points by SR2Z 3 days ago | 0 comments
DeepSeek announced that each GPU they have makes about $38 an hour in revenue during off-peak times. That's $331k a year, and DeepSeek is cheap.
I think that it's not only possible but likely that the frontier labs are profitable on an ongoing basis, and it's mainly the research and capex which make them look like impending disasters.
Realistically, OpenAI and Anthropic are each probably worth ~$1T each. Anthropic can already easily justify that based on revenue alone. OpenAI isn't far off. Consider how quickly these companies are growing and the fact that they maintain a small but significant edge, I don't think the bubble is going to pop. Just deflate a little.
[–] mschuster91 3 days ago
[–] SR2Z 2 days ago
Having a frontier model is worth something beyond just inference costs, I'm sure. It seems likely that the labs will maintain at least a small lead.
And that is the much more relevant problem. Capable open-weight models from China, European labs or competitors with cash to burn (i.e. Meta!) erode all the moat that the major labs have, and tooling functionality like MCP, CLI interfaces or integration into third-party tools like IDEs quickly gets cloned.
Simply printing money by running inference does not justify valuations in the billions of dollars, because anyone can do so.