Substrate
ai

Elon Musk Leases World's Largest Supercomputer to Anthropic

Elon Musk has leased the world's largest supercomputer, known as Colossus 1, to Anthropic. The move comes three months after Musk publicly described Anthropic as "evil," "misanthropic," and opposed to Western civilization on X. The lease is intended to generate revenue for SpaceX ahead of its planned public offering.

FO
fortune.com
2 sources·May 9, 5:00 PM(11 hrs ago)·2m read
|
Elon Musk Leases World's Largest Supercomputer to Anthropicbgr.in
Audio version
Tap play to generate a narrated version.
Developing·Limited corroboration so far. This page will refresh as more sources emerge.

Elon Musk leased the world's largest supercomputer to Anthropic on Wednesday, three months after he wrote on X that the company was "evil," "misanthropic," and that the AI lab hated Western civilization. The supercomputer, known as Colossus 1, contains roughly 220,000 Nvidia GPUs.

It was built in 2024 to train Grok, Musk's AI assistant. Grok has not filled the facility's capacity. According to estimates from an analyst at New Street Research, the lease will generate $3 billion to $4 billion in annual revenue for SpaceX, with more than $2.5 billion in cash profit.

The margins are high because the data center is already built and the main operating costs are electricity and staffing. "He’s not going to want multiple billions of dollars of GPUs sitting idle," the analyst told Fortune. " The announcement coincides with the dissolution of Musk's AI company xAI into SpaceX, forming SpaceXAi.

SpaceX is expected to begin its public roadshow next month after filing a confidential S-1 on April 1. The company is targeting a valuation between $1.75 trillion and $2 trillion.

The deal positions SpaceX as a provider of cloud infrastructure with a marquee AI customer. This gives the upcoming IPO a credible cloud-infrastructure business that it lacked a week ago. A former head of Google Cloud AI, now CEO of a defense AI startup, told Fortune that controlling data centers currently means controlling the application of artificial intelligence.

The executive described the arrangement as a "wedding of convenience" that may leave both sides stressed. SpaceX would capture the 30%-plus margins that frontier AI labs currently pay to providers such as AWS, Microsoft Azure or Google Cloud. Analysts told Fortune that framing SpaceX as the fourth hyperscaler could support a higher valuation multiple for the IPO, closer to those of Alphabet, Microsoft and Amazon than to Boeing or Lockheed Martin.

" The clause was not included in the formal press release and it is unclear whether it appears in the contract. Anthropic is on track to generate more than $40 billion in annualized revenue, far outpacing Grok's estimated contribution of less than $1 billion.

The lease addresses Anthropic's need for additional compute and SpaceX's excess capacity. A managing partner at Deepwater Asset Management put the odds of the deal still existing in two years at 80 percent. The executive cited Musk's history of changing his mind as the main risk factor.

The former Google Cloud AI executive said frontier AI labs typically avoid single-sourcing their compute infrastructure and will likely develop contingency plans. Both parties will continue to face pressure in the rapidly evolving sector.

Key Facts

Colossus 1 lease
to Anthropic on May 6 2026
$3-4 billion
projected annual revenue for SpaceX
220,000 Nvidia GPUs
in the supercomputer built in 2024
SpaceX S-1 filed
April 1 targeting up to $2T valuation
Reclaim compute clause
if AI harms humanity per Musk X post

Story Timeline

4 events
  1. 2026-02-09

    Musk posted on X calling Anthropic evil, misanthropic and opposed to Western civilization.

    1 source@FortuneMagazine
  2. 2026-04-01

    SpaceX filed a confidential S-1 targeting $1.75-2 trillion valuation.

    1 source@FortuneMagazine
  3. 2026-05-06

    Musk announced the lease of Colossus 1 supercomputer to Anthropic.

    1 source@FortuneMagazine
  4. 2026-05-06

    Musk dissolved xAI into SpaceX to create SpaceXAi.

    1 source@FortuneMagazine

Potential Impact

  1. 01

    SpaceX gains $3-4 billion in projected annual revenue from the lease.

  2. 02

    The lease provides SpaceX with a major AI customer ahead of its IPO roadshow.

  3. 03

    Anthropic secures access to substantial additional computing capacity.

  4. 04

    Musk gains a contractual right to reclaim compute from one of the leading AI labs.

  5. 05

    SpaceX may be valued using a hyperscaler multiple rather than a rocket company multiple.

Transparency Panel

Sources cross-referenced2
Confidence score75%
Synthesized bySubstrate AI
Word count469 words
PublishedMay 9, 2026, 5:00 PM
Bias signals removed4 across 2 outlets
Signal Breakdown
Editorializing 1Framing 1Loaded 1Speculative 1

Related Stories

IMF Reports AI Increasing Cyberattacks on Financial SystemsBenzinga
ai13 hrs agoFraming55Framing risk55/100Rewrite inherits IMF's alarmist framing on AI cyber risks and uses lede_misdirection by centering the IMF report and Barclays CEO reaction over the substantive vulnerability details.Click to jump to full framing analysis

IMF Reports AI Increasing Cyberattacks on Financial Systems

The International Monetary Fund said advanced AI models can dramatically reduce the time and cost to identify and exploit vulnerabilities, outpacing traditional defenses. Barclays CEO called a new AI tool a serious issue for banks. Emerging-market users are increasingly treating…

YA
Benzinga
CoinDesk
3 sources
NGA Director Announces New AI Framework and Launches Rapid Capabilities Officeforbes.com
ai21 hrs ago

NGA Director Announces New AI Framework and Launches Rapid Capabilities Office

Lt. Gen. Michelle Bredenkamp outlined the agency's blueprint for becoming an AI-first organization in her first major speech since taking charge in November 2025. The National Geospatial-Intelligence Agency is finalizing the framework to align with the Department of Defense AI st…

forbes.com
Variety
Breaking Defense
3 sources
**Trump Administration Considers Structured Reviews for Frontier AI Models Following Anthropic Decision on Mythos**Benzinga
ai17 hrs agoFraming55Framing risk55/100Rewrite inherits consensus framing that centers process and official reactions over the substantive AI decision, with mild valence skew favoring regulation.Click to jump to full framing analysis

**Trump Administration Considers Structured Reviews for Frontier AI Models Following Anthropic Decision on Mythos**

Anthropic developed an AI model called Mythos so capable at finding software vulnerabilities that the company decided against public release. Vice President JD Vance warned tech leaders that the technology could enable cyberattacks on critical infrastructure such as small-town ba…

The Washington Post
Benzinga
2 sources