Broadcom Backs $63 Billion Debt Package for Anthropic AI Chips

Four Wall Street names and one alternative asset manager have lined up $63 billion to finance Broadcom's AI chips and networking gear for Anthropic, in one of the largest debt deals yet tied to the AI buildout. Bank of America, Blackstone, Citi and Morgan Stanley will provide the financing, according to a Reuters report dated October 26. The money buys hardware; the structure buys comfort.

Broadcom is supplying "residual value support" to the lenders, who are extending first-lien loans and senior debt. That phrase amounts to a significant guarantee. Chipmakers have offered such guarantees before, absorbing potential declines in the value of assets such as advanced chips and data infrastructure. Broadcom did not respond to a request for comment.

How is the $63 billion split?

The first-lien and senior debt portion totals roughly $43.5 billion, or 69% of the financing, according to two sources familiar with the matter. Bank of America, Citigroup and Morgan Stanley are arranging that tranche. Blackstone provides the committed portion of the second-lien financing, which ranks below senior debt and carries the corresponding risk. The companies did not disclose the full loan breakdown.

The transaction runs through a newly formed special purpose vehicle. That vehicle will purchase Broadcom's AI accelerators and networking equipment and lease them to Anthropic, the companies said on Monday. Deliveries begin in 2027, under a previously announced strategic compute capacity agreement involving Broadcom.

What does the deal say about AI finance?

It confirms that banks and alternative asset managers now occupy a central role in financing AI computing infrastructure, as technology companies race to secure capacity for increasingly complex workloads. Anthropic needs the compute; Broadcom needs buyers; the lenders need someone to stand behind the collateral. Broadcom stands behind it.

The structure is not without precedent in this saga. Apollo and Blackstone previously agreed to finance a $35 billion expansion of Anthropic's AI computing capacity under an earlier agreement with Broadcom. Broadcom's backstop in that deal reached up to $29 billion, according to a filing.

Moody's offered a qualified endorsement. The rating agency described Broadcom as one of the strongest large technology companies, but warned that risk would increase if multiple transactions were undertaken over a short period to meet the sizable funding needs of frontier AI labs. One guarantee is manageable. Several, stacked, are a different question.

What hangs over the timing?

Anthropic filed for a potentially record-breaking IPO on Monday. The filing arrives as disclosures of steep losses and security concerns threaten to test investor appetite for the AI heavyweight. The chip financing and the public offering are separate instruments serving the same purpose: keeping the compute pipeline funded.

The arithmetic is straightforward. A frontier lab with steep losses needs tens of billions in equipment; lenders will fund it only while a chipmaker absorbs the downside. Broadcom has now done that twice — first with a backstop of up to $29 billion, now with residual value support across a $63 billion package.

The next test will come if further labs demand similar structures. Moody's has already named the condition: risk rises with frequency. How many of these guarantees Broadcom chooses to write may determine whether AI infrastructure finance remains a bank business or becomes a chipmaker liability.