Elon Musk's SpaceX Signs Anthropic Compute Deal

Elon Musk's SpaceX Signs Anthropic Compute Deal, Reversing Months of Public Hostility

In one of the more surprising pivots in recent AI industry history, Elon Musk's SpaceX and Anthropic announced on May 6, 2026 that Anthropic will gain access to the full compute capacity of Colossus 1, SpaceX's AI data center in Memphis, Tennessee. The deal hands Anthropic more than 220,000 Nvidia GPUs and over 300 megawatts of new compute capacity within one month — enough electricity to power more than 300,000 homes — and arrives just weeks before SpaceX's anticipated IPO roadshow. The announcement marks a dramatic about-face for Musk, who had publicly called Anthropic "misanthropic and evil" and written on X in February 2026 that "Anthropic hates Western Civilization."

What the SpaceX–Anthropic Compute Deal Actually Includes

Under the terms of the agreement, Anthropic gains access to all of the compute capacity at the Colossus 1 facility in Memphis. The site, originally constructed for Musk's xAI venture, delivers more than 300 megawatts of power and over 220,000 Nvidia GPUs — a substantial injection of infrastructure for a company that has been struggling to keep pace with surging demand for its Claude AI models.

The practical effects were immediate. According to TechTimes and Coindesk, Anthropic doubled Claude Code's rate limits for paid and enterprise customers and removed peak-hour usage caps for Claude Pro and Max subscribers as a direct result of the deal. Tom Brown, co-founder of Anthropic, confirmed the rapid rollout on social media: "In the next few days, we'll be ramping up Claude inference on Colossus. Grateful to be partnering with SpaceX here."

The partnership also carries a longer-range ambition. In a blog post, SpaceX stated: "As part of this agreement, Anthropic also expressed interest in partnering to develop multiple gigawatts of orbital AI compute capacity." That would represent a significant expansion into space-based AI infrastructure, though it remains at the exploratory stage.

It is worth noting that the Colossus 1 facility has not been without controversy. According to TechXplore, xAI installed dozens of natural gas-burning turbines to power the site, claiming no federal permit was required because they were for temporary use. The move drew protests from civil rights groups over air pollution concerns in the Memphis area.

Why Anthropic Needed the Compute — and Why Musk Said Yes

Anthropic's compute crisis is rooted in growth that outpaced even the company's own aggressive projections. Speaking at Anthropic's developer conference in San Francisco on May 6, 2026, CEO Dario Amodei explained that the company had planned for 10-fold growth, but revenue and usage increased 80-fold in the first quarter on an annualized basis. "That is the reason we have had difficulties with compute," Amodei said. That growth trajectory is reflected in the company's financials: Anthropic's revenue run rate surpassed $9 billion in January 2026, up from roughly $1 billion at the end of 2024, according to Futurum Group.

Anthropic has been building out infrastructure aggressively across multiple fronts. Its broader compute portfolio includes deals with Amazon, Google, Broadcom, Microsoft, and Nvidia. According to Blockonomi, Anthropic previously secured up to 5 gigawatts of compute capacity from Amazon. The company has also committed to spending more than $100 billion on Amazon Web Services over the next decade and has a $50 billion U.S. AI infrastructure investment underway with Fluidstack, starting with data centers in Texas and New York. Even against that backdrop, the scale of demand growth made additional capacity urgent.

For Musk, the motivations appear equally practical. According to Axios, SpaceX acquired xAI in January 2026 in a deal valuing the two combined companies at $1.25 trillion, and on the same day as the Anthropic announcement, Musk declared on X that xAI "will be dissolved as a separate company" and rebranded as SpaceXAI. That consolidation leaves SpaceX holding significant compute infrastructure that needs to generate revenue. Adding a high-profile AI company as a named compute customer — just ahead of what could be one of the largest IPOs in corporate history — strengthens SpaceX's pitch as an AI infrastructure provider. According to Coindesk, SpaceX filed confidentially with the SEC on April 1, 2026, for an IPO targeting a valuation of $1.75 trillion to $2 trillion, with the public S-1 expected by late May and a roadshow set for the week of June 8.

Axios also reported that SpaceX has a separate deal with coding startup Cursor that gives it the option to buy the startup for $60 billion, further signaling the company's ambitions in the AI software and tooling space.

From "Evil" to Business Partner: The Context Behind Musk's Reversal

The timing and backdrop of the deal add layers of strategic complexity. According to Al Jazeera, Musk spent the prior week in federal court in Oakland, California, testifying in his ongoing lawsuit against OpenAI and its CEO Sam Altman. The Anthropic deal — announced while Musk was actively litigating against his most direct AI rival — allows him to simultaneously monetize idle compute capacity and align with a competitor to OpenAI, all in the same week.

The reversal from Musk's earlier rhetoric is stark. As recently as February 2026, he wrote on X that "Anthropic hates Western Civilization" — a claim he appears to have set aside entirely. According to CNBC, Musk said after meeting with senior Anthropic staff that he came away with a markedly different impression. "Everyone I met was highly competent and cared a great deal about doing the right thing. No one set off my evil detector," Musk said. He also offered a conditional endorsement of Anthropic's flagship model: "So long as they engage in critical self-examination, Claude will probably be good."

Musk did not entirely abandon his characteristic caution, however. He attached a notable condition to the arrangement: "We reserve the right to reclaim the compute if their AI engages in actions that harm humanity."

The deal also arrives against a difficult regulatory backdrop for Anthropic. According to CNBC, the Pentagon declared Anthropic a supply chain risk in March 2026, blacklisting it from work with the U.S. military after the two sides failed to reach an agreement on how the company's models could be used. Anthropic subsequently filed suit against the Trump administration in San Francisco and Washington to reverse that designation. Partnering with SpaceX — a company with deep ties to U.S. defense and government contracting — does not automatically resolve that dispute, but it does reposition Anthropic within a different part of the national security ecosystem.

What Comes Next for SpaceX, Anthropic, and the AI Infrastructure Race

The immediate near-term milestone is SpaceX's IPO process. With a confidential SEC filing already in place and a roadshow scheduled for the week of June 8, the Anthropic deal lands at a moment when SpaceX has every incentive to demonstrate that its AI infrastructure assets are generating real commercial revenue. The combined valuation target of $1.75 trillion to $2 trillion would make it one of the most significant public offerings in market history, and the ability to point to Anthropic — one of the fastest-growing AI companies in the world — as a compute customer adds a tangible data point to that story.

For Anthropic, the path forward involves managing an infrastructure buildout of unusual scale and complexity. The company now draws compute from SpaceX's Colossus 1, Amazon, Google, Broadcom, Microsoft, Nvidia, and Fluidstack, while simultaneously pursuing the speculative goal of orbital compute capacity with SpaceX. How effectively it can coordinate that portfolio — and whether demand growth continues to outpace even these expanded resources — will be a central operational question for the remainder of 2026.

What is clear is that the AI compute race has reached a point where ideological rivalries and public feuds are increasingly subordinated to infrastructure necessity. Musk's three-month journey from calling Anthropic "evil" to leasing it his most powerful data center is a data point, not an anomaly.

For more tech news, visit our news section.

What This Means for Your Productivity

Expanded compute capacity translates directly into faster, more reliable AI tools for the people who use them every day. The removal of Claude's peak-hour usage caps and the doubling of Claude Code rate limits are early signs of what happens when AI infrastructure catches up to real-world demand — and the race to build that infrastructure is only accelerating. Whether you use AI for deep work, health research, or daily task management, the underlying compute battles being fought right now will shape what those tools can do for you. Join the Moccet waitlist to stay ahead of the curve.