OpenAI in talks to lease 10 gigawatt Ohio data center with Nvidia backing- The Information

OpenAI is reportedly in advanced talks to lease a 10 gigawatt data center campus in Ohio, with total buildout costs potentially reaching at least $500 billion and possible backing from Nvidia. The first phase is expected to begin in 2028, and OpenAI would control the equipment under a long-term lease. The deal underscores a massive AI infrastructure buildout and could be supportive for Nvidia and related data center supply chains.

Analysis

The key read-through is not just “more AI capex,” but a potential multi-year re-acceleration in compute demand that stretches the supply chain far beyond GPUs. A 10GW campus implies a step-function increase in demand for power infrastructure, transformers, switchgear, cooling, networking optics, and permitting expertise; those bottlenecks likely matter more in the next 12-24 months than the chips themselves. That makes NVDA an obvious beneficiary, but the second-order winners may be the picks-and-shovels vendors with pricing power in grid equipment and datacenter interconnect, while hyperscaler-style buildout pressure could tighten lead times across the entire AI hardware stack.

The market is likely underestimating execution risk on the timeline. A 2028 first phase means the near-term upside is mostly narrative and order-book optionality, not immediate revenue, so the stock reaction may overshoot on headline enthusiasm and then fade unless there is a clearer financing, power, and deployment roadmap. If the project advances, the bigger implication is a longer-duration demand floor for next-gen accelerators and memory, which should support semi multiples through cyclicality—but only if supply growth remains disciplined.

Contrarian view: the $500B buildout number is so large that it can become a signal of capital intensity risk rather than unambiguous bullishness. The market may start to question whether AI returns on invested capital can absorb this level of infrastructure spend, especially if energy costs, depreciation, and utilization remain uncertain. That creates a bifurcation: semiconductor vendors with already-contracted demand benefit, while names tied to pure AI monetization need proof that capex is translating into cash flow, not just capacity.

Key Decisions for Investors

• Stay long NVDA, but structure via 3-6 month call spreads rather than outright stock; the trade captures narrative-driven upside while limiting drawdown if the market starts discounting execution risk before revenue shows up.
• Add to SMCI only on weakness if there is evidence of follow-on datacenter capex orders; otherwise keep it as a tactical trade, since the multiple is vulnerable if investors conclude this buildout is too far out on the curve.
• Initiate a basket long in AI infrastructure beneficiaries against a short in higher-quality AI application names that rely on near-term monetization; the market may reward hard-asset enablers before it rewards software monetization stories.
• Watch for a pullback in NVDA/SMCI after the first headline spike and buy the dip only if there is confirmation of power/land/financing milestones; absent that, the risk/reward is better on implied-volatility selling than chasing spot.
• Set a 6-12 month alert on utility and grid-equipment suppliers tied to datacenter load growth; the strongest asymmetric opportunity is likely in the non-obvious infrastructure layer, not the obvious chip winner.