AI · Oct 19, 2026
Microsoft ships Guardian Agents, saying frontier models ‘can’t be trusted’Anthropic’s leaked S-1 shows a lab racing ahead on clouds owned by its rivals
· Updated · Signal Desk newsroom
Analysis of reporting by TechCrunch. How we work

A leaked Anthropic S-1, reported by The Wall Street Journal, says the company struggles to secure capacity from AWS, Google and Microsoft, which increasingly see it as a competitor, and could spend more than $750 billion on cloud contracts. Revenue is surging; losses, concentration and liability risks are large.
Key takeaways
- The filing says AWS, Google and Microsoft increasingly treat Anthropic as a competitor, complicating capacity.
- Anthropic says it could spend more than $750 billion on cloud contracts, some still under negotiation.
- It reported $16.3 billion in first-half revenue against an $80.7 billion net loss, with two customers about a fifth of sales.
- It told investors third-quarter revenue reached $16.8 billion, up 45% from the prior quarter.
- The filing warns damages from harmful AI behavior could exceed its insurance coverage.
What the filing says
According to an S-1 filing leaked to The Wall Street Journal and summarized by TechCrunch, Anthropic is struggling to secure enough cloud capacity from AWS, Google and Microsoft, which increasingly see it as a competitor rather than a customer. The company says it could spend more than $750 billion on cloud contracts, some still under negotiation, and is leasing its own data centers and working with third-party operators it worries have limited operating histories. An S-1 is the registration statement a company files with the SEC before going public; Reuters and the Financial Times reported the filing’s broad numbers in late September, and Anthropic is reportedly aiming for a November IPO. Anthropic did not immediately respond to TechCrunch’s request for comment.
Anthropic reported $16.3 billion in revenue in the first half against an $80.7 billion net loss, per TechCrunch’s summary, with two customers making up about a fifth of sales. An earlier Verge read of a prospectus preview cited a $42 billion net loss and $518 billion in compute plans (Anthropic’s catastrophic-risk filing), so reported figures vary by document and period. Anthropic has told investors third-quarter revenue hit $16.8 billion, up 45 percent from the prior quarter.
When the landlord is a competitor
Every frontier lab runs on rented compute from companies that also sell their own models and assistants. For years the arrangement suited both sides; the filing describes it fraying. When a supplier is also a competitor, capacity can become a negotiating lever, and growth can be capped by someone else’s priorities. For enterprise Claude customers, constraints at the infrastructure layer could translate into rate limits or reliability issues.
The revenue trajectory is striking: 45 percent quarter-over-quarter growth at this scale indicates rapid customer adoption. But revenue concentration combined with a very large cost base leaves the business exposed to a few relationships on both the buying and supplying sides. The workaround, leasing its own facilities and relying on younger data center operators, carries risks Anthropic acknowledges; outages or construction delays would hit a product embedded in other companies’ workflows.
Liability on the record
The filing also warns that damages from harmful AI behavior could exceed Anthropic’s insurance coverage, an unusually candid disclosure that follows earlier risk-factor language about existential harm (extinction warning in IPO pitch). Public investors will now price catastrophic-risk disclosures alongside growth, which could force clearer accounting of safety risk or create pressure to soften it once quarterly expectations set in. Anthropic’s governance structure, with external trustees choosing most of the board (trustees in spotlight), adds another variable. More in our Anthropic coverage, business coverage and funding coverage.
Who feels it
- Prospective investors
- Rapid growth paired with huge losses, concentration and cloud dependence.
- Enterprise Claude customers
- Capacity limits could mean rate limits or reliability issues.
- AWS, Google and Microsoft
- Their supplier-competitor role is now explicit.
- Data center operators
- A large demand signal, with doubts about newer operators.
What to watch
- Whether the public S-1 changes cloud and liability language.
- New compute deals with AWS, Google or Microsoft before the IPO.
- Anthropic’s own data center lease announcements.
- Whether the November IPO timing holds.
FAQ
Why is Anthropic struggling to get cloud capacity?
Its leaked S-1 says AWS, Google and Microsoft increasingly see it as a competitor.
How much could Anthropic spend on cloud?
More than $750 billion on cloud contracts, according to the filing.
What is Anthropic’s revenue?
$16.3 billion in the first half and $16.8 billion in the third quarter, per reports.
When is Anthropic’s IPO?
It is reportedly aiming for November.
Companies: Anthropic
More on Anthropic
AI · Oct 13, 2026
After months off the frontier, Google ships Gemini 4 Argon with benchmark claims and aggressive pricingAI · Sep 30, 2026
The four rules in the White House AI commitment, and the missing fifth: penaltiesAI · Oct 20, 2026
Anthropic’s draft federal AI bill would let a judge halt a model release over mass-casualty riskAI · Oct 8, 2026
Anthropic’s new usage policy sharpens election rules, loosens voter targeting and bars sustained cruelty toward Claude