The Claude maker's IPO prospectus warns that local bans on new data center construction could slow its expansion
Anthropic wants to go public. Its prospectus also lists, at length, the ways the plan could go sideways.
The AI company behind the Claude models submitted a Form S-1 registration statement to the SEC on October 26, 2026, a step that signals preparation for a potential IPO. One of the risks it flags most clearly involves zoning boards, not rival chatbots.
A 400-page prospectus with a long list of worries
The filing runs roughly 400 pages. More than 90 of those pages cover risk factors.
Among the risks, Anthropic singles out moratoriums on new data center construction in various US jurisdictions. It names them as a specific threat to its growth.
The company also points to local opposition. That resistance is tied to public health concerns, environmental impacts, and questions about how energy gets allocated.
The filing warns that community pushback, regulatory delays, and environmental scrutiny could halt development projects. That exposure is sharper because Anthropic’s infrastructure remains primarily based in the US.
The numbers behind the risk
Anthropic carries at least $417 billion in minimum payment obligations for compute and data center capacity, a figure pegged to the end of 2025. That is a contractual commitment, not a wish list.
When additional cloud commitments and direct leasing obligations are included, the total climbs further. The company’s commitments extend beyond $542 billion and reach upwards of $808 billion.
On the income side, Anthropic reported first-half 2026 revenue of $16.3 billion, a significant increase from the prior year.
The same period brought net losses totaling $80.7 billion.
Anthropic has committed to paying for capacity on a massive scale, and its ability to turn that capacity into revenue depends partly on whether facilities actually get built and powered.
A hybrid strategy to reduce dependence
Anthropic has historically leaned on the major cloud providers. Its key partners include Amazon Web Services, Google Cloud under Alphabet, and Microsoft Azure.
The filing describes a strategic shift toward a hybrid model. Anthropic is leasing dedicated data centers under the brand Anthropic Direct.
The goal is to reduce reliance on the hyperscalers. It also helps the company secure the specialized hardware that frontier model training and deployment require.
Anthropic is also pursuing international expansion. The aim is to cut the risk that comes from concentrating operations in the US.
