Nvidia plans to provide $250 billion in guarantees for OpenAI data centers, marking a new phase in AI infrastructure integration

Nvidia is negotiating with OpenAI over financing guarantees of up to $250 billion for a data center project.

On July 26, according to The Wall Street Journal, Nvidia is in talks to provide around $250 billion in guarantees to help OpenAI lease computing power from a massive data center project. This deal would assist OpenAI in leasing a 10-gigawatt data center hub under development by SoftBank Group in Ohio. SoftBank founder Masayoshi Son previously stated that the project could cost up to approximately $500 billion.

This potential deal completely shatters previous market rumors about “Nvidia cutting back on a $100 billion-level capital commitment to OpenAI.” From shelving a $100 billion direct investment due to OpenAI’s IPO preparations to transitioning into a $250 billion project financing guarantee, Nvidia has not only significantly increased its capital commitment but also substantially upgraded its support approach.

Bypassing equity games—$250 billion guarantees reshape capital binding model

Previously, there had been market rumors that Nvidia had excluded a $100 billion-level funding plan for OpenAI. According to early planning, Nvidia had once considered a $100 billion investment, but this was postponed due to OpenAI’s IPO progress and adjusted to the current $30 billion round.

Now, the emergence of the $250 billion guarantee plan shows that Nvidia has not reduced its financial support for OpenAI, but has fundamentally changed its approach. Shifting from direct equity investment to project financing guarantees, Nvidia cleverly avoids the pre-IPO valuation games and the risk of equity dilution with OpenAI. This model enables the leverage of large-scale infrastructure projects with lower capital occupation, and the deep binding can ensure that OpenAI’s huge future computing power purchase orders continue to flow to Nvidia, maximizing both its commercial benefits and ecosystem control.

Anchoring supercomputing hubs and government electricity, accelerating the rollout of a “trillion-token factory”

The core of this negotiation is that the guaranteed funds will be strictly designated for OpenAI’s mega data center hub project, aiming to build one of the world’s largest AI computing centers. Notably, the project also involves power resources controlled by the US government, signaling an expansion in AI infrastructure and a release of cooperation signals between government and enterprises.

This aligns closely with the “trillion-token factory” blueprint recently proposed by Nvidia CEO Jensen Huang. Huang has made it clear that the focus of AI is shifting from model training to inference, and data centers will transform into AI engines that process unstructured data. To support this vision, OpenAI must build a data center hub of unprecedented scale, but heavily capital-intensive infrastructure faces major financial and energy constraints. By providing massive guarantees and introducing government power resources, Nvidia is effectively clearing both funding and energy obstacles for OpenAI’s computational expansion.

Financial exposure test and profound impact on AI infrastructure landscape

Although Nvidia’s current cash flow is abundant, a $250 billion guarantee is still massive and will inevitably have a significant impact on its financial statements. As the guarantor, Nvidia will undertake considerable contingent liability risks. If OpenAI’s data center project fails to generate the expected returns or a debt default occurs, Nvidia’s financial risk exposure would be forced to expand. Thus, the final terms of the guarantee, risk-sharing mechanisms, and the likelihood of regulatory approval remain the main points of contest at this negotiation stage.

From an industry perspective, this move will further consolidate Nvidia’s moat in the AI computing power field. Coupled with Nvidia’s recently disclosed $95.2 billion in supplier purchase commitments and approximately $90 billion in ecosystem investments covering more than 145 AI industry chain companies over the past 16 months, Nvidia is trying to establish its absolute dominance in next-generation AI infrastructure through a combination of financial instruments, hardware capacity, and energy barriers. This not only secures long-term demand from its largest customer, but also dramatically raises the entry threshold for competitors like AMD and Intel.