Anthropic Drove “the Entire Acceleration in AWS and Google Cloud,” per Gil Luria. That’s a Big Risk for Amazon and Alphabet
Anthropic's stock fell 20% from its IPO debut, impacting Amazon and Alphabet's stakes. Analyst Gil Luria claims Anthropic drove recent AWS and Google Cloud growth, posing a risk to their revenue. Amazon's AWS revenue grew 37% YoY, with Anthropic committing $100B over 10 years. Alphabet's Google Cloud also benefits from Anthropic's $121.3B commitments. Amazon's free cash flow dropped to $7.7B in 2025, with $220B in expected capital spending.
How this was made
The 30-second read
Why it matters
If Anthropic’s compute spending pace slows, hyperscalers may face utilization and margin pressure from already-funded capacity, while levered data-center operators face higher credit risk.
Market read
Traders may reassess AI compute demand durability and hyperscaler cloud growth assumptions ahead of upcoming quarterly reports.
What to watch
The article relies on analyst framing and read-across; it does not show new AWS/Cloud guidance changes, nor does it quantify how much of capacity is truly incremental versus already contracted.
Background
The piece connects Anthropic’s IPO performance and long-term compute commitments to hyperscaler cloud growth, then highlights capex and financing exposure for Amazon, Alphabet, and levered data-center builders.
Ticker impact
Anthropic’s stock slid after its IPO, and the article links its compute commitments to future spending risk that could affect its partners’ demand.
Likely volatility and downside risk as investors weigh whether commitments translate into sustained compute spend.
The piece frames Anthropic’s cash coverage and spending pace as the key variable, with market attention on partner capacity and upcoming quarterly tests.
Amazon owns about 15% of Anthropic and the article argues Anthropic drove AWS acceleration, raising risk to AWS revenue growth and capacity utilization.
Likely pressure as the market prices in slower AWS demand and potential margin/capex inefficiency.
The article cites AWS growth and Anthropic’s long-term commitments, then highlights Amazon’s capex and a facility that could become stranded if Anthropic buys less.
Alphabet owns 13.7% of Anthropic, and the article claims Anthropic drove acceleration in Google Cloud, making it a direct read-across risk.
Likely downside bias as investors reassess Cloud growth durability tied to Anthropic commitments.
The piece ties Anthropic deals to Google Cloud acceleration and then flags the scenario where Anthropic buying slows while Alphabet has already funded capacity.
The article says compute-paying ability matters to CoreWeave because it borrows heavily to build data centers for AI customers like Anthropic.
Likely downside as leverage amplifies any demand slowdown risk from major AI compute customers.
CoreWeave is mentioned as exposed, but the article does not provide CoreWeave-specific new disclosures or numbers beyond the general borrowing exposure.
Market effects
Reinforces a key AI-infrastructure trade: hyperscaler cloud growth and data-center utilization can hinge on a few large AI customers’ compute commitment pace.
Primarily US mega-cap and AI-infrastructure sentiment spillover, with broader risk appetite effects for AI capex beneficiaries.
Global AI compute supply chain and data-center financing conditions may be repriced if large AI labs slow incremental spending.
Counterpoint
Anthropic’s cash coverage and long-dated commitments could mean near-term compute buying remains stable, limiting the stranded-capacity risk for AWS and Google Cloud.
Key entities
- companyAnthropic
AI lab whose IPO performance and long-term compute commitments are framed as a driver of AWS and Google Cloud acceleration.
- companyAmazon
AWS hyperscaler with a stake in Anthropic and capex exposure to compute demand.
- companyAlphabet
Google Cloud hyperscaler with a stake in Anthropic and exposure to compute-driven demand acceleration.
- companyCoreWeave
Data-center builder described as heavily borrowing to build capacity for AI compute customers.



