The $500 Million Auction Testing AI Infrastructure Credit

Jeremiah Bauman |

In private credit, a sizable sale can bring a useful dose of reality to a market accustomed to working with estimates. According to Bloomberg reporting, Dutch bank ING is seeking buyers for roughly $500 million of data-center loans through a competitive bidding process, with offers due November 19. The sale could provide a revealing look at what investors are willing to pay for existing exposure to the AI buildout. Markets have a peculiar fondness for actual transactions, particularly when everyone has spent the preceding month explaining why someone else’s valuation is wrong.

The financing behind AI infrastructure has expanded rapidly as developers and technology companies race to secure computing capacity, land, and electricity. That expansion requires lenders willing to finance projects whose construction schedules, customer commitments, and operating economics can stretch across many years. Investors have become more attentive to how those pieces fit together, especially when a delay in securing power can postpone revenue while financing expenses continue arriving on schedule. Interest payments have always displayed an admirable commitment to punctuality.

The political backdrop has added another layer of uncertainty. Following Democratic gains in the midterm elections, investors sold technology and AI-related shares amid concerns about tougher oversight and potential restrictions on data-center development. Those concerns could affect assumptions about permitting, construction schedules, and project costs, although the policies ultimately adopted will determine the practical consequences. For lenders, the question becomes how much protection a loan provides if a project takes longer or costs more than originally expected. 

Against that backdrop, the reported ING offering deserves attention. The portfolio includes a dozen loans, with exposure to Vantage Data Centers, CyrusOne, and a $25 million slice tied to the New Mexico campus associated with the Stargate venture backed by OpenAI, SoftBank, and Oracle. Bids must reportedly be at least $10 million, and ING declined to comment when contacted by Reuters. These details establish the outlines of the offering, while leaving the bank’s motivation open to interpretation. 

Banks regularly originate loans with the intention of distributing some of their exposure. A sale can help manage concentration, release balance-sheet capacity, or accommodate changing internal risk limits. ING’s decision could reflect any combination of those considerations, along with its assessment of market conditions. The offering alone gives us limited grounds for deciding how concerned the bank is about the underlying projects.

What makes the process interesting is the opportunity for buyers to assign prices to a collection of data-center loans at the same time. Each bid reflects a judgment about the borrower, the contractual protections, the collateral, and the return required to accept the risk. A prospective buyer also has to consider competing investments and how much AI-related exposure already sits elsewhere in the portfolio. Enthusiasm for the technology can remain intact while the investment committee decides it has heard quite enough about another campus requiring several hundred megawatts.

The eventual prices will require careful interpretation. A loan trading below its face value could reflect a coupon that has become less attractive as market yields changed, concerns about the borrower, limited liquidity, or specific problems with the project. Prices near par could suggest continued appetite for those particular loans, depending on their terms. The useful information comes from understanding why buyers demand a given price and whether similar requirements begin appearing in other transactions.

That distinction matters when considering the cost of future financing. If investors consistently require greater compensation for construction risk, customer concentration, or uncertain power availability, developers may face more expensive borrowing and stricter loan terms. Some projects may still offer attractive economics under those conditions, while others may need more equity, stronger guarantees, or additional time. A discount in this auction would be one piece of evidence to examine alongside those broader developments.

The November 19 deadline could therefore give participants a useful reference point, provided bids lead to completed transactions. Public visibility will depend on what pricing information becomes available, and any reported results will need to be evaluated loan by loan. The process may also reveal substantial differences between projects that tend to be grouped together under the same AI label. A completed facility serving a financially strong customer presents a different set of questions from a construction site still waiting for a reliable route to electricity.

For equity investors, these credit-market developments deserve a place in the analysis. Financing terms can affect project schedules, cash flow, returns on invested capital, and the amount of new equity a company may need to issue. Those consequences can travel through the supply chain to equipment manufacturers, power providers, networking companies, and other businesses participating in the buildout. A promising order book becomes more valuable when the customer can finance the work and pay the invoice.

In managing portfolios, I want to understand that chain of obligations. Who is borrowing, who is ultimately paying, what has been contractually committed, and how much flexibility exists if conditions change? Those questions help us evaluate the durability of a theme while maintaining discipline around valuation, position sizing, and concentration. On the client side, the financial plan and anticipated cash needs help determine how much uncertainty the portfolio can reasonably carry.

ING’s reported auction offers an opportunity to learn more about a market financing an extraordinary expansion. The results could support confidence in certain borrowers, reveal greater caution toward others, or show that investors remain willing to lend when the compensation is sufficient. Each outcome would help refine the investment picture. The work continues as we compare the enthusiasm surrounding AI with the financial arrangements required to build it.

If you would like to discuss what these developments mean for your portfolio, call or email me. We can review your exposure to the AI buildout, the financial strength of the companies behind it, and how those investments fit your broader financial plan. We’ll keep watching what lenders are willing to finance and what borrowers are willing to pay, both tend to become refreshingly candid when actual money is involved. Your capital, our expertise, a bespoke creation.