Nvidia Puts $1.5 Billion Into SoftBank-Linked Data Center Project for OpenAI

Nvidia is investing $1.5 billion in SB Energy, the SoftBank-linked power and data center developer behind OpenAI’s planned Ports-Pike campus near Cincinnati, Ohio. The deal strengthens Nvidia’s position as the sole supplier of compute infrastructure for the project and underscores how the race to build AI systems is now colliding with the harder challenge of securing enough electricity to run them.

The investment comes as SB Energy prepares a massive buildout on federal land that once supported uranium enrichment for the U.S. nuclear weapons program and Navy submarines. The site’s proposed scale, financing, and power needs make it one of the clearest examples yet of how artificial intelligence is reshaping infrastructure, energy markets, and industrial policy at the same time.

What Nvidia is buying into

Nvidia’s new commitment is aimed at SB Energy, a company that sits at the center of a large AI infrastructure effort involving SoftBank and OpenAI. The arrangement gives Nvidia a financial stake in a project that is not just about data center construction, but about the power generation and compute capacity needed to support frontier AI models.

In practical terms, Nvidia is helping underwrite the physical foundation of a large OpenAI-linked campus while also locking in its own role as the exclusive provider of compute hardware for the site. That includes the systems that will power training and inference workloads once the facility comes online.

The company also disclosed that it may provide up to $105 billion in credit tied to the buildout, according to filings with the U.S. Securities and Exchange Commission. That financing support adds another layer to the deal, making Nvidia not only a supplier and investor but also a potential lender helping to accelerate deployment.

Why the Ports-Pike campus matters

The Ports-Pike project matters because it reflects the scale of investment now required to support modern AI. Large language models and advanced AI applications consume extraordinary amounts of compute, and compute consumes enormous amounts of power. The size of this project shows how those demands are pushing beyond ordinary data center planning into utility-scale infrastructure.

The campus is expected to begin at 4.25 gigawatts and could expand to 8 gigawatts, an enormous figure by data center standards. For comparison, this would place the facility in a league usually associated with regional power systems rather than a single computing site.

Such numbers highlight a new reality in the AI boom: access to chips is no longer enough. Developers now need land, transmission capacity, fuel supply, and long-term financing to bring large-scale AI projects to life.

How the financing is structured

The deal combines direct equity investment with substantial credit backing, a structure that suggests a long-term industrial relationship rather than a simple minority stake. Nvidia’s role is especially notable because it connects the company’s dominant position in AI chips to the broader financing ecosystem now emerging around AI infrastructure.

This is not just a bet on a data center operator. It is a wager on the economics of scaling AI in an era when the cost of power plants, transmission, and cooling systems can rival the cost of the chips themselves.

What the SEC filing suggests

The SEC filing indicates the project has reached a level of maturity where capital commitments need to be formalized and quantified. By setting out possible credit support, Nvidia appears to be signaling confidence in the project’s timeline and in demand for the compute capacity it will deliver.

The filing also helps explain why the company wants to secure exclusivity at the site. In a market where access to high-end accelerators is strategic, being the only supplier at a multi-gigawatt campus ensures both revenue and influence over the deployment architecture.

Who owns SB Energy and why does that matter?

SB Energy’s investor base matters because it ties the project to two of the most important names in AI capital formation: SoftBank and OpenAI. Their involvement suggests the campus is part of a broader ecosystem of interconnected bets on model development, cloud-scale compute, and power-intensive infrastructure.

SoftBank’s relationship to Nvidia adds further context. The Japanese conglomerate had previously held $5.8 billion in Nvidia shares, but sold that position in November to free up capital for other AI investments. In effect, SoftBank appears to have rotated out of one kind of AI exposure and into another.

SoftBank’s sale of its Nvidia stake last year was widely understood as a move to redeploy capital into additional AI bets, and the latest project shows how quickly those bets are being translated into concrete infrastructure.

That shift matters because it illustrates how investors are moving from pure chip ownership toward the broader stack: data centers, energy systems, and AI deployment platforms.

Why the power plant is such a big part of the story

The power plant is central because the data center cannot function at the proposed scale without dedicated electricity generation. SB Energy plans to build a 9.2-gigawatt natural gas plant on the site, an eye-catching number that reflects the immense energy appetite of the project.

The estimated cost of the plant is $33 billion. That price tag is important not only because of the scale of the facility but because natural gas power plants have become dramatically more expensive to build. BloombergNEF has estimated that construction costs have climbed 66% over the last two years, a trend that makes large projects harder to finance and slower to complete.

The result is a feedback loop: AI is driving demand for electricity, while the rising cost of electricity infrastructure is becoming a limiting factor for AI expansion.

