Over the past 11 months, Anthropic has aggressively inked computing deals totaling a staggering $517 billion. The immediate catalyst behind this procurement frenzy is the surging demand triggered by the explosive growth of its Claude Code and Cowork products this year.

On September 6, based on a comprehensive analysis of public statements and earlier reports, since October last year, Anthropic has signed agreements securing at least 14.8 gigawatts of computing capacity, accessible over the coming years. Combined with the previously secured 1 to 2 gigawatts, the total scale has expanded further.

This figure far exceeds the target Anthropic disclosed to investors in December 2025, when the company stated it would spend $180 billion on server rentals by 2029. Based on currently announced agreements, total expenditures over the next decade or so could reach as high as $517 billion, encompassing cloud capacity leases, chip purchases, and data center rental agreements.

It is worth noting that Anthropic may hold additional undisclosed agreements, and the capacity under signed contracts may not be fully utilized.

Moreover, despite Anthropic's annualized revenue surpassing $65 billion, exceeding OpenAI's $40 billion, its computing reserves still lag behind OpenAI's planned target of 30 gigawatts.

Core Suppliers: Amazon and Google Dominate the Landscape

Anthropic's early partners and investors, Amazon and Google, remain its largest computing suppliers, providing a combined 11 gigawatts of capacity, with costs exceeding $300 billion over the next decade or so.

Over the past year, Anthropic has also extended its reach to partners previously closer to OpenAI:

Microsoft: Last November, Anthropic announced a deal with OpenAI's long-time backer, committing to pay at least $30 billion to lease roughly 1 gigawatt of Microsoft Azure servers (powered by Nvidia chips).

SpaceX: This May, Anthropic announced an agreement with Elon Musk's SpaceX, paying $1.25 billion monthly to rent AI servers, with the contract extending until May 2029, carrying a total potential cost of up to $45 billion.

Lambda and Nscale: Just last month, Anthropic signed $80 billion cloud contracts with startup Lambda and UK-based Nscale, locking in at least 460 megawatts of capacity for six years.

Building Own Data Centers: Betting on Long-Term Control

Beyond leasing, Anthropic is also accelerating its own data center deployment.

Last November, Anthropic announced plans to co-invest $50 billion with cloud startup Fluidstack to build data centers in Texas and New York, with some facilities already operational this year.

Additionally, Anthropic signed a lease with former Bitcoin miner TeraWulf, which is developing an AI data center in Kentucky that will ultimately provide 401 megawatts of computing power.

On the chip front, Anthropic signed an agreement with Google in April to procure multi-gigawatt-scale tensor processing units (designed by Broadcom), and in July added a 2-gigawatt order for AMD chips.

Gap with OpenAI: Revenue Leads, Computing Lags

Although Anthropic has overtaken OpenAI in revenue—according to Bloomberg, its annualized revenue has surpassed $65 billion, exceeding OpenAI's over $40 billion as of July—the gap in computing scale remains significant.

OpenAI told investors in April it plans to have 30 gigawatts of computing capacity by 2030. Sources indicate OpenAI expects to spend around $750 billion on computing before 2030, up from its earlier forecast of $665 billion.

OpenAI has signed agreements with Oracle, Microsoft, CoreWeave, and others. OpenAI executives say these commitments have established a competitive advantage over rivals.

In contrast, Anthropic's total announced computing plans appear to remain well below OpenAI's target level.

The Key Variable Before an IPO

Whether computing capacity keeps pace with growth directly impacts Anthropic's profitability and customer retention.

Anthropic and its CFO, Krishna Rao, have prioritized computing agreements over the past year. With its IPO prospectus expected to become public within the coming weeks, investors are most eager to see the latest disclosures on computing capacity and spending.

However, these agreements are not set in stone. Data center construction faces multiple hurdles, including site selection, grid access, government approvals, local opposition, and component shortages.

Contract terms also carry uncertainty. For instance, in the SpaceX and Anthropic deal, either party can cancel with 90 days' notice.