News

Public · Published

Broadcom Partners With Wall Street Bigwigs To Forge $63B Deal To Power Anthropic's AI Hardware

Broadcom and a group of Wall Street financial institutions announced a $63 billion financing arrangement to supply hardware to artificial intelligence startup Anthropic, using a special purpose vehicle to buy Broadcom's AI accelerators and networking gear and lease them to Anthropic, with deliveries starting in 2027.

Published:

Updated:

What happened

Broadcom and a group of Wall Street financial institutions announced a $63 billion financing arrangement to supply hardware to artificial intelligence startup Anthropic, using a special purpose vehicle to buy Broadcom's AI accelerators and networking gear and lease them to Anthropic, with deliveries starting in 2027.

Confirmed

Global impact / market context

This shows how banks and asset managers are helping fund the expensive hardware behind artificial intelligence, letting a chipmaker lock in a huge order while spreading risk across lenders. For investors, it ties Broadcom's future revenue to Anthropic's ability to pay lease and debt obligations.

Analyst inference

The package uses layered debt, where first-lien lenders like Bank of America, Citi and Morgan Stanley get first claim on assets in a default, while Blackstone holds the riskier second-lien portion. That structure points to strong lender appetite for AI infrastructure, but also to concentrated exposure if Anthropic's growth slows.

Analyst inference

What to watch

  1. Deliveries of Broadcom's specialized XPU AI accelerators and high-speed networking solutions to Anthropic are slated to begin in calendar year 2027, so watch for confirmation that the schedule holds. Confirmed
  2. Watch Anthropic's ability to generate enough cash to cover lease payments and debt service; any shortfall could pressure the special purpose vehicle and its lenders. Analyst inference
  3. Watch whether other AI hardware deals copy this layered-debt, special purpose vehicle model, which would signal lenders are willing to take on more AI infrastructure risk. Analyst inference

Evidence