How natural gas fits the AI buildout

Natural gas is often treated as a practical bridge fuel for large power users because it can provide firm generation that is easier to scale than intermittent renewables alone. For an AI campus that needs steady, high-volume electricity around the clock, that reliability is attractive.

But the economics are getting tougher. Higher equipment costs, labor shortages, permitting delays, and fuel-market uncertainty are all raising the difficulty level for developers attempting to construct giant power plants tied to data centers.

In this case, the project’s location adds another layer of significance. The facility will be built on land owned by the U.S. Department of Energy, giving the plan a public-land backdrop that ties the AI race to America’s historic industrial and nuclear landscape.

What is the significance of the Ohio site?

The Ohio site is significant because it was once used to enrich uranium for U.S. nuclear weapons and Navy submarines. That legacy makes the land a symbol of past strategic infrastructure, now being repurposed for a new technological era centered on AI computing.

The transformation is striking: a place once associated with the Cold War and military supply chains is being positioned to support the next wave of digital intelligence systems. The continuity is not in the technology itself, but in the scale of national importance attached to the site.

For local and regional stakeholders, that also means the project could reshape land use, utility demand, and industrial development patterns around Cincinnati and the broader Ohio-Kentucky-Indiana corridor.

How big is this compared with other infrastructure projects?

It is unusually large even by the standards of the AI boom. Data centers are already among the fastest-growing categories of infrastructure spending in the United States, but a campus measured in gigawatts rather than megawatts sits far beyond most existing facilities.

Here is a simplified comparison of the project’s major components:

The numbers are so large that they blur the line between private tech infrastructure and regional energy development. The project is effectively a combined compute-and-power industrial complex.

Why this deal says so much about the AI economy

The Nvidia-SB Energy deal reveals that the AI industry’s bottleneck has shifted. A few years ago, the conversation centered on model quality, product adoption, and who had the best algorithms. Today, the harder question is whether enough physical infrastructure can be built fast enough to support the next generation of models.

This is why chip companies, energy developers, and AI labs are increasingly intertwined. Each depends on the other: AI labs need compute, chipmakers need customers with massive deployment plans, and power companies need anchor tenants large enough to justify billion-dollar plants.

At the same time, the deal shows how concentrated the AI buildout has become. A small number of players are effectively deciding where the largest compute clusters will be built, how they will be powered, and which hardware architectures will dominate them.

The business logic behind exclusivity

Nvidia’s exclusive supplier role at the Ports-Pike campus is more than a procurement detail. It is a strategic moat. By ensuring that one of the world’s largest prospective AI sites uses its systems exclusively, Nvidia deepens its grip on the high-end infrastructure market.

That arrangement may also simplify integration, maintenance, and deployment planning for SB Energy and its partners. But it also locks the project into Nvidia’s ecosystem at a time when the broader market is watching for any signs of competition in AI hardware.

What happens next?

The next phase will likely involve permitting, engineering, financing, and phased construction. Because the project spans both compute and energy infrastructure, progress is likely to be measured in stages rather than a single groundbreaking moment.

If the campus and power plant move forward as planned, the facility could become one of the most important physical anchors of OpenAI’s long-term infrastructure strategy. If delays hit, the reasons will likely involve the same factors that are making the project so notable: cost, power availability, and the sheer complexity of building at this scale.

For now, Nvidia’s investment is a strong signal that the company believes the demand is real enough to justify putting more capital directly into the infrastructure behind it. It is a reminder that the AI race is no longer just about software breakthroughs. It is about who can assemble the land, energy, chips, and financing needed to run them at industrial scale.

Timeline of the deal

As the AI industry grows more power hungry, deals like this are likely to become more common. But few will be as large, as complex, or as closely tied to the future of frontier AI as Nvidia’s new investment in SB Energy.

Why investors are paying attention

Investors are watching this story because it connects three of the biggest forces in the market: AI demand, power scarcity, and infrastructure inflation. Any one of those can affect returns. Together, they define the economics of the next phase of the AI boom.

For Nvidia, the move expands the company’s influence beyond chips and into the capital structure of the projects that buy them. For SoftBank and OpenAI, it adds a heavyweight partner with both technical and financial power. For the broader market, it offers another sign that AI is becoming an infrastructure story as much as a software story.

And for the energy sector, it is a warning and an opportunity at once: the companies that can deliver reliable power at scale may become indispensable to the future of AI.

In that sense, the Nvidia deal is about more than one data center in Ohio. It is about the industrial transformation required to keep artificial intelligence expanding, and the mounting cost of doing so.