Comment by joegibbs
14 hours ago
Anthropic at $1t for an IPO vs Google at $23b in 2004 sounds insane but Google's revenue at the time was $2.7b while Anthropic's already at $47b, so a valuation at about 20x vs 10x revenue. Anthropic also has very high revenue growth (50x since 2024), it doesn't seems quite as insane as it could be.
That is revenue. What is the net profit?
If you are growing revenue at a high rate then taking profit is a misallocation of resources. That is short-term thinking. It is much better to reinvest in revenue growth.
You can take small profit now or much larger profit later. Insisting that companies need to be profitable even when growing revenue rapidly is failing the marshmallow test.
The point is that the unit economics are way worse because inference is expensive. Cost of goods sold matters, even if you're reinvesting profits.
13 replies →
I guess net isn't the relevant measure, but what are the unit economics? Are they actually making money selling tokens?
1 reply →
That's not how classical valuations worked though. I was taught the rough shorthand for valuing a company was profits / real_interest_rate, treating the company like a perpetuity. Revenue ain't in it. Now we have a bunch of "We'll make it up on volume companies!" like https://www.youtube.com/watch?v=CXDxNCzUspM
1 reply →
> That is short-term thinking.
Then why IPO? Isn't that even shorter term thinking?
1 reply →
While good to ask, that is less relevant so long as they can maintain runway.
Please don't ask those rational questions, revenue is all that maters.
Doesn't inference have very good profit margins* but all the losses come from training?
* For now, when they don't have to compete much against companies like DeepSeek who supplies inference at 1/10th of the cost
> Doesn't inference have very good profit margins* but all the losses come from training?
There is no source for this. Amodei just pulled a hypothetical explicitly distanced from Anthropic out of his ass and kickstarted some citogenesis when people half-remembered that number and started quoting it as truth.
The only material claim of Anthropic is that they would "turn an operating profit of $559 million in the June quarter ... The company might not remain profitable for the full year as it plans spending increases due to its vast computing needs." with an explicit disclaimer that: "It is unclear what accounting methods Anthropic has used to book revenue and costs, as the company isn’t yet required to follow the financial-reporting requirements of a public company."
https://www.wsj.com/tech/ai/mind-blowing-growth-is-about-to-...
This is the exact same quarter where xAI is giving them deeply discounted compute, as such the numbers cannot be projected out to the later quarters once Anthropic has to actually pay xAI for the compute they use.
Finally, there's the reality that were the revenue numbers any good, Anthropic would just publish them and leapfrog OpenAI. That they do not provide clear GAAP numbers suggests the numbers are bad.
They reported 559 million in Q2 of this year. OpenAI on the other hand, is nowhere near this.
because of the mutual sweetheart deal with SpaceX
SpaceX gave em discount for the pre IPO quarter so they can show profit
Anthropic signed a deal to lease compute that is the bulk of SpaceX revenue
2 replies →
Not using GAAP, so this is just PR for them.
I suspect the answer is: in the future after Moore's law somehow inevitably does its thing.
That has worked in the past for tech infrastructure, so there is clearly a gamble that it does that again here.
These AI companies will be able to jack prices way way up once companies and users are fully addicted to doing everything with their AI.
What’s defensible about Anthropic’s revenue? It seems like OpenAI and others are equivalent. Open weight models are catching up. Google has ads networks, video platforms, and so much more.
I am skeptical that Anthropic and OpenAI can defend their dominance for long enough to make meaningful gaap accounted profits
Anthropic seems to have clawed its way to being the best AI and charging for itself. Microsoft had to slash the Anthropic budget… which it exceeded while being the exclusive host of OpenAI.
Google seems to have a good B2B and internal leveraging AI to make $. OpenAI/Microsoft seems to have squandered an early product lead.
And then you have the Muskiverse, where we have an rocket ship company that buys surplus cyber trucks, operates a space ISP, an AI company that produces virtual fetish porn and makes money renting GPUs to Anthropic, a rando dying social network and a tunnel company to cock-block public transit.
I may be underestimating the market for AI anime porn, but I think Anthropic is probably the best in class product right now. Google and AWS are probably the best positioned sellers of AI. SpaceXAI is the dark horse because they are likely enriching the dear leader more. OpenAI is fucked.
> an AI company that produces virtual fetish porn and makes money renting GPUs to Anthropic
Whatever Anthropic is paying is too much, since it means xAI will get to observe Anthropic's software, weights and operations in detail. It's probably contractually prohibited from doing so, but I doubt that would stop Musk, given what's at stake.
1 reply →
Anthropic is profitable unlike OpenAI though. Sure they'll owe a lot of money for probably decades, but if they remain profitable moving forward, it will be worthwhile.
That profit figure is a pre IPO marketing claim, not an audited and GAAP accounted number. And there is already a lot written about how Anthropic exaggerated revenue compared to OpenAI.
https://www.forbes.com/sites/josipamajic/2026/03/25/openai-a...
1 reply →
The question still remains whether they will be defensively profitable when things settle down.
I don't think open weight models are likely to overtake or match frontier models in the next year or so when it comes to doing the most difficult tasks, but I do expect a lot of people who are currently funneling wheelbarrows of money to Anthropic to realize that they can achieve the vast majority of things they are doing with LLMs just as well with much cheaper open weight models.
1 reply →
I think their real moat is the grip they have in corporates.
Once they are in, they will catch most of the opportunities.
They will defend it the way any good monopoly always does: buying the competition. Case in point is Facebook: it is just a social network, the way they really stay on top is buying other companies and paying for people to spend even more time on their properties.
They bought insta and WhatsApp but at the time neither were really social networks. Insta was a popular filtering app. WhatsApp is just recently turning into a social network with their status updates.
Looks like it'll be more like $2t
https://polymarket.com/event/anthropic-ipo-closing-market-ca...