Broadcom Could Finance $60B of AI Spending, Anthropic Expands Compute

Broadcom is moving deeper into AI financing, with as much as $60 billion in funding being assembled for customers including Anthropic.

The structure reportedly includes about $42 billion of senior secured financing and another $18 billion of junior debt, with Blackstone involved in the latter. The financing has not yet been formally announced.

<iframe src=”https://widgets.coincodex.com/w/47bbfa32-2b94-4410-8e53-8d82293cfe12?site=coinpaper&mode=light” width=”100%” height=”420” frameborder=”0” referrerpolicy=”no-referrer-when-downgrade” style=”border:0;background:transparent;border-radius:0px;”></iframe>

A separate Anthropic filing also showed Broadcom agreeing to provide up to $42 billion in loans to help fund infrastructure spending tied to the AI company’s enormous compute commitments.

That makes Broadcom more than just a chip supplier. It is increasingly helping customers finance the infrastructure required to buy and deploy those chips.

Broadcom Is Helping Fund Its Own Demand

Anthropic is expected to become a major customer for Broadcom’s custom AI accelerators, while its infrastructure plans already include access to next-generation Google TPU systems built with Broadcom.

Financing those deployments could make it easier for Anthropic to expand faster while locking in years of future hardware demand for Broadcom.

The structure fits a much broader shift in AI infrastructure financing through debt, private credit and off-balance-sheet vehicles.

Chipmakers Are Becoming Financiers

That creates an uncomfortable question for investors: how much AI demand comes from customers with independently funded budgets, and how much is being enabled by the companies selling the hardware?

It is the same issue behind growing scrutiny of Nvidia’s push to mobilize hundreds of billions for AI infrastructure.

Big Tech is also using increasingly complex financing structures to support data-center expansion, including guarantees, leases and special-purpose vehicles. Some of that exposure sits outside traditional balance sheets, a trend explored in the rise of hidden AI infrastructure liabilities.