done too many tequila shots and started blabbing about basically giving loans to anyone. 100% true story. But again, good luck timing the crash.
> IMO that's anti-AI Psychosis, the evil twin of believing GPT-4o was sentient.
IMO that’s a particularly scraggly straw man. Even Ed Zitron, who we can stipulate is among the most cynical, thinks that some value will be left after the bubble either deflates or bursts.
So really, after all of these companies are wiped off the map because the bubble popped and they went broke, what will be left? Let's get specific here. Let's make some hard falsifiable predictions.
I predict bumpy IPOs for Anthropic and OpenAI, maybe even ending in acquisition instead. I predict anything with a PE over 100 is in trouble. But I also predict anyone with a PE of 40 or less is going to be just fine. Like Michael Burry said, just like Cisco, now running a PE in the high 30s after going through some things. Finally, coding agents are here to stay and they will only get better and cheaper, but they are unlikely to replace common sense meatbags.
So what are your predictions?
Me? I didn't downvote you for it and you are projecting a lot onto me without a sound basis.
Either way, my predictions are all personal, because as someone with his own problems I don't really have the spare energy to give a fuck what happens to the US economy or its tech industry. In practice for the rest of the world, I think it will all be dwarfed by the USA's failings in the Strait of Hormuz.
Though as a man in his fifties who has spent his life in the tech industry, I am slightly invested in the possibility of Larry Ellison's humiliation. Bring that on.
But also, if you don't want to talk about this stuff, why comment about it?
Is it? I don't think this is really true.
For example he's spent some time talking about Oracle's relatively more dangerous exposure to OpenAI. Here he is back in April:
https://shows.acast.com/the-tech-report/episodes/how-openai-...
Here's S&P in July, downgrading Oracle specifically because of their exposure to risks from OpenAI.
https://www.spglobal.com/ratings/en/regulatory/article/-/vie...
> and he's been predicting it for a while now (2024)
Like I said elsewhere, the first accurate, detailed, correct predictions of the subprime crisis were made four years before it finally unfolded. There's likely at least one more round of funding to come for both OpenAI and Anthropic. The big systemic risk to the USA is if the bubble bursts in 2028, if you ask me; that is shaping up to be a restless year.
> but if it's the gospel you want and/or need to hear,
It's not, especially? I listen every now and then; I don't even use a podcast app as a rule, so I'm far from a subscriber or follower. I am more interested in what Cal Newport has to say. I am probably slightly less bearish about AI's long tail value than Zitron is, as it goes.
> But also, if you don't want to talk about this stuff, why comment about it?
I'm happy to talk about it. I just don't care enough to make predictions beyond the personal, because I'm just not that invested and other people are better at it than me.
To the extent that I think personal observations scale up to the rest of the world, I would say that leads me to think that on-device AI will very significantly derail optimistic consumer AI revenue predictions (in particular OpenAI's), that cloud-based agentic coding is closer to its useful limits than people so far understand, that local (on-prem or boutique-hosted if not necessarily on-device) LLMs will do more and more of that work, and that as a result the total addressable market for cloud AI in the next five years is closer to its apex than people in the industry think.
But I am not going to bother to make more specific predictions about the fates of the two big AI companies or the value of the market, because I am not invested in them or the celebrity OpenAI/Anthropic employee influencer aspect of it. I have no team (and I am trying to avoid the products as much as possible).
I have learned in my life not to have too much anxiety about things that won't affect me or I am not close enough to influence.
My limited interest in AI is in the potential offered by smaller models, and I personally think people in the tech industry are thinking in a shallow, FOMO way, obsessing about shiny "frontier" model baubles and what a handful of overpaid loudmouths think, when they should be spending that energy exploring running open weights models and open source tools.
You know what created the AI backlash as well as anyone, and it is: AI and AI people.
If e/acc voices were not so abrasively, obtrusively YOLO about their technology, if their entire take on what they earn millions to do was not so easily reduced to “yeah it sucks that your job will go away, learn AI I guess LOLz” then there would be far less to have a backlash against.
