NVIDIA is reportedly in talks to back OpenAI with a guarantee worth approximately US$250bn for a proposed Ohio data centre development, according to The Wall Street Journal.
The newspaper reported that as a whole, the development could exceed US$500bn, including the NVIDIA chips intended for deployment across the facility.
If completed, the project would rank among the largest data centre developments announced to support AI workloads.
As OpenAI is an unprofitable private company, backing from NVIDIA could help the developer obtain debt on more favourable terms.
People familiar with the discussions have told the Wall Street Journal that OpenAI has been in advanced negotiations to lease the site for several weeks.
Ohio project puts data centre financing in focus
According to reports, NVIDIA’s proposed guarantees would support OpenAI’s plans to lease capacity at a 10GW campus being developed by Japanese investment holding company SoftBank in southern Ohio.
The US Department of Energy says SoftBank, alongside subsidiary SB Energy, plans to build at least 9.2GW of natural gas generation to supply power for 10GW of data centre capacity.
Parts of the agreement have already been announced through US President Donald Trump’s US-Japan Strategic Trade and Investment Agreement, which includes US$33.3bn in Japanese funding.
The reported guarantee would strengthen SB Energy’s ability to secure financing by reducing lending risk.
Investor concerns grow alongside AI infrastructure spending
While AI infrastructure investment continues to accelerate, financial markets are showing greater caution over the level of borrowing supporting these developments.
The Financial Times reports that credit default swaps, which act as financial contracts used by investors to hedge against the risk of a borrower failing to repay debt, are becoming more expensive for companies closely associated with AI infrastructure.
The newspaper says prices for credit default swaps linked to Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and NVIDIA have reached record highs, citing data from the London Stock Exchange Group.
Furthermore, shares in South Korean chipmakers SK hynix and Samsung have fallen by 10% and 12% respectively during a wider market sell-off centred on AI-related companies.
Growing concern centres on the amount being spent on data centres, AI accelerators and memory technology as companies compete to expand capacity.
The Bank of England also highlights potential macroeconomic risks.
-
In its Financial Stability Report for July 2026, the Bank of England warned that an AI-related equity price correction could result in a 2.2% reduction in UK GDP.
Sean Peche, Founder and Portfolio Manager at Ranmore Fund Management, urges investors to remain disciplined when assessing long-term expectations.
He says: “In all the bubbles we’ve experienced, everybody at the time said, ‘This is the biggest thing ever to happen’, but the truth is we just don’t know.
“AI is a very, very fast-moving world. Every day there is something new coming out.
“The future quite often turns out to be different from what you expect. So, the key is not to pay too much for those expectations. If the future is unforecastable, don’t overpay for it.”
NVIDIA expands its data centre footprint
The reported Ohio project is not the only large-scale data centre commitment involving NVIDIA.
The Financial Times separately reported that the company has signed leases worth up to US$50bn covering a 1GW data centre under development at Beacon Point in Texas.
The newspaper spoke to five people familiar with the matter, whom it did not identify.
They said that NVIDIA is leasing the entire facility being built by Hut 8.
The developer, formerly focused on cryptocurrency mining, now concentrates on AI infrastructure as demand for high-density computing capacity continues to grow.
Speaking when announcing NVIDIA’s quarterly results in May 2026, Jensen Huang, Founder and CEO of NVIDIA, said the expansion of AI computing facilities is increasing at an exceptional pace.
He described the buildout of AI factories as “the largest infrastructure expansion in human history”.
He added: “NVIDIA is uniquely positioned at the centre of this transformation as the only platform that runs in every cloud, powers every frontier and open-source model and scales everywhere AI is produced – from hyperscale data centres to the edge.”
McKinsey estimates that global spending on data centres could reach US$7tn by 2030, illustrating the scale of investment expected across power, facilities, networking and compute infrastructure as demand for AI capacity continues to expand.
https://datacentremagazine.com/news/nvidia-openai-us-500-bn-ohio-ai-data-centre-talks
