Nvidia in Talks for $3 Billion Stake in SoftBank's SB Energy

A Strategic Infrastructure Play

Nvidia is negotiating an investment of up to $3 billion in SB Energy, the SoftBank Group subsidiary building large-scale power and data center infrastructure across the United States. The deal would tie the chip maker directly into one of the industry's most ambitious data center projects, a sprawling campus in Ohio designed to serve OpenAI's compute requirements.

The proposed equity stake represents roughly half of what SB Energy aims to raise in an initial public offering planned for as early as next month. Founded in 2019, SB Energy has positioned itself at the intersection of two infrastructure booms: renewable energy and AI-driven data center expansion. OpenAI already holds a stake in the company.

Nvidia's involvement extends beyond the equity investment. The chip maker has been in discussions with both OpenAI and SB Energy about providing approximately $100 billion in credit support for the Ohio project, a financing structure that would underpin one of the largest single data center developments in the United States. The arrangement would see Nvidia commit half of its $3 billion investment when the Ohio deal is formally signed, with the remainder tied to SB Energy's public market debut.

Scaling Back from Initial Ambitions

The current financing framework marks a significant revision from earlier plans. Nvidia had previously discussed guaranteeing as much as $250 billion to support the Ohio campus, a figure that would have dwarfed most infrastructure commitments in the technology sector. The company has since recalibrated, with initial credit guarantees now expected to come in below $120 billion.

The shift reflects both the complexity of financing such a large project and the evolving dynamics between Nvidia, OpenAI, and their partners. Data center projects of this scale require not only capital but also long-term commitments on power supply, chip procurement, and operational coordination. Nvidia's dual role as both investor and hardware supplier gives it unusual leverage in shaping the project's economics.

SB Energy's planned IPO, targeting at least $5 billion in proceeds, would provide additional capital to fund the Ohio campus and other data center developments in its pipeline. The company has been building several campuses in parallel, responding to surging demand from AI companies that require dedicated infrastructure for training and deploying large language models.

Asia's Role in the AI Infrastructure Buildout

The deal underscores SoftBank's continued influence in global AI infrastructure, even as its portfolio has faced scrutiny in recent years. SoftBank founder Masayoshi Son has repeatedly signaled his intention to position the conglomerate at the center of the AI revolution, with investments spanning chips, software, and now the physical infrastructure that underpins the technology.

For Nvidia, the investment represents a strategic extension of its core business. The company has increasingly moved beyond selling chips to shaping the ecosystems in which those chips operate. By taking an equity position in SB Energy, Nvidia gains visibility into long-term demand patterns and ensures its hardware remains central to the most significant AI infrastructure projects.

The Ohio campus, once operational, would rank among the largest concentrations of AI compute capacity in the world. OpenAI's partnership with SB Energy signals the company's need for dedicated infrastructure as it scales beyond shared cloud platforms. The arrangement also highlights a broader trend: the largest AI developers are moving toward owning or controlling their infrastructure rather than relying solely on hyperscale cloud providers.

Neither Nvidia nor SB Energy has confirmed the terms of the discussions. The timing of any final agreement remains uncertain, though the planned IPO timeline suggests SB Energy is moving quickly to lock in capital commitments. If the deal proceeds as outlined, it would mark one of the largest single investments by a chip maker in data center infrastructure, setting a precedent for how hardware companies participate in the capital-intensive buildout required to support the next generation of AI systems.