Being lectured about the future by people who do not have a fucking business plan for how they will repay a trillion dollars and who might actually crash the economy does tend to grate on the nerves of the reality-based. Being told again and again that we will be ruled over by two firms that ultimately amount to the corporate equivalent of trust fund kids, that is annoying.
If you want to convince people otherwise, find an analyst who is not churning out AI slop.
As to the “predicting it for years” thing, the first correct-with-specifics predictions of the subprime crisis were published in 2004, by a pretty fringe outlet (karmabanque) and its author, Max Keiser. I remember not being shocked at all when it finally happened, or being shocked at LIBOR rigging. Because Max Keiser presented his reasoning on his crazy radio show, and told his listeners what signs to look out for.
If someone is right for the right, well-informed reasons and presents that reasoning, it doesn’t always matter all that much if their background is unconventional. They tend to be dismissed, and they were back then. “People will always need houses” is what we were told, as if that was enough to ward off massive structural problems.
Another read: Ed provides great content to help people afraid of AI self-soothe and pretend that it's not going anywhere.
He does think the market is going to crash.
Like I say, I find his tone enjoyable, some of his predictions are interesting (and he has already been proved right on its risks to Oracle for example).
I'm not interested in self-soothing and I am not afraid of AI. I am even a bit less bearish than Zitron. I am concerned about a world that is fucking stupid enough to fall for the elements of grift, but I am insulated enough from the consequences, for now, that it's not my primary concern.
If SotA does not go up in price (ala open weights), then everyone will continue to shill digital garbage AND the stock market will presumably get ravaged. So I'm really crossing my fingers about cost increase and- full circle- this hope is why I like to self-soothe with these articles (i.e. I cannot attest the articles' veracity but it makes me feel good that a human writer is speculating convincingly about instability).
That said, I sense that the ravaging is a forgone conclusion, unless someone pulls a(nother) rabbit out of a hat. I'd qualify agentic coding was a rabbit, so it could happen again I suppose.
Ed Zitron helps you see the broad absurdities in this situation and more importantly helps you understand that at various levels there is chicanery: they need more money constantly so they make weird unprovable claims to people who lack the skills to critique them, they exploit not jusT FOMO but actually fear of their own products to market them, their suppliers are making such ruthless use of SPVs as to conjure a possible apparently entirely legal Enron-scale disaster, they are attempting to scare the US government into backing them, and they have tried to get the index funds to back them before they have any hope of profit. None of it is normal; much of it is operating in the face of whatever still passes for principle in the US tech and finance industries. The broad picture is obscured by smoke.
But as fascinating as it is, it's only background information and it only broadly supports one's instincts; anything can still happen. For me, the main instinct it provides some support for is that the market is so terrifyingly lacking in verifiable fundamentals and stable predictable pricing that ultimately people will be driven to open weights and local models to keep control of their businesses; anything else is not a sane bet. So it broadly supports my instincts about my local-first approach, to find sustainable value.
The thing about pulling rabbits from hats is that the rabbit was always there; it's not a new rabbit. With a bit of understanding about how GPTs and LLMs work it is clear that code analysis and generation is not just a good application, it is the utmost practical and profitable one, because nowhere else in broad human effort do we communicate only in context-free grammars. Nothing else is actually going to prove to be more valuable financially).
(Formal pure maths is the "purest" one which is why we see its value in chugging through maths puzzles. But is there as much money to be found there?)
The fund is still well up because approximately 25% of the fund’s assets were invested in Anthropic, which has done well but is not very liquid yet.
No matter how you slice it the whole sequence of events last week was an unmitigated disaster. He’ll likely never manage other people’s money ever again.
> When this star of San Francisco arrived in New York during his fund-raising tour around last summer, however, he received a relatively cool reception, according to three people from whom he tried to raise money, who declined to be identified talking about a private fund. They said they viewed him as a lightweight and a one-hit wonder. The asset management colossus Blackstone, the world’s largest investor in hedge funds, passed on investing, according to two of those people.
> One wealthy New York investor who did take the meeting welcomed Mr. Aschenbrenner into his downtown office, and then gave him a grilling, according to the investor, who declined to be named publicly because he had agreed to keep the contents of the fund’s pitch private.
