September 26, 2026
OpenAI is pushing the boundaries of artificial intelligence infrastructure with plans to lease a colossal 10-gigawatt data center in Ohio, a project that could redefine the scale of AI computing. According to exclusive details from The Information, this ambitious facility would mark OpenAI’s largest infrastructure commitment to date, with Nvidia stepping in as a financial guarantor—an unprecedented move for the chip giant.
The proposed data center, developed on federal land in southern Ohio by SB Energy—a subsidiary majority-owned by SoftBank, a key OpenAI investor—would come with a staggering price tag. At full capacity, the project could exceed $500 billion in costs, with OpenAI securing a 20-year lease. Nvidia’s role as a financial backer, using its balance sheet to guarantee payments, signals a bold new direction in AI infrastructure financing. This initiative mirrors the stalled Stargate project announced in early 2025, which also involved Oracle and SoftBank but failed to gain traction.
The Ohio site, once home to a uranium enrichment facility in Pike County, is slated to begin its first phase—an 800-megawatt operation—by 2028. If realized, this facility would dwarf existing AI data centers, providing the computational power needed to train next-generation AI models. The sheer scale of the project underscores OpenAI’s long-term vision for AI dominance, even as negotiations remain fluid and subject to change.
Nvidia’s involvement as a guarantor for both the lease and project financing is a game-changer. Historically, the chipmaker has focused on supplying hardware rather than underwriting infrastructure costs. By leveraging its balance sheet to back this $500 billion endeavor, Nvidia is signaling its confidence in OpenAI’s roadmap—and its willingness to take on financial risk to secure a foothold in the AI ecosystem.
This partnership could set a precedent for future AI infrastructure projects, where hardware providers like Nvidia play a more active role in funding data centers. If successful, it may encourage other tech giants to explore similar arrangements, accelerating the development of AI supercomputers.
SoftBank’s SB Energy is developing the Ohio site, reflecting the conglomerate’s continued investment in AI infrastructure. However, this isn’t SoftBank’s first foray into large-scale AI projects. The Stargate initiative, announced in January 2025 with Oracle, aimed to create a similarly massive data center but ultimately stalled due to logistical and financial hurdles.
The Ohio project’s revival suggests OpenAI and its partners have learned from past missteps. By securing Nvidia’s financial backing and opting for a phased rollout—starting with an 800-megawatt phase—they’re mitigating risks while keeping long-term scalability in mind.
The timing of this data center announcement is notable, coming just days after OpenAI confidentially filed paperwork for an initial public offering (IPO). A project of this magnitude could bolster investor confidence by demonstrating OpenAI’s ability to secure long-term infrastructure commitments. It also positions the company as a leader in AI compute capacity, a critical factor for maintaining its competitive edge against rivals like Google and Microsoft.
AI’s growing computational demands are driving a shift toward larger, more centralized data centers. OpenAI’s Ohio facility, if completed, would represent a new benchmark for AI infrastructure, far surpassing the capabilities of today’s largest supercomputers. This trend reflects the industry’s need for massive, dedicated resources to train increasingly complex models like GPT-5 and beyond.
Nvidia’s decision to back OpenAI’s data center financially marks a strategic pivot. Traditionally, the company has profited from selling GPUs and AI accelerators. By guaranteeing infrastructure costs, Nvidia is deepening its ties with key AI developers while potentially opening a new revenue stream. This move could also pressure competitors like AMD and Intel to explore similar partnerships.
A project of this scale inevitably raises questions about energy consumption and regulatory approvals. The Ohio site’s 10-gigawatt capacity would require significant power resources, likely necessitating partnerships with local energy providers. Additionally, repurposing a former uranium enrichment facility may involve environmental assessments and community engagement to address potential concerns.
OpenAI’s 10-gigawatt data center is designed to meet the escalating computational demands of advanced AI models. The facility will provide the infrastructure needed to train next-generation systems, ensuring OpenAI remains competitive in the AI race.
Nvidia is acting as a financial guarantor for the data center, backing both the lease and project financing with its balance sheet. This is a first for the chipmaker at this scale and reflects its commitment to supporting AI infrastructure growth.
The Ohio data center shares similarities with the stalled Stargate project, including its massive scale and involvement of SoftBank. However, OpenAI’s phased approach and Nvidia’s financial backing may improve its chances of success.
The first phase, an 800-megawatt facility, is expected to launch by 2028. Full buildout to 10 gigawatts would take longer, depending on funding and construction progress.
OpenAI’s planned 10-gigawatt data center in Ohio represents a monumental leap in AI infrastructure, with Nvidia’s financial backing adding a new dimension to the project’s viability. If successful, this facility could set a new standard for AI compute power, enabling breakthroughs in model training and deployment. The partnership also highlights Nvidia’s evolving role in the AI ecosystem, transitioning from a hardware supplier to a key financial enabler of large-scale projects.
As negotiations continue and the 2028 launch date approaches, the tech industry will be watching closely. This project could either cement OpenAI’s leadership in AI or serve as a cautionary tale about the challenges of scaling infrastructure at this magnitude. One thing is certain: the future of AI will be built on foundations like these.
