Nvidia Corp. has reduced the financing guarantee it may provide for an OpenAI data center project in Ohio [1], [2].
The move signals a shift in how the world's leading AI chipmaker manages its financial risk as the industry scales its physical infrastructure. Because these data centers require massive upfront capital, any change in guarantees can impact the speed of AI deployment and the stability of partnership agreements.
The project is being developed by SB Energy, a company owned by SoftBank [2], [3]. Initial reports indicated that Nvidia had considered a financing guarantee of $250 billion [1]. However, the company has since scaled back this commitment due to concerns from investors regarding Nvidia's risk exposure [1], [3].
Reports on the new guarantee amount vary across sources. One report states Nvidia will provide a guarantee of up to $105 billion [2]. Another source indicates the amount was reduced to under $120 billion [3]. These figures represent a significant decrease from the initial $250 billion figure [1].
While most reports focus on the reduction, some earlier discussions suggested the potential for a guarantee as high as $500 billion for the lease of the SoftBank-owned facility [4]. The current trend, however, shows a tightening of financial commitments as the company balances its role as both a hardware supplier and a financial backer.
Nvidia remains a critical partner for OpenAI, providing the H100 and Blackwell chips necessary to train and run large language models. The Ohio facility is intended to house these systems at a massive scale to meet growing demand for generative AI services.
“Nvidia has since scaled back this commitment due to concerns from investors regarding Nvidia's risk exposure.”
This reduction suggests that Nvidia is facing increased pressure from shareholders to decouple its hardware sales from the massive capital expenditures required for data center construction. By lowering its guarantee, Nvidia limits its potential liability if the OpenAI project fails to generate expected returns, reflecting a more cautious approach to the 'AI bubble' risks often cited by market analysts.