> What was Mr. Aschenbrenner’s plan if the A.I. revolution didn’t pan out quite as hoped? The hedge-fund founder had no detailed response, the investor recalled. Mr. Aschenbrenner simply truly believed it would all work out.
(as quoted in https://www.bloomberg.com/opinion/newsletters/2026-08-03/hed..., originally from https://www.nytimes.com/2026/07/31/business/situational-awar...)
It actually wasn't an unmitigated disaster. He was overleveraged and apparently had the dumbest hedges (he was long AI and short software and when AI was down software was up so his hedges didn't hedge).
Citadel swooped in and bought the public equity portfolio. That's the definition of mitigated. There was no propagation of the fund's losses to other financial institutions. This made for great headlines but it was a nothingburger as far as markets are concerned.
> He’ll likely never manage other people’s money ever again.
He didn't close the fund so he's still managing other people's money. And while nothing about his profile would ever persuade me to be his LP, you're underestimating the fact that a pristine track record isn't required to raise money. Given this kid's profile and the fact that he made an obvious mistake (even if it was big), there are probably people out there who still have an appetite for parking money with him.
Didn't hold back Sam Altman. And he's not the only one...
You can read in their investor letter exactly what happened and where they stand and explicitly they did not sell every liquid thing nor did they sell all their public equities. Are you just lying or stupid?
2) The bullish AI side is full of grifters and folks that were block-chain and NFT “experts” before they became AI “experts.” 99% of the folks in AI know almost nothing about AI apart from thinking it’s cool and having played around with it a bit.
The folks that correctly call BS on a thing tend to not be deep in the thing. Thats how they see things that are completely obvious to anyone but those so deep in they can’t see what’s right in front of them. That’s playing out big time right now with AI.
The only folks that don’t see a massive AI bubble ready to burst right now are those that have drunk so much Kool-Aide that they long since stopped having any clarity in judgment.
The implosion of “situational awareness” last week due to a complete lack of situational awareness that most Wall St pros called total amateur hour is a textbook case of this unfolding.
At the beginning of this whole thing, when I still had an X account to log in with, I used to scroll back on an AI influencer's profile to see how many tweets I'd have to go past before I saw "ETH" or "NFT".
>They are two-party round-trips: a hyperscaler invests in an AI lab that is also its cloud customer, so the investment comes back as cloud revenue.
They may not label as circular financing but this is still the exact same thing he’s bringing awareness to in his article.
You called him him an influencer, waved at "practitioners" and declared that token factories will be fine because that is your opinion.
If his analysis is wrong, identify the error.
So his point is that big % of cloud revenue of Microsoft/Google/Amazon come from companies that:
1)are very unprofitable
2)need to raise staggering amount of capital to survive
3)are financed by their suppliers and that money is circling back to them
Your counter-argument is this:
>> It truly doesn't matter whether closed source Frontier lab models are spewing tokens or large foreign open weight models are doing it, the token factories will be just fine, and that's really all I care about.
This might be true but there are 2 majors questions here. One is exposure to Anthropic/OpenAI. If they go bust/can't IPO at expected price it's a big loss hyperscalars will need to admit. The second question is how much of that cloud revenue comes from training. This part of the demand is going shrink or disappear in the bad scenario.
His fundamental error I think, illustrated here https://www.youtube.com/watch?v=C0Gcx-6hJJw&t=196s is he thinks AI is just another tech product to hype rather than a comparable revolution to the industrial one.
His criticism to AI has two angles.
The weak angle is on AI usability. I think he is wrong there; AI is clearly useful. Now, there is a discussion if it is multi-trillion dollar useful; I think it isn't, but it is useful nonetheless.
Now, there is a strong angle, which is the economic viability of AI, and the gargantuan amount of money being burned in what is a very risky bet. There, his arguments have proven so far rock solid.
The fact that you (as all his critics) chose to attack only the weak angle says something.
Of course, he may just be deceiving himself or his audience about this belief. I don’t think he is actually dumb. But this alternative is equally problematic.
I think Zitron is wrong in the usefulness of AI; but then again, so are the people hyping AI way beyond its capabilities.
The numbers he brings up in the article are solid though.
