OpenAI’s latest infrastructure deal shows how the AI race is moving beyond models and chips into land, electricity, buildings, and long-term financial commitments. The company has signed a 20-year lease with SoftBank subsidiary SB Energy for the "PORTS-Pike" campus in Ohio, a project described as the largest data center project announced to date.
The arrangement gives OpenAI around 8 gigawatts of IT capacity. The Wall Street Journal puts the project’s gross total at 10 gigawatts when cooling and infrastructure are included.
What OpenAI is leasing in Ohio
The Ohio site is partly located on a former US Department of Energy uranium enrichment facility. It is expected to draw power from a 9.2-gigawatt gas plant owned by the US government and financed by Japan as part of a trade agreement.
For OpenAI, the lease is not simply a real estate transaction. It is a way to secure finished compute capacity over a long period as demand for AI infrastructure grows. OpenAI says it only pays for finished capacity, and the first 800 megawatts are slated to come online in 2028.
The scale matters because AI systems require more than chips. They need sites where those chips can be installed, powered, cooled, connected, and operated. That is why the Ohio data center lease is being treated as a major marker in the AI buildout.
Nvidia’s role goes beyond supplying GPUs
Nvidia is backing the project with up to $105 billion, but the structure is narrower than a simple rent guarantee. According to the Wall Street Journal, Nvidia is not guaranteeing OpenAI’s lease payments. Instead, it is backing the residual value of finished data centers in the first construction phase, which covers 4.25 gigawatts of IT capacity.
If OpenAI were to walk away, SB Energy would first need to seek another tenant and then try to sell the facilities. Nvidia would only cover the difference in value after that process, with the guarantee capped at $105 billion.
In exchange, Nvidia becomes the exclusive chip supplier for the first half of the site. Nvidia is also investing $1.5 billion in SB Energy.
Nvidia CEO Jensen Huang framed the issue with the acronym "LPS," meaning "land, power, and shell." In his view, those basics have become the key constraint for AI expansion, replacing chips and networking gear as the biggest bottleneck.
The numbers point to a much wider commitment
Huang expects around 1.5 million GPUs per system generation, or $150 to $200 billion in revenue. Across all sites, he puts OpenAI’s commitments through 2030 at roughly 12 gigawatts of Nvidia compute.
The Ohio package could grow further. If Nvidia exercises its option on the remaining 3.75 gigawatts in Ohio, the package grows to about 16 gigawatts worth roughly $600 billion.
The source also notes that Nvidia had originally wanted to backstop the entire project with around $250 billion. That plan was scaled back after a 5 percent stock drop and pressure from investors. The final $105 billion cap is below the $120 billion figure that had most recently circulated, and it covers asset value rather than the full lease.
Why off-balance-sheet AI obligations matter
The deal also highlights a broader financial issue in the AI sector. A Wall Street Journal analysis found that nine tech companies, including Alphabet, Meta, Microsoft, and Nvidia, hold around $3 trillion in mostly AI-related obligations that do not appear on their balance sheets.
That happens because leases are recorded only once payments begin, while purchase commitments are recorded only when goods are delivered. Leases that have not yet started add up to $1.2 trillion, four times as much as a year earlier. At Alphabet, purchase commitments rose from $332 billion to $811 billion within three months.
The contracts are described as nearly impossible to cancel. Alphabet and Amazon have recently reported negative free cash flow, and Morgan Stanley analysts warn that investors can barely gauge these companies’ actual debt levels anymore.
The bigger signal for AI infrastructure
The OpenAI Ohio data center lease shows how AI growth is becoming tied to long-duration infrastructure contracts and complex financing structures. Compute capacity is no longer just a matter of buying hardware. It also depends on sites, energy supply, building shells, construction timelines, and who is willing to absorb financial risk before capacity is ready.
For OpenAI, the lease secures a large pipeline of future capacity. For Nvidia, the arrangement protects chip demand while placing the company deeper inside the financing of the AI buildout. For investors, the deal adds to a growing question: how much AI spending is already committed, even when it has not yet reached the balance sheet?