Broadcom is reported to be considering a $42 billion financing arrangement to support Anthropic's custom AI chip leasing program.

Broadcom has agreed to provide up to $42 billion in financing tied to Anthropic's massive compute commitments, according to Anthropic's IPO prospectus. The money could cover approximately one-third of Anthropic's $125.2 billion five-year obligation to lease tensor-processing-unit capacity.

The arrangement effectively positions Broadcom as both a semiconductor supplier and a lender to one of its largest future customers. Anthropic is expected to become Broadcom's largest custom-chip customer by 2027.

Broadcom's custom accelerators and networking hardware are gaining strategic importance as AI labs seek alternatives to Nvidia GPUs. The company projects its AI semiconductor revenue will reach roughly $115 billion in 2027 and potentially $230 billion in 2028. This outlook reflects a broader surge in AI chip revenue, with Broadcom reporting more than $30 billion in AI bookings in its latest quarter.

However, the financing structure introduces an unusual dynamic: part of the demand supporting these forecasts may itself be enabled by Broadcom's own capital.

Anthropic disclosed approximately $518 billion in future cloud, computing, and infrastructure commitments in its IPO filing. Despite generating nearly $4.6 billion in revenue in 2025, the company spent about $7.33 billion on compute and infrastructure that year, illustrating how aggressively leading AI labs are expanding capacity before revenue fully catches up. Anthropic's broader IPO filing also showed more than $8 billion in operating losses.

Broadcom's financing could ease that buildout while locking in years of future chip demand. This structure aligns with a wider trend in AI financing, where chipmakers, hyperscalers, and private-credit investors increasingly use loans, guarantees, and leases to fund data centers.

The arrangement carries potential risks for Broadcom. If Anthropic's growth slows, financing a customer that simultaneously represents a growing share of Broadcom's AI revenue could amplify exposure. The agreement includes potential conflicts around pricing, hardware availability, and repayment obligations, according to Reuters. Nvidia faces similar questions as Wall Street evaluates whether GPUs can support large-scale chip financing.