How the global AI boom is driven by thousands of newcomers, diluting Big Tech's infrastructure dominance; US data center credit deals hit $178.5B in 2025 so far
In the golden hills of Puglia, a southern region of Italy known today for its olive trees, white cliffs and turquoise mediterranean coves …
Context & Ripple Effects
The AI infrastructure buildout had already expanded beyond a narrow set of hyperscalers: prior coverage put U.S. capacity that was built, underway, planned or stalled above 80GW and highlighted return-on-investment constraints in the rapid data-center expansion. This report adds financing evidence that the buildout is drawing a broader set of borrowers and capital providers.
It also arrives as large platforms confront a power bottleneck for their planned capacity additions, documented in coverage of Big Tech's growing power requirements. The significance is less that infrastructure leadership disappears than that access to financed capacity is becoming less exclusive.
First-order effects
- New AI entrants and independent infrastructure developers gain more routes to fund data-center projects, reducing their reliance on Big Tech-owned facilities.
- The $178.5B in U.S. credit deals makes lenders and private-credit providers more consequential counterparties in deciding which AI infrastructure projects advance.
Second-order effects
- Big Tech faces greater competition for power, sites, construction capacity and financing terms as more well-capitalized projects seek the same inputs.
- Credit underwriting will put greater weight on contracted demand, utilization and asset life, intensifying scrutiny already raised by ROI concerns around the buildout.
Third-order effects
- If diversified financing persists, AI infrastructure could evolve from a hyperscaler-led capex race into a more financialized, multi-owner market in which developers and lenders hold greater leverage.
- That shift also raises the stakes of a mismatch between capacity commitments and AI revenue: leverage can accelerate deployment, but may expose weaker projects when demand or power delivery falls short.
The trend: AI computing is shifting toward a broader, credit-funded infrastructure ecosystem, even as power availability and project economics constrain how quickly it can scale.