Is xAI the new cloud now?

XAI and Anthropic announced on Wednesday A surprising partnership This led to the Claude maker purchasing “all of the computing capacity in the Colossus 1 (xAI) data center,” approximately 300 megawatts of which allowed Anthropic to immediately lift usage limits. It’s a huge deal for xAI, potentially worth billions of dollars. Most importantly, it was monetized immediately One of the company’s most impressive achievements,Transforming XAI from a consumer to a computing provider.

It’s tempting to see this arrangement as a shot at OpenAI amid the ongoing litigation. But Musk’s explanation for the X is that xAI technology already exists Moved training to a newer data center, Colossus 2And xAI simply didn’t need both.

In the short term, there is clear business logic at work. Existing xAI products mostly focus on Grok, which is exactly what happened Declining usage Since the image generation disasters earlier this year. If building an xAI data center is much more than Grok needs to operate, the partnership with Anthropic adds a lot of green to the balance sheet. This is especially useful as the company, along with SpaceX, accelerates toward an IPO. More broadly, having Anthropic as a customer makes it easier to believe SpaceX’s orbital data center plays a role It might actually work.

But beyond the short-term benefits, the humanitarian partnership sends an unusual message about where Elon Musk’s priorities really lie. This suggests that the company’s real work may be more about building data centers than training AI models.

It’s rare to see a major tech company handle computing resources this way when companies like Google and Meta, which also serve as training models, are building more data centers. It’s easy to miss this point, because many of these companies act as enterprise AI vendors, online services, and cloud providers simultaneously. But when they have to choose between selling more of the computing available to customers and keeping some to build their own tools, they reliably choose Door 2.

Just last month, Sundar Pichai admitted in a phone call that Google Cloud’s revenues were lower than they could have been because the company was “capacity constrained” — and when given the option to lease its GPUs or use them to develop AI products, Google chose the AI ​​products.

Facebook faced a more extreme version of the same constraint, spinning up a brand new cloud machine just to make sure it would have enough GPU power to chase Mark Zuckerberg’s AI ambition. As he put it when announcing Meta Compute in January, “How we engineer, invest and partner to build this infrastructure will become a strategic advantage.”

The key word there is “strategy”. Both Zuckerberg and Pichai look forward to a future in which artificial intelligence powers the world’s most popular and profitable systems. Computing power is not just a way to meet the demand for inference today, but to build the products of tomorrow – and running out of compute means missing out on this opportunity.

By focusing on data centers (terrestrial and otherwise), xAI is positioning itself more like NuCloud: buying GPUs from Nvidia and leasing them to model developers like Anthropic. It’s a much tougher business, pressured by chip suppliers and shifting demand cycles. Reviews of most new active clouds reflect this fact: xAI was With a value of $230 billion In the January financing round; CoreWeave, which oversees a similar amount of computing power, is worth it Less than a third of that.

Musk’s version of the new cloud is more ambitious, as you might expect. Some data centers could be in space, at least by 2035, if all goes according to plan. xAI will manufacture its own chips In TerraFabwhich will take away some but not all of Nvidia’s pricing power. But none of this changes the basic economics of the new cloud business.

Recently February all handsXAI had real ambitions in software. This was the show that unveiled the orbital data center project, but it also sparked major programming ambitions (it has since been boosted by Index partnership) and interesting ideas such as leveraging the use of computers in large-scale digital twins (in the unfortunately named Macrohard project). These are the types of long-term projects that need dedicated computing resources to succeed. As long as xAI sells large amounts of computing to its competitors, it’s hard to think such new ambitions have much of a future.