That said, if Zitron irks you, you can watch the most recent Patrick Boyle video on the big tech debt: https://youtu.be/NufJ7g63KSY?is=2Eh70U7gleoriLO9
This is a guy that comes purely from the financial angle as has no obvious ideological stance in being either pro or against AI. Also pretty balanced in his delivery.
Some numbers he brings up should make even AI hypers question the sanity of this whole thing.
But why do you think everyone in AI is hyping it beyond its capabilities? That's a specific San Francisco-driven AGI cult and a handful of annoying billionaires. I'm doing what Patrick Boyle said is the biggest use case: running a one person consultancy on $400/month of AI services. It's working far better than I expected. And if the rates go up too much down the road, I'll pivot to running locally.
I suppose if you still consume influencer content, it's pretty bleak slop right now too, but then there are occasional slopcore geniuses that manage something entertaining so I say let the future work itself out. No worries, it will.
As for the sanity of the gold rush phase of anything... Are you kidding me? Really...
The global GDP annually is ~$120T. $2T is not that big a number at that scale. It's interesting that the critics stick to the domestic tech GDP when the global tech GDP is ~$20T to insist AI hitting $2T annually is impossible.
The thing about "hyping AI way beyond its capabilities" is that the future capabilities will no doubt be greater than the current ones so they may just be talking about the future a bit.
There's a bit of a tipping point in usefulness between AI being a bit worse than humans and a bit better, like for mathematical theorems that's probably happened where being not very good was kind of useless and just recently they are proving lots of things. That will probably gradually happen in other fields creating a lot of economic value.
I avoid making predictions on tech:
- AI may speed up on improvements and be a singularity moment, truly a new industrial revolution.
- AI may have only incremental improvements in the next few decades with diminishing returns.
- AI improvements may plateau and fizzle out.
All those are possible scenarios, and if you dig you may find evidence and historical precedence for all of those. I don't think making wild bets that can tank the whole economy based on a FOMO-fueled prediction that everything will work out in a best-case scenario is healthy.
As an example of the predictability, Hand Moravec wrote quite a well argued paper in 1989 predicting human level hardware capabilities would be available in inexpensive machines around the mid 2020s which I think was fairly spot on. He also argued once those capabilities were there, software guys would figure things to use if for.
You can look as past quotes like "that sound you hear is the slow deflation of the bubble I've been warning you about since March" in July 2024 and say he was over pessimistic on the economics but it's hard to prove he still is.
There's a fundamental issue that in standard finance the present value of an investment is the discounted value of the future cash flows which are hard to estimate with a developing tech like AI but Ed I guess thinks they'll be low and optimists high. I'm not sure he'd even be up on the concept of discounted future cash flows.
If you look at the other end of financial knowledge with Warren Buffett, he's bought in, in spite of not being a tech enthusiast, through buying shares in Google. https://finance.yahoo.com/technology/ai/articles/buffett-say...
If you assume that this will be "industrial revolution" situation... maybe? It doesn't help if this happy scenario takes 30 years to materialize. The money needs to be there by sometime in the next few years.
Also, every risky bet has a happy scenario that if successful, all the gamblers become gorillionaires. In the happy scenario, every GME hodler would have wife-changing money by now, with their shares in Gamestop worth infinite dollars.
And yet, his arguments on the economic viability of AI are rock solid. He has many haters, and I am still to see a good counter argument to the numbers he goes through.
In fact, it is a shame that it falls to a "videogame reviewer and PR guy" with no tech or finance background to ask the questions that the press that reports on those companies should be asking.
I'll just pick a nit, on the topic of demand concentration, it cites that "MIT NANDA study / 95% AI pilots failed" number without considering the provenance (and, frankly, accuracy and relevance) of that number: https://archive.md/AvSYL
AI build out is a risk. Is anyone serious saying it isn't? It might blow up, certainly some egregiously overpriced endeavors will revert to mean, but risk taking is what businesses and investors do for a living. But for giggles, I threw Nvidia's latest 10Q into Sol 5.6 to analyze it as Patrick Boyle suggested that was the only way to understand the games they're playing. And its summary:
"financial distress risk is very low; earnings-volatility risk is moderate to high. NVIDIA’s debt is trivial relative to earnings and liquidity. The main downside scenario is not creditors forcing distress—it is an AI-demand slowdown or regulatory shock colliding with enormous supply commitments, concentrated customers, and investment exposure. Even then, its margins, cash generation, net-cash position, and discretionary buybacks provide a substantial cushion."
