Anthropic expects to spend at least $518 billion over the next decade on artificial intelligence infrastructure through agreements with six partners, according to a confidential IPO prospectus reviewed by Reuters.
The planned investment ranks among the largest AI infrastructure commitments disclosed to date. Anthropic told prospective investors that about 80% of the amount is either non-cancelable or requires payment regardless of actual usage.
The AI company said the commitments reflect growing demand for computing power, which it expects will become a major constraint on the development and deployment of advanced AI systems.
Anthropic said future demand for AI could exceed available computing capacity and would be “limited principally by the availability of compute.”
Under long-term infrastructure agreements, Anthropic expects to spend at least $111.1 billion with Google, $110 billion with Amazon and $31.4 billion with Microsoft over periods ranging from seven to 10 years.
England Beat Czech Republic as Kane and Gordon Secure Nations League Win
The agreements require Anthropic to make payments regardless of how much computing capacity it ultimately uses.
Anthropic said its commitment to Google runs from April 2026 through July 2033, while its agreement with Amazon covers May 2026 through April 2036.
“If our actual spend falls short, we must pay Google the difference,” the company said in the prospectus, adding that similar terms apply to its Amazon agreement.
Its Microsoft commitment of $31.4 billion runs from November 2026 through May 2033 and is non-cancelable except in cases involving an uncured material breach by Microsoft.
Anthropic is also carrying around $161.2 billion in equipment lease obligations related to Broadcom, most of which are described as non-cancelable.
The company said the Broadcom lease arrangements can generally be terminated only in the event of a default.
Anthropic did not immediately respond to a request for comment on the confidential filing.
Additional partnerships with xAI and AMD
The prospectus also disclosed additional agreements that could significantly expand Anthropic’s computing capacity.
A partnership with Elon Musk’s xAI could result in up to $84.5 billion in spending on Nvidia-based computing capacity through 2029. Unlike some of Anthropic’s other commitments, the xAI-related arrangements are largely cancelable with 90 days’ notice.
Anthropic also disclosed a deeper relationship with Advanced Micro Devices. AMD has committed to purchasing up to $5 billion of Anthropic stock and is expected to provide AI computing capacity worth more than $20 billion.
The scale of Anthropic’s planned infrastructure spending is comparable with OpenAI’s $500 billion Stargate project, whose costs are expected to be shared by OpenAI, SoftBank, Oracle and MGX.
Anthropic shifts toward dedicated infrastructure
Anthropic said it is increasingly moving away from a cloud-only model and building dedicated AI infrastructure, including data centers and directly leased computing equipment.
The company also highlighted risks arising from its dependence on Amazon, Google and Microsoft.
Those companies simultaneously act as investors, customers, cloud providers, distributors and competitors because they are developing their own AI models.
Anthropic said those overlapping relationships could create incentives that “may not be fully aligned” with its interests.
“If the compute we have access to from third parties is curtailed, repriced, or terminated … our business, financial condition, and results of operations could be adversely affected,” the company said.
The disclosure illustrates the enormous capital requirements emerging across the AI industry as companies compete for access to advanced chips, data centers and computing capacity.
Anthropic confidentially filed for an initial public offering with the US Securities and Exchange Commission in June. The filing has not yet been publicly disclosed.
