A day after filing for its IPO, Anthropic is getting another boost for its computing infrastructure. Chipmaker Broadcom announced a $63 billion financing deal to fund AI chips and networking equipment to be supplied to Anthropic, the developer of Claude. According to reports by Bloomberg and Reuters, Bank of America, Citigroup, Morgan Stanley and investment firm Blackstone are taking part.
How the deal works
Instead of Anthropic buying the chips itself, a special purpose vehicle (SPV) has been set up to buy AI accelerators and networking equipment from Broadcom and lease them to Anthropic. The banks are arranging the senior debt, and Blackstone is providing the committed portion of the junior debt. According to Reuters sources, the senior debt comes to about $43.5 billion, roughly 69% of the deal.
Broadcom itself is taking on part of the risk: it has agreed to absorb a possible decline in the value of the equipment, up to the amount of the senior debt. It is a kind of safety net for the lenders, in case the chips lose value faster than expected. Deliveries are expected to begin in 2027 under a previously announced compute capacity agreement.
New financing for the AI era
The deal illustrates what funding AI labs looks like today: not just equity raises, but complex financial structures in which chipmakers, banks and investment funds share the risk. It is not the first: according to reports, a previous $35 billion expansion was financed by Apollo and Blackstone, with a Broadcom backstop of up to $29 billion.
Rating agency Moody's called Broadcom one of the strongest large technology companies, but warned that risk would rise if similar deals piled up over a short period to meet the AI labs' enormous funding needs. In other words: as long as demand for AI keeps climbing, this structure works. If it slows, whoever guaranteed the value of the chips may discover how fast hardware becomes obsolete.
