AI Financing Gets Creative

July 26, 2026 - 21:16

Dealmakers are scouring both public and private markets for new ways to fund the massive artificial intelligence infrastructure buildout. As demand for capital rapidly outstrips traditional data center financing methods, the strategies are getting increasingly inventive.

John Greenwood, Goldman Sachs' global head of infrastructure and real asset investing, noted that the scale of required investment is forcing a rethink. The typical model of a single company borrowing against a single facility is no longer sufficient. Instead, firms are bundling multiple projects, tapping into asset-backed securities, and structuring joint ventures that blend debt with equity from pension funds and sovereign wealth funds.

One emerging approach involves "data center REITs" that package income streams from multiple facilities into publicly traded securities. Another trend is the rise of pre-leasing agreements, where major cloud providers commit to long-term capacity before construction even begins, giving lenders the confidence to provide financing.

Greenwood also pointed to the growing role of private credit funds, which are stepping in where traditional banks hesitate due to regulatory constraints. These funds can offer flexible terms but at higher interest rates, reflecting the risk of a rapidly evolving technology sector.

The creative financing push comes as estimates for global AI infrastructure spending over the next decade range from one trillion to two trillion dollars. While the demand is clear, the challenge remains matching that need with sustainable, long-term capital structures.