> You can be right for the wrong reasons

Also known as a broken clock being right two times a day.

The fallacy is assuming ahead of time that they are actually right this time.

If you start assuming they’re right and also admit that their logic doesn’t support that conclusion, there isn’t even any argument left. You’re just implying that you believe one thing is true and anyone who agrees with you, regardless of the reason, is also correct.

I don’t find that persuasive in any way.

You would think any investor would prefer to be informed that the numbers don't match expectations even if they then decide to go ahead anyway. People like to gamble sometimes even though the likelihood of losing money is very high.

It's not that the stock market isn't overvalued. It is that everything else is more overvalued. Especially real estate.

Friends businesses don't need our money, because banks are generous with giving out loans and friends prefer to keep ownership intact while being in debt instead of diluting it.

It fundamentally comes to investments/commitments divided by revenue over certain low number of years. Now add in any financing costs, overhead and actually operational costs.

Math ain't looking pretty. Just like it did not look pretty with WeWork. Might have worked, but still not great business to be in.

Being hyperscaler really does look very high capital and questionable margin bussiness at moment. With lot of competition.

>> Many people doubt. And we know that they do all sorts of funny book keeping shenanigans like SPVs and delayed deprecation.

Every time I see Anthropic and other AI companies getting ready to go public, I cannot stop thinking about the entire Enron debacle and all of their "unique" book keeping shenanigans that has far reaching impact, even to this day.

People keep telling me, "Nah this is totally different man, you can't compare the two!" But is it really??

but they also predicted failed IPOs

You're discussing something entirely different, which is fine, but you're not refuting anything that the parent poster said.

All they were saying is that they're terrible at predicting risk. Examples were given. It doesn't matter why they're bad at predicting risk, they simply are.

> None of all the examples you list here except CoreWeave has a comparably high cost base.

Wut? I'm pretty sure SpaceX spends a lot of money shooting rockets into space, and Cloudflare has pretty high infrastructure costs given that a very large portion of all Internet traffic goes through their networks.

It’s not “good research” when they are wrong >50% of the time. If you invested money in every IPO they told their subscribers to avoid you’d be very rich today.

I feel like the market staying irrational/gullible longer than you can stay solvent argument doesn't really make sense when all you're doing is NOT investing in something. It seems an aphorism more suited to puts and calls, puts especially, than it is to deciding not to invest at all.

> the market can stay gullible, for longer than you stay solvent.

This is essentially just what people say when current reality is different from what they want it to be. Sort of an unfalsifiable escape hatch response to “well observable reality contradicts your claim”

Track record is extremely relevant when the "argument" is mostly about judgement and subjective opinions.

In fact track record is perhaps one of the most important things of all when deciding wherever to even engage with the argument in the first.

Part of their argument is, "Look how reliable the speaker is."

TFA does not contain an argument beyond reporting on the statement

Their track record is directly relevant.

The rush towards open weight models defeats this thesis. There’s no prize for being “ahead.” That’s academically interesting but a terrible business plan. 99% of AI use cases don’t need the latest and greatest model, and businesses increasingly understand this.

> it's the fact that they have to train the next iteration right away due to competition

That it would be profitable if only there were no competition is not a really strong starting point for an IPO, though, is it?

(It is, however, the winner-takes-all world view Anthropic still seems to operate with)

> if this bet holds, a user could get the same or more done with a cheaper model (running-cost wise) while paying the same subscription

I think there will be a few forces driving prices lower if anything.

Everyone is currently looking at their spend on AI and the revenue growth for their own companies that this spend is not yielding.

I also think that if there is any competition, they won’t be able to lock in prices where they stand. Especially with open source models.

No doubt that if they wanted to they could be profitable.

But profitable enough to justify the ludicrous valuation?

> their current moat is the fact that they're ahead

the problem is that they are not ahead. they are ahead on producing one-shot demos, which are useless party tricks, but when it comes to shipping real products these large models are just a big expensive waste of time and what you need is deterministic systems that integrate small efficient purpose built models in the places where they are actually the best solution for the problem.

the anthropic game plan was to wrap the entire software industry up in a magic box called "AI" and then rule the world but these people are delusional and their house of cards is rapidly collapsing. the technology is real, but they didn't invent it and they don't own it. it is kind of amazing to think of a $2T company with zero IP ... they didn't invent a single thing that they are selling, and for the most part they are implementing it worse, and serving it less efficiently than at least a half-dozen other competitors.

