Apollo's $35 Billion AI Chip Private Debt Deal Begins Trading
0xBroomberg
Apollo and Blackstone's $35 billion AI chip financing package — the largest private credit deal ever — is set to begin drawdowns and shift into tradeable markets, a pivotal step from lockup to liquidity.
What is this deal actually doing?
A special-purpose vehicle (SPV — a shell entity created for a single transaction) is borrowing $35 billion to buy custom AI chips co-developed by Google and Broadcom.
The chips are then leased to AI company Anthropic. Broadcom backstops Anthropic's payments on the most senior tranche of debt.
In plain terms = this is not a conventional loan. It is a "borrow money, buy chips, lease them out for rent" structure — the chips themselves serve as collateral.
How does the money arrive, and when can it trade?
The debt uses a delayed-draw structure: the borrower pulls funds in roughly 16 installments as chips are delivered, spread over more than a year.
By early next year, about $15 billion is expected to be drawn and tradeable. By summer 2027, that figure should reach roughly $24 billion.
This means → the market will not absorb a $35 billion shock all at once. Staggered drawdowns give buyers and sellers a pricing window.
From private placement to 144a — how does liquidity open up?
The debt was initially issued in the private placement market, where the buyer pool is small and liquidity is thin.
Once drawn, tranches will migrate to the 144a market — a secondary market where qualified institutional investors such as insurers and mutual funds can trade debt securities, significantly boosting liquidity.
In plain terms = at first only a handful of large buyers could touch this debt. Once it moves to 144a, far more asset managers and dealers can step in — think of it as going from a VIP-only room to an open trading counter.
What does this deal signal for AI financing?
Tech companies are flooding both public and private credit markets to raise the hundreds of billions of dollars needed for data-center construction and chip procurement.
This reflects a structural shift in AI infrastructure financing — away from equity-only or bank-loan models, toward securitizing chip assets and tapping credit markets to fund compute.
As more AI chip debt enters the 144a market, pricing and liquidity for this asset class face a real-world test — what buyers are willing to pay will determine whether this financing pathway scales.
Content is for reference only, not financial advice.