So broadly nothing I didn't more or less know already.
(2) is the point of this article. OAI and Anthropic are spending by far the most money of anyone in the space, as the article rightly notes, but they have no path to becoming profitable, meaning they cannot occupy that position forever. The other entities that rely on their spending to support their own margins - in this case the major cloud providers - are vulnerable to revenue collapse if OAI and Anthropic fail.
The premise most would disagree with is that the labs have no path to profitability. Two points support this: demand for inference is functionally infinite, or at least is so great that it is not meaningful to discuss its limits; and the labs are profitable on inference and are only taking losses to compete with one another. Some would extend this further and say that once the tech is good enough it will be able to drastically reduce their costs by some combination of speeding up research and creating efficiencies to reduce compute spend.
These are valid criticisms. But “AI has gotten better since he started saying ‘AI bad’” is not a reason to ignore the fact that major cloud computing providers are taking on massive new debt while becoming increasingly dependent on only two customers who face meaningful margin pressures. Unless OAI and Anthropic can find a durable moat and a means to exert pricing power, this is a serious issue going forward. That is true whether we wind up with a machine god (although we might have bigger problems in that case) or if we plateau at current capabilities.
With AI, by contrast - at least in its current state - there is no benefit to be gained from using the same model provider as somebody else. Switching is trivial for most use cases. Since they can’t capture consumers using network effects, the labs only have the levers of price and quality to pull to acquire and retain customers. To pull the price lever, they have to reduce their revenues; to pull the quality lever, they have to increase their expenditures. Indeed, they are sowing the seeds of their own demise by making inference cheaper and more efficient: since they can’t exercise pricing pressure, efficiency gains will be passed on to the consumer, which is unsustainable if your GPU debt is priced based on yesterday’s efficiency expectations.
If I’m using ChatGPT and I decide I want to use DeepSeek instead, I am only a couple of keystrokes away from doing it, and that’s if I have never used DeepSeek before.
As for your “corner the market” scenario, it’s possible, but unlikely. It is too easy to enter; even if you somehow got all of the major players to commit to growing their margins - and somehow manage not to violate the antitrust laws in the process - a newcomer could spoil the party far easier than it could in an industry like mobile phones (where you need tons of components, manufacturing capacity, network relationships, etc.) or ride sharing (where you need a large user base to justify your existence).
But the real sales pitch is that AI overtakes everything. That the YC cohort of 2032 will be just CEO, sales guy and a massive AI bill. It’s not entirely impossible IMO, either.
Then you totally get network effects. All your company documents, discussions, context etc are in there, your agents/employees whom you finally taught to do the job right. And I’m guessing the REST API for extracting your data is absent.
> https://www.bbc.com/news/technology-48227381
> https://www.forbes.com/sites/lensherman/2019/08/22/ubers-dub...
> https://americanaffairsjournal.org/2019/05/ubers-path-of-des...
Cory Doctorow, who is close to Ed Zitron and writes a lot in the same way about AI's economics, was of the same opinion: https://doctorow.medium.com/no-ubers-still-not-profitable-2b...
I used to believe this, so I'm not sure what to think of the AI market.
It also bears mentioning that if any particular AI lab manages to survive and succeed in the way Uber has, there will be several multibillion dollar corporate gravestones behind it. In fact, I don’t even think that nobody can be the Uber of AI. I just think it can’t be OAI or Anthropic. The debt is too great and it’s priced under old assumptions.
But the economics of it, at least from the outside, smell funny.
Some people are anxious that AI will take people's jobs. I'm not. I say this as a daily user. LLMs can be very useful, but they need to be carefully steered. It feels like a superpower the more I am an expert on the subject matter. What I am afraid of is that I suspect that once the dominoes start to fall, the economic downturn that it will spawn will be very, very painful.
Zitron is a bit histrionic, and this may put off people that don't like his style.