>> Their inference margins are already high

So they make a profit on each burger they sell you, but however they keep limiting how many they are willing to sell you?

> Open source is a challenge but their current moat is the fact that they're ahead, and as long as they stay ahead they'll be able to charge a premium.

Being ahead isn't much of a moat in a game where playing catch-up is significantly easier than being ahead. Also, how much ahead of Chinese labs are they at best? A year? Two? That's a moat for ants!

Agree with you overall, but:

> Unlike the dot-com era, this is a bubble that everyone saw coming.

As far as I recall, the dot-com bubble was widely seen as such prior to it bursting. E.g. Greenspan and Buffet were both quite vocal about the irrationality of the values being assigned to companies.

Interesting take. How high is that probability within the next five years, do you think?

Nvidia probably doesn't really have really serious liabilities as long as it has actually produced all the GPUs it has sold on a buy-and-hold basis.

It's fabless, right? So they haven't even built out wasted manufacturing capacity.

As long as the GPUs they have sold are really sitting in a warehouse in Taiwan or somewhere in the USA, they are making properly concrete money.

The fact that nobody else is making real money, that will eventually come back to bite them, but Nvidia will just see its business shrink; it will have to scale back if the bubble bursts, but it will get out with much of that money intact.

My prediction is the same. I think MS, Google, and Amazon will be the winners here. They might take a hit when OpenAI and Anthropic blow up, bu they can absorb it, and are well positioned to provide those services in a reasonably sustainable fashion. I suspect the costs of using AI will be considerably higher at that point (subscriptions will be extremely limited or just non-existent). Meta does not benefit in this case since it doesn't provide compute, but they have money to waste (as they did with Metaverse, VR, etc).

This. The product is the stock, the tech is the fomo theater to sell the product.

If they are a year from AGI, then the private investors can hang on and use their new super intelligence on their own tasks to make themselves considerably more wealthy now and into the future, so why would they want to sell that power to anyone else?

> These companies believe they're a year from AGI, and that's how they'll be profitable.

What I don't get is that if Company A is a year from AGI (I don't think that's true, but then again, I don't think anyone even has a good definition of AGI), then surely Companies B–ZZ are only a few years (at most!) from AGI as well. There is no moat. Or if there is a moat, the moat is hardware scarcity – but that's systemically outside of the control of these companies.

Most here need a finance lesson.

In intrinsic valuation, the value of the operating assets is determined by ebit(1-t) - reinvestment.

Gross profit doesn’t mean a whole lot if a) your operating income is negative b) your reinvestment needs are mega. You have to make some interesting assumptions Re. Anthropic to get a 2T valuation.

I almost have to hold my nose reading finance posts on here.

BullSHIT.

You sure they're not including training costs in those probability claims? I've not seen confirmation one way or the other on that.

Remember, when you see “EBITA”’think “bullshit earnings”

Careful, there are word plays used here:

Anthropic's annualized revenue run rate reached $65 billion at the end of July 2026, up from just over $5 billion in August 2025 (Anthropic, CNBC) Anthropic made nearly $4.6 billion in revenue in 2025, 12x more than in 2024 (IPO prospectus, via Reuters, September 2026) Anthropic's preliminary Q2 2026 revenue was more than $11.5 billion, ahead of OpenAI's $6.7 billion (Bloomberg and The Wall Street Journal, August 2026)

https://explodingtopics.com/blog/anthropic-revenue

So, the Q2 revenue was $11.5 billion. The claimed $65 billions are "annualized revenue run rate". Please show me where that term is defined reliably. It's a BS marketing term to obfuscate real revenue after the tokenmaxxing hype has ended.

Thanks, good catch. Updated my comment.

I take no position on the correct valuation of Anthropic, but I think if you've used coding agents, it's pretty clear that it's not just hype; they really are a useful technology that can massively increase software engineering productivity.

