Anthropic's IPO Filing Reveals $518 Billion AI Buildout Plan with Largely Non-Cancelable Contracts

AI developer Anthropic expects to spend a minimum of $518 billion over the next decade to build out its artificial intelligence infrastructure, according to a confidential IPO prospectus seen by Reuters. The filing reveals that the vast majority of this spending is locked into long-term, non-cancelable contracts, signaling a massive capital commitment to secure the computing power necessary for developing advanced AI models.

Binding Infrastructure Deals

Anthropic's filing details one of the largest AI infrastructure commitments on record, with approximately 80% of the total sum designated as non-cancelable or subject to payments regardless of usage. The company is securing its access to processing power, which it identifies as the primary constraint on future AI development.

The specific long-term obligations detailed in the prospectus include:

• $111.1 billion with Alphabet's Google between April 2026 and July 2033.
• $110 billion with Amazon between May 2026 and April 2036.
• $31.4 billion with Microsoft between November 2026 and May 2033.
• $161.2 billion in largely non-cancelable equipment lease obligations related to Broadcom.

For the cloud service deals, Anthropic stated it must pay the difference if its actual spending falls short of the committed amount. The company filed confidentially for an IPO with the U.S. Securities and Exchange Commission in June.

Strategic Rationale and Market Context

Ad

The AI lab justified the enormous expenditure by stating that future demand for advanced AI is expected to exceed the available supply of computing resources. This plan places Anthropic's ambitions on a scale comparable to OpenAI's reported $500 billion Stargate supercomputer project.

The filing also disclosed a deepening relationship with Advanced Micro Devices (AMD), which has committed to purchase up to $5 billion of Anthropic stock and supply over $20 billion in AI computing capacity. Additionally, Anthropic has an agreement with Elon Musk’s xAI for up to $84.5 billion in Nvidia-based capacity, though this deal is largely cancelable with 90 days' notice.

Shifting Strategy and Acknowledged Risks

Anthropic is increasingly shifting from a cloud-only model toward building its own AI infrastructure, including dedicated data centers and directly leased chips. However, the prospectus flags a key strategic risk in its heavy dependence on Amazon, Google, and Microsoft.

The company noted that these tech giants act simultaneously as investors, customers, cloud providers, and direct competitors, creating incentives that "may not be fully aligned" with Anthropic's own interests. The filing warns that if its access to third-party computing is "curtailed, repriced, or terminated," its business and financial condition could be adversely affected.