SpaceX supplies computing power to Anthropic as Musk sues OpenAI

Anthropic has reached an agreement to use computing resources from SpaceX, a company owned by Elon Musk, marking a reconciliation with a former critic and benefiting both in the high-stakes race for artificial intelligence.

Under the deal announced Wednesday, Anthropic will use the full computing power of the Colossus 1 facility in Memphis, Tennessee, which houses more than 220,000 Nvidia processors and provides the maker of the Claude chatbot with 300 megawatts of new capacity within a month. That amount of electricity is enough to power more than 300,000 homes, as Anthropic, led by Dario Amodei, seeks to boost capacity for its Claude Pro and Claude Max AI assistants for subscribers.

The announcement came on the same day Anthropic held its developer day and unveiled a new feature for the Claude AI system called "dreaming." The tool allows the AI system to review work between sessions, detect patterns, and update stored preferences and other user context files.

Available now as a research preview, "dreaming" comes with agent management software — AI programs that perform tasks with minimal human intervention. Earlier, Amodei warned that AI could wipe out 50% of all human input jobs within 1-5 years.

Driven by the additional capacity from the SpaceX deal and similar agreements, Anthropic said it is doubling the rate limits of Claude Code for paid plans, removing peak-hour usage caps for Pro and Max accounts, and sharply increasing the volume of requests developers can send to Claude Opus models.

The company also expressed a desire to collaborate with SpaceX on developing multi-gigawatt space-based orbital data centers — one of Musk's key goals and a major driver behind SpaceX's initial public offering (IPO) — which are expected to be extremely expensive and technically challenging.

In a statement on X, Musk said he decided to lease computing power to Anthropic after spending time with their leaders last week. He was impressed by their efforts to ensure Claude AI is "good for humanity."

"No one triggered my bad-detector," wrote Musk, who is suing OpenAI and CEO Sam Altman in court, accusing them of violating the mission to develop AI for the benefit of humanity and maintain nonprofit status.

Last week, Musk alleged in court that in late 2022, Altman tried to "steal the charity" and said, "That turned out to be true."

Musk, a co-founder of OpenAI, also claimed that OpenAI would not exist without him. The world's richest man said he founded OpenAI after a conversation with Google co-founder Larry Page, whom he believed lacked concern for AI safety.

"I said, 'What if AI wipes out all of humanity?' He said that's fine as long as artificial intelligence exists. I said that's crazy. And then he called me a 'speciesist' because I care about humans more than AI… The reason OpenAI exists is because Larry Page called me a 'speciesist'… What is the opposite of Google? A nonprofit open-source organization."

OpenAI's lawyers argue that Musk's lawsuit is an attempt to undermine the company led by Altman to benefit his own xAI, which is part of SpaceX.

Musk's deal comes as the administration of U.S. President Donald Trump signed new agreements with other AI giants, excluding Anthropic. On Tuesday, tech giants Microsoft, Google, and Musk's xAI agreed to give the federal government access to inspect their new AI tools amid concerns about hacking tied to Anthropic's technology. Last week, the Department of Defense also announced agreements with 10 major AI companies to use their technology on classified networks. Anthropic was not part of that agreement.