If you want a different, more balanced analysis, let me recommend you this: https://youtu.be/NufJ7g63KSY?is=Ojgb5pzrI-wg9wbo
Patrick Boyle's more recent video goes from a different angle and was very interesting for me, that have only a passing, layman's understanding of corporate accounting and investments.
And it's not like those numbers are static - they are probably up several times on a year earlier.
In July 2024 Anthropic's annualized revenues were ~$0.5 bn when Ed wrote
>And yes, that sound you hear is the slow deflation of the bubble I've been warning you about since March...
>How does GPT – a transformer-based model that generates answers probabilistically (as in what the next part of the generation is most likely to be the correct one) based entirely on training data – do anything more than generate paragraphs of occasionally-accurate text?
which illustrates how spot on he is with understanding AI.
For example, this article claims: "Every single story you’ve read about the “incredible growth” of these cloud platforms is an embarrassing misread of three companies that are misleading investors that will more than likely be forced in the next year or two to have to restate revenues, cut remaining performance obligations, and admit that they’ve drastically overbuilt capacity. "
In july 2024, Zitron wrote at length about how the economics of OpenAI were likely to collapse in the next 1 to 2 years [1].
It seems like the nearly inevitable collapse of generative AI is always 1 to 2 years away, but it's just the details of the intricate financial argument that change.
> I am hypothesizing that for OpenAI to survive for longer than two years, it will have to (in no particular order):
And listed various things like a technological breakthrough or more fundraising. I would bet he’s right that they’ll fundraise by the end of the year if they can’t IPO and pass the bag to retail investors.
What the AI bulls don’t seem to understand is that AI could be a great technological achievement AND an impending economic collapse because the valuations of these companies are absurd, basically requiring them to fulfill 10% of the country’s GDP within the next couple years. The achievements are impressive but vastly outpaced by the mania.
>What the AI bulls don’t seem to understand is that AI could be a great technological achievement AND an impending economic collapse
This doesn't seem to be Ed Zitron's position, though. He is always minimizing the use cases of AI, and seems to think virtually all of the demand is due to the technology industry manufacturing consent.
One of my clearest memories of the first tech bubble was of a number of people who accurately identified that it WAS a bubble and then predicted it would go pop 12-18 months before it did.
They attracted plenty of scorn and derision for being 80% right from people who were 100% wrong.
Some also lost a bunch of money - short selling really explodes in your face if you time it badly. It's not enough to know that it is a bubble, you have to be able to know when market sentiment will finally turn which is a gigantic gamble.
-- Isaac Newton
He seems to have quite a nuanced take that AI is a big deal but it's tricky to figure who will profit.
https://www.thestreet.com/investing/warren-buffett-alphabet-...
Personally I'm of the mindset that the current AI prices are too rich and that AI is very useful. Much like high speed internet in 2000. The prices were too high but the services themselves are great.
I have seen such people - notably Jim Chanos.
But, Zitron was vocal at the time where fawning over AI companies was basically mandatory everywhere else. He was the one bringing in numbers and, well, passion rather then fear when arguing that point.
On security, Brucr Schneier was also calling AI threat claims overblown repeatedly.
So the ~one guy who doesn't gets around.
- Huge subscribe CTA at top
- In-text subscribe CTA
- scroll through that get pop up in your face full page subscribe CTA
- close that and continue scrolling to yet another subscribe CTA
was enough to make me close the page and ask my agent for a summary rather.
When reading an article you’re supposed to analyze the actual thesis, not just evaluate if the person is an oracle or not
Very often predictions are useless to evaluate an analysis, you need to look at the underlying assumptions and facts, and see how they reflect reality. Ignore the person, look at the story and decide for yourself if it makes sense or not
Economists have entered the chat.
The first 2 page lengths are 70% taken by subscription and premium callouts.
I didn't read the article. Was too distracted and annoyed.
Does this strategy actually work on people?
> What is it you think you’ve gotten yourself into? Because I think you’re being sold a lie.
Hundreds of thousands of dollars of realized gains.
And the question I want to ask to anyone paying $70/year for Ed's newsletter: what has throwing the AI baby out with the bathwater gotten you?
That unless is pretty interesting.
Gotcha.