Anthropic reports $33.2 billion in revenue and $895 billion in compute commitments

The scale of AI infrastructure

Anthropic recently filed a revised IPO prospectus with the US Securities and Exchange Commission, revealing the massive scale of its operations. The company reported $33.2 billion in revenue for the nine-month period ending September 30, 2026. This figure represents a year-over-year growth rate of 1,206% as reported by gHacks.

To support its AI models, Anthropic has entered into nearly $895 billion in commitments for compute and infrastructure. These obligations include $320 billion in existing contractual duties and roughly $451 billion in signed leases that have not yet started. The company also expects to incur about $447.5 billion in commitments by the end of 2030.

When including additional payments to secure enough compute capacity, Anthropic estimates its total tab could exceed $1 trillion. The filing notes that many of these deals, particularly with Microsoft, Google, Amazon, and Broadcom, are non-cancellable. Some of these agreements even require payment regardless of whether Anthropic actually uses the capacity.

Slowing growth and revenue mix

While the company has seen explosive growth in the past, the new filing suggests that the pace is beginning to ease. Revenue rose nearly 2.5 times between the first and second quarters of 2026, but the third quarter saw a more modest increase of 46% over the second quarter. During that same period, the costs for compute and infrastructure climbed by 77%.

Anthropic explicitly warned investors that it does not expect revenue to continue growing at its previous rates. The company stated it "cannot assure you that it will continue to grow at all" in the prospectus. This transparency highlights the difficulty of maintaining triple-digit growth as the AI market matures and competition increases.

The company's revenue streams are also becoming clearer as it prepares for public offering. Cloud marketplaces made up about 44% of the revenue for the nine-month period. Notably, the company confirmed that no single customer accounts for more than 10% of its total revenue, suggesting a diversified client base.

The risk of hyperscaler reliance

A significant portion of Anthropic's business model relies on large cloud providers, often called hyperscalers. The filing identifies this reliance as a primary risk for the company. This is because companies like Amazon, Google, and Microsoft are not just distributors; they also build competing products or host models from Anthropic's direct rivals.

This dynamic creates a complex environment where Anthropic must balance its need for infrastructure with the risk of being overshadowed by its own partners. The company must generate enough revenue to meet its substantial and growing financial obligations to secure the necessary hardware. The filing notes that the ability to meet these costs depends directly on significant revenue growth.

To fund these massive obligations, the company may consider debt or equity financing. However, the prospectus warns that relying on these methods over the long term could lead to unfavorable borrowing terms. It could also result in the dilution of existing shareholders, which is a key consideration for potential investors in an IPO.

What comes next for the company

Looking ahead, Anthropic expects many of its signed leases to begin between 2026 and 2028. Some of these agreements are still subject to specific conditions or termination rights. The company is currently navigating a transition from securing capacity to integrating that capacity into its core product offerings for users and developers.

For the average user of the Claude AI model or an API customer, the filing does not indicate any immediate changes to products or pricing. The focus remains on the backend infrastructure and the financial stability required to keep the models running. The company's primary challenge will be converting its massive infrastructure investments into sustainable, long-term profit.

As the company moves toward its public debut, the market will be watching how it manages its $1 trillion compute tab. The success of the company will depend on whether it can maintain a lead in the AI space while managing some of the largest infrastructure contracts in history. The next few years will determine if the current spending can be justified by the value provided to its customers.