Data Center Loan Auction Tests AI Financing Confidence
A $500 million sale of debt tied to Stargate-linked sites will show whether lenders still price AI infrastructure as low-risk.
Updated on Nov. 5, 2026 in Data Centers

Live Poll
Do you support building more AI data centers in your community?
An anonymous seller is offering a roughly $500 million portfolio of data center loans in a single auction, with bids for a dozen loans due Nov. 19. The sale is a bids-wanted-in-competition process, and the development is a current market transaction rather than a product or research result.
Why it matters
The auction follows a Democratic sweep of U.S. midterm elections that raised doubts about many data center projects, and it comes as the Nasdaq 100 fell as much as 6.4% on Thursday amid investor concerns about whether AI revenue can justify spending. Taken together, the sale signals that lenders are testing how much confidence remains in financing the physical infrastructure behind the AI buildout.
The loan portfolio is a bundle of roughly $500 million across a dozen loans, with a $25 million slice backing a Stargate data center site in New Mexico. Bids must be at least $10 million, and the final pricing and buyer identities remain unreported.
The players
Oracle
Enterprise database and cloud company whose infrastructure business is a major tenant of AI data center projects, including Stargate.
Goldman Sachs Group Inc.
Global investment bank that committed to the SpaceX chip financing deal with Apollo before pricing it, and is looking to syndicate portions to other investors.
Apollo Global Management Inc.
Alternative asset manager that co-committed with Goldman Sachs to the SpaceX chip financing and is looking to syndicate portions to other investors.
CoreWeave Inc.
AI-focused cloud infrastructure provider whose junk bonds saw some of the biggest declines in the high-yield market.
Vantage Data Centers
Data center operator whose debt is included in the loan portfolio being offered for sale.
The details
Data center construction is typically financed with debt secured against the buildings and the long-term leases signed with tenants. In this sale, the seller is offering those loans through a bids-wanted-in-competition process, in which prospective buyers submit competing offers for the debt rather than negotiating one-on-one. The portfolio also includes debt tied to operators Vantage Data Centers and CyrusOne, and Oracle has sent the developer of the New Mexico site a force majeure notice, a contract clause that excuses performance when extraordinary events intervene.
Timeline
Bids on the data center loans are due Nov. 19.
Goldman Sachs and Apollo committed to the SpaceX chip financing in October before pricing it.
The Nasdaq 100 fell as much as 6.4% on Thursday.
The Tech Race
The AI infrastructure race is now being financed through layered debt, with the Stargate project and its operators at the center of the pressure. This auction shows that lenders are willing to sell exposure to the same data center assets that frontier AI companies depend on.
For most people the effect arrives through the cost and pace of AI services that depend on these facilities being completed on schedule. A stalled New Mexico site or a tightening of data center credit could delay the capacity that cloud and AI products rely on, though the extraction does not give a timeline for any such delay.
The takeaway
Watch the Nov. 19 bid deadline, since the clearing prices on these loans will show whether lenders see data center debt as a safe asset or a discounted one. The outcome will offer a concrete read on whether investor doubts about AI spending are spreading from equities into the debt markets that fund the buildout.
What happens next
Bids on the dozen loans are due Nov. 19, which will show whether the offered debt clears at or below its face value.
Further reading
For the wider context on how these facilities are built and financed, see the Data Centers section.
Source note: This article includes information reported by FA Magazine.
Live Poll
Do you support building more AI data centers in your community?









