The AI buildout is increasingly funded by debt, and private credit is playing a much bigger role than it did even a year ago.
In the second quarter of 2026, Apollo and Blackstone led a $35 billion vehicle to buy chips and lease them to Anthropic. UBS called it the largest private credit deal on record. For comparison, direct lenders made about $23 billion of loans to private-equity-backed U.S. companies that quarter. One AI financing was larger than all the traditional PE-backed direct lending PitchBook counted during the same period.
And the spending that needs financing is still growing. S&P Global forecasts Alphabet, Amazon, Meta, Microsoft, Oracle, and SpaceX (xAI) will spend about $870 billion this year, and more than $1.3 trillion in 2027, with debt, leases, new shares, and other financing helping fill the gap.
This is where private credit comes in. It is lending negotiated directly between borrowers and nonbank lenders, and it grew rapidly after 2008 as banks pulled back from parts of the corporate lending market. Direct loans to U.S. borrowers grew roughly 25x, from $96 billion in 2010 to $2.4 trillion in 2025. Over the same period, lending to U.S. software and technology companies grew roughly 45x, from $22 billion to more than $1 trillion.
Private credit usually costs more than traditional bank or syndicated loans, but borrowers pay that premium for speed, certainty, confidentiality, and financing they might not get elsewhere.
One thing that stood out in our research was how little extra lenders appear to charge for AI risk. In January, the BIS found that private loans to AI-related companies carried an average spread of 6.2 percentage points above the base rate, compared with 6.1 points for other private credit loans. Meanwhile, equity markets reflect much higher expectations for AI-related growth. The BIS said the small difference in loan pricing could mean lenders are underestimating AI-related risks, equity investors are overestimating future profits, or both.
Private credit is exposed not just to the companies building AI, but also to the software companies it could disrupt. If AI spreads quickly, it could undermine existing software borrowers by taking away customers and revenue. On the other hand, if AI takes longer to pay off, AI borrowers may face revenue shortfalls when their loans come due.
The structure of some AI financings adds another layer to the picture. CoreWeave borrowed for five years against customer contracts averaging about three, while Meta’s Hyperion bonds run to 2049 against leases renewed every four years.
The market is also already distinguishing among these deals. Apollo bought debt financing chips leased to xAI at 99 cents in December 2025, and after SpaceX bought xAI, it traded as high as 106. Project Jupiter, an $18 billion AI data-center financing, later traded at 89 to 91 cents as investors reassessed the project's credit.
So rather than ask whether AI itself is a bubble, our team at Social Capital looked beneath the surface at a different question. Is private credit, which is funding a growing portion of the AI buildout, a bubble?
In our Deep Dive on Private Credit, we explore this fundamental question. Here's a look at what you will learn inside:
- In October 2025, JPMorgan's CEO warned, "When you see one cockroach, there are probably more." This case study follows a chain of bankruptcies linked to private debt. How likely is it that history will repeat itself?
- The six key factors that separate the newer branch of private credit from the traditional model that emerged after 2008, and how hyperscalers leverage the new way.
- Our refinancing map shows what different types of refinancing can reveal about borrower strength and stress.
- In early 2026, major private credit funds paid out more in withdrawals than they raised from new investors for the first time. What happens if that pressure continues?
- Two AI financings depended on tenant payments. One rose above face value while the other fell toward 90 cents. Why?
Hope you enjoy reading! Link below.