As far as I know the circular deal fears are more around the chip purchasing agreements, because there is a speculative component to building out datacenters for training which may not produce models that are worth the training cost, but the revenue from inference is real and it has truly grown a lot over the past two years.

With the NASDAQ rule change, even more so.

https://ir.nasdaq.com/news-releases/news-release-details/nas...

That’s a good question that I assume will be cleared up in the prospectus, but I have to assume that enough of it is just cash coming in that they and OpenAI are likely the fastest growing companies of all time, or at least very high up the list.

That’s a risk for sure.

> this is the greatest invention of all time

I think it's a little early to make that statement. Could be. But at this point I think things like printing press, agriculture, writing ... they're all many times more significant.

I agree with you about the greatest invention of all time part, the rest is where my crystal ball gets hazy.

> I’m buying, this is the greatest invention of all time and Anthropic has garnered a pretty solid lead right now.

Depends on your perspective, but that's definitely true if you're selfish and have been paying people for intellectual labor. Lots of layoffs to fantasize about (even though that represents vast amounts of pain and destruction for many people).

Well, unlike the economy I can’t grow in size rapidly. (And if I did I’d just have another major problem.)

Induced demand and limitless cheap human-quality labor can possibly grow the pie very fast, though even then, you start running into the earth’s finite resources eventually. But it is possible that they are worth much more than $2 trillion long before that.

But that is why I said you can’t just safely assume 15x multiples to continue.

The article isn't about AI. It's about investment. And that quote I think tells us something about the man. Who is likely the same man he was in 2021

I agree it’s not rocket science, but I think there’s more to consider here.

First, a lot of usage can switch to open models (like is happening, I agree), and yet frontier model usage can still grow (like is happening). As long as the market for frontier level intelligence keeps growing, the cost of frontier intelligence is marginally profitable to the user, and the open weight models don’t catch up, I don’t see any reason for frontier level usage to fall. Indeed, we haven’t seen any of that so far, despite the rise in open weight usage. My own product has started using open weight models, but my usage of frontier models also continues to grow, because they keep unlocking new capabilities. Maybe that’ll end, but I’m skeptical. I think the problem space that can benefit from the best intelligence money can buy is unbounded.

Plus the labs can still offer cheap non-frontier models to compete with the open weight models (like Luna and Haiku), and they have a massive advantage in terms of available compute.

Secondly, I don’t think that the big labs need mass layoffs to justify their valuation. That’s perhaps one path, but in my view a more likely one is that we get greater increases in productivity and economic growth, and they capture a large share of that. If global GDP grows 2% faster than it otherwise would over the next decade, that’s an extra $175 trillion in economic value created over the next decade. If it grows 5% above baseline growth, it’s half a quadrillion. GDP in 2036 would be more than $100T higher per year than if GDP grows at historical rates. AI companies don’t have to capture very much of that to justify a $2T valuation.

Now, will any of that happen? Who knows, it’s definitely unrealistic and unprecedented, but then again, so is much of what I see happening in the world today. Time will tell.

Anthropic, once it goes public.

While I think SpaceX is still horribly over-valued they at least have a diverse line of businesses that are very hard to copy.

Anthropic is a one trick pony and the world is discovering lots of other cheap ponies that do the same trick.

I suspect SpaceX is going to be the opposite of what will happen with Anthropic. With SpaceX I think the overwhelming majority of people didn't initially understand what Starship is, how far along it is, or what it even means in terms of its impact on the pathway of getting humans to be a multiplanetary species. So it was greatly undervalued and zoomed upward.

With Anthropic I think a lot of people haven't really realized the trend line of cost of "AI" (both in literal $ terms, and compute costs for local operation) or what open weight systems will do to the market in the "long run", where long run in this case means the next year or two. I don't think any LLM company is going to come out a winner in the longrun, because the price for LLM output is going to approach $0 with no chance for a moat whatsoever.

They made it convenient to get value for the specialist masses and I don't think anyone else managed to achieve it yet. That's their moat.

I've seen firsthand attempts at integrating open source in place of it and they all failed. Time is money and it costs premium. Touching the ground running is a huge advantage. That's why cloud services stuck.