Nvidia is investing about $1.5 billion in SB Energy to secure up to 8 gigawatts of AI computing capacity at an Ohio campus designed for OpenAI participation.

SB Energy, backed by SoftBank, is developing large-scale power and data center campuses to support AI workloads, with OpenAI involved as a tenant.

The broader plan envisions at least 10 gigawatts of new power generation and roughly $4.2 billion in Ohio grid upgrades to back AI data centers.

Observers note possible market implications, including DePIN-like models and the need to watch high-yield credit spreads and AI enterprise adoption timelines.

Nvidia’s financing approach mirrors a strategy of funding ecosystem partners, a move that has drawn scrutiny over fund flows.

Nvidia trimmed its debt/credit guarantee exposure from about $250 billion to under $120 billion, backing the initial 5 gigawatts of capacity to reduce balance-sheet risk.

Analyst perspectives suggest Nvidia’s dual move—an equity stake in SB Energy plus a debt-guarantee cap—positions it strategically while raising concerns about cross-asset volatility and implications for decentralized compute and tokenized infrastructure.

The deal signals a structural shift in AI infrastructure finance toward institutional syndication and external credit platforms, with potential ripple effects across equities, debt, energy, and digital assets.

OpenAI remains the anchor tenant but lacks an investment-grade rating and GAAP profitability, prompting the need for external credit enhancements and a long-term master lease facilitated by Nvidia and SoftBank.

Key risks include grid interconnection delays, permitting hurdles for new generation facilities, and the challenge of monetizing AI software at scale to sustain large capital outlays.

A broad syndicate including Goldman Sachs, Morgan Stanley, Apollo, Blackstone, BlackRock, Brookfield, and KKR is coordinating to reduce reliance on Nvidia’s balance sheet and support vendor-financed capital expenditures.

The deal reflects a wider trend of tying chips, power, and data-center development together to secure AI infrastructure.