Nvidia is reportedly discussing a $250 billion financing plan for a significant OpenAI data center in Ohio. This development comes shortly after CEO Jensen Huang expressed support for a contentious “open source” AI model.
The funding would enable OpenAI to lease a 10-gigawatt project in southern Ohio, developed by Japanese billionaire Masayoshi Son’s SoftBank through its energy arm. This project is the largest in the industry’s history, costing a total of $500 billion, as noted in a recent Wall Street Journal article.
OpenAI has been negotiating to lease the site for weeks, and other companies like Anthropic, Microsoft, and Google are also discussing the project with Commerce Secretary Howard Lutnick.
However, this potential deal has raised concerns about Nvidia monopolizing the AI boom. Industry insiders, including some Wall Street traders, have warned that these funding arrangements are “cyclical,” as Nvidia invests in companies using its chips.
In a contrasting move, Huang recently posted on X, urging for an open-source AI model. This letter was co-signed by major tech firms such as Microsoft, Meta, IBM, and Palantir Technologies. They formed the Open Secure AI Alliance, advocating against early restrictions on open models that could hinder competition or innovation abroad.
Huang’s remarks are seen as pivotal in the growing debate surrounding open-source versus closed-source AI models, a conversation stirring tension in both Silicon Valley and Washington.
Huang and other industry leaders argue that an AI model without a single company controlling it could accelerate development and reduce power concentration. Conversely, opponents, including those from Anthropic and OpenAI, claim that public access to AI blueprints could allow malicious actors to create harmful tools.
Anthropic CEO Dario Amodei recently reiterated his concerns at a Senate committee hearing, stating that the open-source path poses significant dangers.
OpenAI’s Sam Altman also has reservations about open-source AI. He previously mentioned that if they did not adopt certain open-source elements, the global landscape could become dominated by China’s model.
In his letter, Huang recognized the risks linked to open-source AI models but argued that, ultimately, AI will foster “innovation and prosperity.” He added that while open weights do present clear dangers, banning them isn’t a viable solution.
The alliance group announced plans to develop open-source AI tools intended to help defend against cyber attacks.
This Ohio project poses considerable risks for Lutnick and the Trump administration. Under a Japan-U.S. tariff agreement, the Japanese government has committed $33 billion for building a natural gas facility managed by SB Energy, a SoftBank subsidiary.
In March, Lutnick, his son, and Energy Secretary Chris Wright began construction on the site, previously a uranium enrichment plant near Columbus. The U.S. will cover SB Energy’s operational costs, with revenue from electricity divided evenly until Japan recoups its investment, after which the U.S. will receive the majority of the income.
Other tech companies are exploring similar open-source models, with Google having financially supported some of Anthropic’s data centers.
While Nvidia and other advocates of open-source argue that proprietary tools can also be misused, they believe democratizing AI through open-source solutions would enhance protection for all.
If Nvidia proceeds with backing the data center project, SoftBank will likely secure better financing terms, as OpenAI, an unprofitable private firm without an investment-grade credit rating, would enhance credibility.
Reportedly, Nvidia has already invested $30 billion in OpenAI and is working on a deal that could fund up to $350 billion in chip purchases for the company.
This ambitious AI campus is expected to consume about 10 gigawatts of energy—enough for millions of homes—and will take years to construct, with completion of the first phase anticipated by 2028.
