SpaceX Leases 220,000-GPU Colossus 1 Data Center to Anthropic
Elon Musk announced Wednesday that SpaceX will provide compute capacity from its Memphis facility to Anthropic after the AI company demonstrated efforts to ensure its technology benefits humanity. The deal supplies Anthropic with 300 megawatts from more than 220,000 Nvidia processors while SpaceX shifts its own AI training to Colossus 2.
New York PostSpaceX will lease its Colossus 1 artificial intelligence data center in Memphis, Tennessee, to Anthropic, the companies announced on Wednesday. The facility houses more than 220,000 Nvidia processors and will give Anthropic 300 megawatts of new capacity within a month. Elon Musk stated that SpaceX will provide compute to AI companies taking the right steps to ensure it is good for humanity.
He added that SpaceX reserves the right to reclaim the compute if their AI engages in actions that harm humanity. Musk said he would provide computing capacity to other AI companies that make similar efforts to favor humanity. He compared the arrangement to how SpaceX launches satellites for competitors with fair terms and pricing.
The decision followed Musk spending time with Anthropic leaders last week. Their work to ensure Anthropic’s Claude AI is “good for humanity” impressed him, he said. “No one set off my evil detector,” Musk wrote.
SpaceX has moved its AI training efforts to Colossus 2. Musk made the comments in a statement on X as Anthropic held a developer day in San Francisco on Wednesday. ” The feature is meant to help AI systems learn by reviewing work between sessions, spotting patterns, and updating files that store user preferences and other context.
Anthropic is doubling Claude Code’s rate limits for its paid plans. It is removing peak-hour usage caps for Pro and Max accounts and sharply increasing the volume of requests developers can make to its Claude Opus models. Anthropic has expressed interest in working with SpaceX to develop multiple gigawatts of space-based orbital data centers.
The collaboration aligns with one of Musk’s key goals and a major driver behind SpaceX’s initial public offering. SpaceX’s IPO is expected in June. It is poised to raise tens of billions of dollars and value the company above $2 trillion.
If the IPO occurs as planned, the company owning a rocket manufacturer, a social media app, and an AI chatbot developer will rank among the world’s top 10 largest publicly-traded companies. On Wednesday, Randi Weingarten, president of the American Federation of Teachers, wrote to the US Securities and Exchange Commission urging scrutiny of SpaceX’s IPO preparations.
8 million members whose retirement accounts are likely to hold SpaceX shares soon after the offering.
“Weingarten cited concerns about the company’s business plans potentially relying on nonexistent or speculative technologies, the reported thoroughness of its accounting, and the adequacy of its board of directors,” according to a letter shared with WIRED. ” AkademikerPension has been divested from Tesla since last year. ABP sold off a stake approaching $600 million in Tesla in 2024.
Both funds are reviewing potential SpaceX investments with caution. Activists demonstrated outside of meetings with investors that SpaceX held in Tennessee and Texas last month. Environmental groups expressed fears that the company’s growth could fuel further pollution.
Musk testified last week in his trial against OpenAI CEO Sam Altman. In February, Musk accused Anthropic’s AI of bias and wrote that he doesn’t think there is anything that can escape the inevitable irony of Anthropic ending up being misanthropic. The lease marks a reversal from those earlier criticisms.
Musk described SpaceX as “life insurance for life as we know it” during last week’s testimony, emphasizing its goal of making humans a multiplanetary species to ensure long-term survival of life and consciousness.
Transparency
Rewrite inherits consensus framing by centering Musk's personal 'evil detector' and humanity tests while burying substantive lease details and tacking on activist criticism and past conflicts at the end.
Lede misdirection: Lede foregrounds the deal announcement and messenger over the substantive 220k-GPU facility and 300MW capacity
The same facts could be read as SpaceX responsibly monetizing surplus supercomputing capacity to accelerate safe AI development aligned with humanity-first principles while funding its core mission of making life multi-planetary.
3 independent outlets report the same core facts. This score blends how many outlets corroborate, their editorial tier, and how closely their facts agree — it measures corroboration, not proof.
Sources framed at 68 → our rewrite 55. We stripped 13 points of framing the sources carried in.
Story details
Related Stories
japantoday.comAnthropic Confidentially Files for IPO After Raising $65 Billion
The artificial intelligence company behind the Claude chatbot submitted its filing on Monday, weeks after completing a funding round that more than doubled its valuation from February. Anthropic did not disclose the size or terms of the offering. The move comes as the global IPO…
thenextweb.comInstagram patches vulnerability allowing account hijacks through Meta AI chatbot
Instagram restored access after hackers used its AI support chatbot to add new emails and reset passwords on multiple accounts. The company confirmed the issue was resolved Monday.
New York PostFlorida Attorney General Sues OpenAI and Sam Altman, Alleging ChatGPT Caused Harm to Users
Florida Attorney General James Uthmeier filed an 83-page civil complaint Monday in the state's 10th Judicial Circuit against OpenAI and its chief executive Sam Altman.