Anthropic locks in $518B in AI infrastructure deals over the next decade
What's the deal? Anthropic has outlined at least $518 billion in long-term infrastructure commitments over the next decade, covering cloud computing, equipment and services. The agreements span major technology and semiconductor companies, underscoring the computing capacity now needed to build and run advanced AI systems.
The breakdown: Broadcom accounts for about $161.2 billion in equipment lease obligations, followed by GoogleDealroom has a profile for this one. Try Dealroom → at $111.1 billion and Amazon at $110 billion in cloud and computing services. Smaller commitments include xAI at up to $84.5 billion, Microsoft at $31.4 billion, and AMDDealroom has a profile for this one. Try Dealroom → at more than $20 billion.
AMDDealroom has a profile for this one. Try Dealroom → also plans to buy up to $5 billion of Anthropic stock, deepening the tie between the two companies.
Why now? Anthropic has said compute availability is becoming a key constraint as demand for advanced AI grows and could eventually exceed supply. Long-term agreements give it more predictable access to the resources needed to train and operate its models.
What's the endgame? The company is also expanding beyond standard cloud capacity. In November 2025 it announced US$32.4B investment in US AI infrastructure with Fluidstack, covering data-centre development in Texas and New York, alongside an $11.6 billion, seven-year agreement with Akamai for computing workloads.
What could go wrong? Around 80% of the commitments are non-cancelable or require payments regardless of actual usage. The xAI arrangement is an exception, with most of it cancelable on 90 days' notice.
The scale raises a core question: whether growth in AI demand can generate enough commercial value to support obligations of this size.
The signal: The commitments reflect a broader expansion of the AI infrastructure market, where demand is rising for data centres, advanced processors, networking equipment, electricity and cooling. AI firms are increasingly locking in capacity in advance to hedge against shortages — a bet that future workloads will justify the spending.
Read more: sphericalinsights.com
Image credit: RyanDonegan
