Blue Owl Capital has quietly become one of the most consequential players in the AI infrastructure buildout, pouring billions into data centre projects that the biggest names in tech would rather not finance themselves. The firm owns equity stakes in special purpose vehicles backing projects including OpenAI's Stargate initiative and a Meta data centre in Louisiana, taking on the riskiest tranches of these deals.
The arrangement is straightforward in theory: AI hyperscalers like Meta get to keep massive capital expenditures off their balance sheets, and Blue Owl gets above-market returns for absorbing that risk. The company's CEO has framed this as a huge arbitrage opportunity, lending to creditworthy tech giants at returns that exceed what those borrowers' profiles would normally command.
The stock market's verdict has been less enthusiastic. Blue Owl's shares have fallen roughly 50% over the past year, a stark signal that investors are not yet persuaded the rewards justify the exposure.
The playbook borrows from real estate, but the tenants are different
Blue Owl's data centre strategy follows a structure the company refined in its real estate business. The firm takes equity positions that generate regular payments, a model designed to reduce risk through predictable cash flows rather than speculative upside.
The critical difference is the counterparty list. In traditional real estate, tenants tend to be established businesses with long operating histories. In AI infrastructure, Blue Owl is placing bets on companies like Crusoe, Oracle, and CoreWeave being around to pay their bills and complete projects that stretch over a decade. If any of those counterparties falter, Blue Owl will be left searching for someone to fill their place in facilities that may not have obvious alternative uses.
Debt markets are cooperating, for now
One pillar of Blue Owl's strategy has held firm: the debt markets. The company has successfully raised debt financing that covers 80 to 90% of project costs on top of its equity positions, and appetite among lenders for tech and AI infrastructure remains strong, with tight spreads reflecting confidence in the sector.
That confidence, though, has a conditional quality. Lenders in this space are primarily focused on capital preservation, and they are watching closely for any crack in the data center story that might justify pulling back. A single high-profile project delay or counterparty default could shift sentiment quickly.
If the debt markets tighten, the effects on Blue Owl would compound. The firm's ability to layer cheap debt on top of its equity is central to making the economics work. Any sign that borrowing costs are rising or that lenders are growing cautious could constrain Blue Owl's appetite to keep deploying equity into new projects.
More money is chasing the same deals
The competitive picture is also shifting. The wave of capital flowing into AI infrastructure has attracted a growing number of players looking to participate in large data centre deals, compressing the returns that made these investments attractive in the first place.
After extensive reporting on data centre financing, the picture that emerges is genuinely uncertain. Blue Owl may well be right that financing AI infrastructure for the world's largest tech companies at premium returns is a generational opportunity. But the firm is also absorbing risk that those same tech companies, with their enormous balance sheets and deep understanding of the AI market, have deliberately chosen not to hold themselves. The data centre companies and their financiers may need to endure some pain in the short term before the thesis plays out, and whether Blue Owl can stay the course through that period remains an open question.