Anthropic reveals $895 billion in future AI infrastructure commitments

Artificial intelligence company Anthropic has disclosed around US$895.4 billion in future commitments for computing power and related infrastructure, revealing the scale of investment required to support the expansion of its AI services.

The figure was disclosed in an updated initial public offering (IPO) filing, which includes purchase commitments and lease agreements for computing equipment and infrastructure.

According to the filing, approximately US$447.5 billion is expected to be incurred by the end of 2030, with around US$69.8 billion due in the 12 months following September 2026.

Broadcom accounts for the largest individual commitment, with approximately US$161.2 billion in equipment lease obligations. Anthropic has also committed at least US$121.3 billion to Google and at least US$110 billion to Amazon for computing capacity and infrastructure.

A separate agreement with SpaceX’s AI infrastructure business could involve up to US$120.5 billion in spending through 2029. Anthropic has also outlined potential additional spending of between US$125 billion and US$150 billion on Google’s Tensor Processing Units, although this estimate is separate from the disclosed US$895.4 billion total.

The commitments reflect the growing demand for computing resources as AI companies develop and operate increasingly powerful models. Access to specialised chips, data centres and electricity has become a major consideration for companies competing in the sector.

However, the headline figure does not represent money already spent or a single unconditional payment obligation. It combines different types of purchase commitments and leases, some of which are subject to conditions or can be terminated. Around US$451.2 billion relates to leases that have been signed but have not yet started.

The scale of the commitments also raises questions about the long-term cost of expanding AI infrastructure and how companies will generate enough revenue to support that investment.

Anthropic’s disclosure puts its future infrastructure obligations among the largest reported by major technology companies. The filing places its total close to Meta’s reported US$866 billion in commitments, while Alphabet’s figure stands at around US$1.06 trillion.

The figures highlight the financial demands of the AI race, as companies secure computing capacity years ahead of expected demand. Whether that investment delivers returns will depend on the growth of AI adoption, the cost of operating these systems and the ability of businesses to turn demand into sustainable revenue.