The Rise of Private Credit and Direct Lending

The most consequential shift in how AI capacity gets funded is the rise of private credit. Over the past few years, direct-lending funds, money raised from pensions, insurers, and sovereign wealth and lent straight to borrowers without going through a bank, have become a dominant source of data-center debt. Where a syndicated bank loan or a public bond once would have funded a large build, increasingly a private credit fund writes the whole check. This is not a niche. It is a structural reallocation of who supplies the debt behind the AI build-out, and it is happening because private credit fits the problem better than the alternatives.

Three features make private credit a natural match for hyperscale construction. First, speed and certainty: a single fund or a small club can commit billions in weeks, with one counterparty to negotiate, rather than syndicating across dozens of banks over months. For a developer racing to energize capacity, certainty of execution is worth paying for. Second, flexibility: private lenders write bespoke terms, blend construction and permanent debt into one unitranche facility, size to a specific asset's cash-flow profile, and structure around the awkward realities of a build that banks' standardized boxes cannot accommodate. Third, capacity: banks face regulatory capital rules that limit how much long-dated construction risk they can hold, and public bond markets want stabilized assets, so private credit fills the gap for large, complex, pre-stabilization financing precisely where the other channels retreat.

Behind the shift is a wall of money that needs to be deployed. Insurers and pension funds have long-dated liabilities and are searching for long-dated, yielding assets to match them. Data-center debt, secured by real property and contracted hyperscaler rent, offers exactly that: a durable, asset-backed yield with a hard-asset floor. Private credit managers raise enormous funds from these investors and then compete to place that capital into infrastructure, and AI data centers have become one of the most sought-after destinations. The result is a self-reinforcing loop: institutional demand for yield pulls capital into private credit, private credit pushes it into data-center debt, and cheap, available debt accelerates the build-out.

The scale has grown large enough that private credit is now shaping the physical map. When a fund decides which developers to back and which markets to finance, it is effectively deciding where capacity gets built. A developer with a committed private-credit partner can move faster and bid harder than one still assembling a bank syndicate, so the flow of private capital increasingly determines who wins land and power. Some of the largest announced financings in the sector are private-credit facilities measured in the tens of billions, structured to fund entire campuses or portfolios in a single relationship.

The trade-off the borrower accepts is cost and control. Private credit is more expensive than a bank loan or a public bond would be for the same risk; the borrower pays a premium for speed, certainty, and flexibility. It also comes with tighter, more actively monitored covenants and a lender that sits close to the business and can intervene quickly if performance slips. For a strong developer racing the clock, that trade is often worth it. For a weaker one, it can be the beginning of a loss of control if the asset underperforms and the covenants bite.

The systemic question this raises is where the risk has gone. It has moved off bank balance sheets, where regulators watch closely, and into private funds held by insurers and pensions, where it is less transparent and less liquid. In good times this is efficient: capital flows to where it is needed without the friction of the banking system. The concern is what happens under stress, when many highly-geared data-center deals, funded by private credit and dependent on continued hyperscale demand, all come under pressure at once. The debt is patient and locked up, which is stabilizing, but it is also concentrated in a young asset class whose long-run demand is still being tested. Private credit has unquestionably accelerated the AI build-out. Whether it has also quietly concentrated risk in the balance sheets of long-term savers is the open question that prudent observers keep in view.