Musk Changes Tone In Surprise Anthropic Alliance

For months, Elon Musk publicly attacked Anthropic. Today, however, the billionaire opens the doors of one of the most powerful supercomputers on the market to them. Behind this surprise agreement between SpaceXAI and the company behind Claude lies a reality that has become central in the AI war: dominance no longer depends only on models, but especially on access to GPUs, energy, and infrastructures capable of powering the next generation of artificial intelligences.

Anthropic Gains Computing Power from Giant SpaceXAI

Anthropic, despite its ban by the Pentagon, has signed a major partnership with SpaceXAI to use all the capacities of the Colossus 1 data center located in Memphis. The deal marks a new stage in the global race for artificial intelligence, where dominance now comes through access to infrastructures capable of powering the most advanced models.

Here are some key points :

• More than 220,000 Nvidia GPUs would be mobilized as part of this agreement ;
• The site would have a power greater than 300 megawatts ;
• Anthropic claims that this partnership will significantly increase its computing capacities ;
• The company states it will use the full computing power of SpaceX’s Colossus 1 data center ;
• The new resources should enhance the performance of Claude Code as well as the Claude Pro, Claude Max, and developer API offerings.

This increase in capacity comes at a time when the industry faces a persistent shortage of advanced GPUs. For large AI labs, securing massive computing capacities becomes a strategic issue as important as model development itself. Anthropic thus seeks to consolidate its position against OpenAI, Google, and xAI in a technological battle where infrastructure has become a major lever of domination.

Elon Musk Turns Coat on Anthropic

The deal also surprises with Elon Musk’s radical change of tone towards Anthropic. In recent months, the billionaire had publicly called the company “misanthropic” and “evil”, accusing some AI players of developing ideologically dangerous models. This time, Musk takes a much more conciliatory position. Commenting on the partnership, he said: “nobody triggered my malice detector”. Such a phrase contrasts sharply with his previous criticisms.

This rapprochement fits within a monetization strategy for SpaceXAI’s infrastructures. Several American media already talk about a potentially massive future IPO for the company, while its computing capacities become a commercial product in their own right. Anthropic and SpaceXAI are even reportedly discussing projects related to orbital data centers capable of powering future artificial intelligence models from space.

This evolution shows how the AI competition is reshaping not only the alliances in the technology sector but also investment priorities in venture capital. Ideological rivalries now take a back seat to the colossal industrial needs of AI labs. As models become more energy- and compute-intensive, companies able to control the infrastructures may become as important as those developing the models themselves.