XAI Will Breakeven by Renting AI Data Center to Anthropic

by

XAI gets to about breakeven by renting half of their AI Data center and Anthropic can serve its models and more money and can get to a better IPO. Both get to IPO before OpenAI. Both stronger against OpenAI. OpenAI IPO is hurt. Anthropic is already at pre-IPO valuation of $1.2 trillion. SpaceX XAI will IPO first with a public S-1 maybe next week. Anthropic will be the second IPO. How viable will the OpenAI IPO be? We will not know until after the lawsuit.

SpaceX/xAI has signed an agreement with Anthropic to provide access to its Colossus 1 supercomputer. Anthropic plans to use this additional compute to directly improve capacity for Claude Pro and Claude Max subscribers. As part of this agreement, Anthropic also expressed interest in partnering to develop multiple gigawatts of orbital AI compute capacity. The compute required to train and operate the next generation of these systems is outpacing what terrestrial power, land, and cooling can deliver on the timelines that matter.

SpaceX valuation will go higher when people realize XAI is no longer a billion a month loss problem. XAI story during the roadshow will be good. There is near term rentals for roughly breakeven and longer term AI for developer business. Keep building more data center and roughly be able to breakeven or better.

Getting more Cursor revenue and margins will also help XAI-SpaceX.

4 thoughts on “XAI Will Breakeven by Renting AI Data Center to Anthropic”

  1. Meta was poaching top talent a lot. I think they are one of the worst to get hands on super powerful AI. They will just use it to make people even more addicted on Facebook. I dont trust OpenAI. If Musk really means, that he wants to build maximum truth seeking AI, that is very good. At Anthropic they seem to be more moral compared to some other companies. I dont want that company, that wins AI race uses AI to dominate us all. There has to be some ethical oversight.

  2. I dont think it is a failure. It is a good model with unique flavor. 4.3 had some bad ratings, but 4.2 is still very high, although not top one.

    Brian made some good points. But still there is a question why not use all that compute for training the best model, to dominate the market and have self driving as fast as possible? Why rent? Did they hit the wall and increasing compute wont work? They are very good at engineering, but it seems on the software side they are behind Gemini and Anthropic. Possible that they cant compete with them.

  3. This implicitly means GROK is probably a failure and they have given up for now trying to compete. It is well established that LLMs scale well with compute capacity. The argument for Colossus was to have the most compute resulting in a superior model.
    Now they basically say that they won’t use this Colossus for their own model and instead sell raw compute. It’s like selling electricity instead of selling products made out of electricity higher up in the value chain.

    • Yes, they are selling computation to competitor because the infrastructure they built is either underused (Grok does not have that many requests) or does not justify the cost (grok might be used but is so inefficient/lags behind so much that will not catch up in any cost effective way that is economically better to scale it down and rent infrastructure to competitors)

      This also means that xAI valuation was astronomically inflated, as their best asset is not the AI model but the infrastructure which is definitely an asset way simpler to replicate than a special “secret recipe” AI model

Comments are closed.