Nvidia and OpenAI Partner on Historic Ohio Data Centre Megacampus
Nvidia is in advanced discussions to provide a USD $250 billion (approximately AUD $385 billion at current exchange rates) financing guarantee to backstop OpenAI’s lease of a proposed 10-gigawatt data centre megacampus in Pike County, southern Ohio. The facility, being developed by SB Energy, a subsidiary of Japanese conglomerate SoftBank, is planned for a site previously occupied by the Portsmouth Gaseous Diffusion Plant, a now-decommissioned uranium-enrichment facility on federal land. When semiconductor hardware procurement costs are included, the total projected expenditure for the campus exceeds USD $500 billion, making it the largest single data centre infrastructure proposal in recorded history.
The deal is structurally unusual because Nvidia, as the world’s dominant graphics processing unit manufacturer, is effectively underwriting the financial obligations of a tenant whose core business activity is purchasing Nvidia’s own products at massive scale. This arrangement has reignited serious debate among financial analysts and institutional investors about circular financing, a dynamic where vendors extend credit to customers specifically to enable them to buy vendor products. Nvidia’s share price declined approximately 5% following the news, reflecting credit market concern about the company’s expanding contingent liability exposure. The first phase of the campus is expected to deliver 800 megawatts of capacity by 2028, with the full 10-gigawatt buildout representing a facility roughly ten times the size of any currently operational data centre complex.
For Australian professionals in engineering, planning, environmental consulting, and infrastructure advisory roles, this development is significant not because it involves Australian soil, but because it signals where global capital and computational infrastructure are heading. The physical demands of frontier AI, in terms of land, energy, brownfield remediation, and grid integration, are now operating at a scale that will reshape industrial land use policy and energy market dynamics internationally, including in Australia, where AI infrastructure investment is accelerating rapidly.
Key details of the Nvidia-OpenAI Ohio megacampus financing arrangement
The 10-gigawatt facility requires an electricity supply equivalent to approximately 8.4 million average United States homes, or roughly 33,600 megawatt-hours of continuous generation capacity. To meet this requirement, the project includes a dedicated USD $33 billion natural gas turbine power plant, which is being funded separately by the Japanese government under a bilateral trade agreement designed to reduce United States tariff pressure on Japanese exports. This is a notable geopolitical dimension of the project: the energy infrastructure underwriting one of the world’s largest AI computing facilities is being financed as a diplomatic concession rather than a purely commercial transaction.
Separate from the USD $250 billion lease guarantee, Nvidia and OpenAI are also reportedly negotiating an additional financing structure worth up to USD $350 billion to fund OpenAI’s chip procurement specifically for this campus. This means the total financial exposure attached to the Ohio megacampus, combining the real estate lease guarantee and hardware financing, could reach approximately USD $600 billion across both arrangements. Nvidia CEO Jensen Huang, in a Bloomberg interview on 27 July 2026, justified the need for this scale by arguing that the global computing industry will need to expand by roughly ten times over the next decade to accommodate an estimated 100 billion autonomous AI agents operating simultaneously. OpenAI CEO Sam Altman, speaking on the Relentless podcast on 26 July 2026, indicated the industry had already reached an inflection point in agentic AI deployment.
The site selection carries its own technical complexity. The Portsmouth Gaseous Diffusion Plant site in Pike County, Ohio, is federally managed land with a documented history of uranium enrichment operations, which typically leaves a legacy of radioactive contamination, heavy metals, and volatile organic compounds in soil and groundwater. Because the site sits on federal land, allocation of power capacity and site access is overseen by the United States Commerce Secretary, currently Howard Lutnick. Competitors including Anthropic, Microsoft, and Google have all reportedly lobbied for access to the site’s power capacity, reflecting how constrained high-voltage power supply has become as a limiting factor for AI infrastructure deployment globally.
The circular financing concern being raised by analysts draws a direct historical parallel to Lucent Technologies during the dot-com era of the late 1990s. Lucent extended billions of dollars in vendor loans to telecommunications startups to enable those companies to purchase Lucent hardware. When the startups defaulted, Lucent absorbed catastrophic losses that contributed to the company’s near-collapse. In Nvidia’s case, the mechanism is structurally similar: Nvidia is guaranteeing the lease obligations of a company whose primary capital expenditure is Nvidia silicon. If OpenAI’s revenue growth fails to match its capital commitments, the contingent liability falls on Nvidia’s balance sheet. For environmental and infrastructure practitioners, the parallel is equally instructive in physical terms: the Ohio project illustrates how quickly contingent financial exposure and contingent environmental liability can compound when brownfield remediation obligations, grid connection costs, and long-term land tenure arrangements are layered across a single site at unprecedented scale.
Australian context: AI infrastructure investment and brownfield redevelopment in Australian regulatory frameworks
Australia is not insulated from the dynamics driving the Ohio megacampus proposal. Domestic hyperscale data centre investment has accelerated significantly across New South Wales, Victoria, and Queensland, with major announcements from Microsoft, Google, and Amazon Web Services collectively committing tens of billions of Australian dollars to local infrastructure over the 2024 to 2030 period. These facilities are increasingly targeting industrial and former industrial land close to established transmission infrastructure, where existing grid connections and zoned land use can reduce development timelines. In practice, this means Australian environmental consultants, planners, and infrastructure advisors are encountering data centre proposals on sites carrying legacy contamination, requiring remediation works to be sequenced alongside civil construction rather than completed beforehand. State planning frameworks in New South Wales and Victoria have introduced specific pathways for critical digital infrastructure, but the pace of incoming proposals is testing the capacity of both regulators and consultants to assess cumulative grid impacts, groundwater risk, and cooling water demand at the scale now being proposed.
References and related sources
- Primary source: www.forbes.com
- cryptobriefing.com
- techpowerup.com
- aiweekly.co
- digitalapplied.com
How iEnvi can help
iEnvi integrates technology and data-driven approaches into environmental consulting. We monitor AI and technology developments that affect how environmental professionals deliver services to clients.
This is an iEnvi Machete news summary. Prepared by iEnvi to summarise the source article for environmental professionals tracking AI, data, and technology developments that affect consulting and project delivery.
Published: 28 Jul 2026
Need advice on this topic? Speak to an iEnvi expert at info@ienvi.com.au or 1300 043 684, or contact us online.
Need advice on this issue? iEnvi provides practical, senior-led environmental consulting across contaminated land, remediation, ecology and environmental risk.