Sources: Broadcom CEO Hock Tan held 11th-hour talks and agreed to higher rates and loan guarantees to secure bank participation in the $63B Anthropic debt deal
The deal, announced on Monday, is the biggest ever to raise money for AI infrastructure — Hock Tan was running out of patience.
Context & Ripple Effects
Broadcom's effort to assemble more than $60 billion of AI-chip financing had been in motion since August, when it was approaching a lender group for the debt. By early October, a Wall Street syndicate was still being assembled, indicating that lining up capital—not just identifying chip demand—was central to the transaction.
The financing vehicle was announced with Bank of America, Blackstone, Citigroup and Morgan Stanley on October 26, structured to fund Anthropic chip purchases. The final lender concessions clarify the price and protections required to move the $63 billion special-purpose vehicle from arrangement to participation.
First-order effects
- Participating banks receive higher rates and loan guarantees, improving the terms under which they join the $63 billion financing vehicle.
- Anthropic gains a funded path to buy chips through the special-purpose vehicle, while Broadcom secures lender participation needed to support the transaction.
Second-order effects
- The revised terms establish that lenders to large AI-compute transactions may demand both higher compensation and bespoke credit protection rather than treating chip demand alone as sufficient collateral.
- Broadcom's financing structure shifts more of the immediate question around Anthropic's infrastructure expansion from access to chips to the bankability of the vehicle purchasing them.
Third-order effects
- If similar structures proliferate, AI infrastructure spending will be increasingly shaped by the credit terms, guarantees and syndication appetite available to chip suppliers and their customers.
- The deal points toward a compute market in which major suppliers help arrange capital alongside hardware, tying infrastructure growth more closely to private-credit and bank risk tolerance.
The trend: AI compute build-outs are becoming finance-led projects, with lender protections and structured vehicles determining how quickly customer chip demand converts into deployed infrastructure.