Nvidia is putting even more money behind the power plants that keep its chips running. The chipmaker has agreed to buy an additional $1.5 billion worth of shares in SB Energy, the SoftBank-backed energy and data center company preparing to go public in the United States, according to Bloomberg and other reports. The move pushes Nvidia’s total equity commitment in SB Energy to $3 billion, cementing the Nvidia SB Energy investment as one of the more unusual financing bets tied to the AI boom.
Summary
Nvidia‘s latest move deepens its financial ties to the physical infrastructure powering artificial intelligence, not just the chips that run on it. The additional $1.5 billion purchase comes ahead of SB Energy’s planned listing and brings Nvidia’s total equity stake in the company to $3 billion, according to reports covering the deal.
The fresh capital injection signals that Nvidia wants a bigger seat at the table before SB Energy hits the public markets. Rather than waiting to buy in during the offering itself, Nvidia locked in its position early, a pattern that echoes similar moves the chipmaker has made across the AI supply chain in recent quarters.
With this tranche, Nvidia’s cumulative equity exposure to SB Energy doubles to $3 billion. That figure represents a direct financial stake in the company that will operate some of the largest AI-dedicated data center campuses in the country.
The structure of the deal matters as much as the size. Nvidia’s participation in the $1.5 billion tranche comes through newly issued Class N non-voting shares, obtained via a private placement paired with a prepaid forward contract, with both instruments carrying the same price as the public offering—putting Nvidia on equal footing with IPO investors financially, though it will hold no voting power. That arrangement lets Nvidia expand its financial exposure to SB Energy without taking on governance influence over the company.
SB Energy is preparing one of the more closely watched listings tied to the AI infrastructure buildout, with a target valuation of roughly $50 billion. The company filed its IPO registration in late August 2026 and is aiming to list on Nasdaq under the ticker SBE.
SB Energy operates as the energy and data center subsidiary of SoftBank, giving it a direct role in building the power-heavy facilities that AI companies need to run large computing workloads. Its position inside SoftBank’s broader portfolio has made it a focal point for investors trying to gauge how AI demand translates into real-world infrastructure spending.
SB Energy is aiming to raise between $5 billion and $7 billion through the offering, a range that would make it one of the larger listings connected to the current AI infrastructure cycle. Trading under the SBE ticker on Nasdaq would give public market investors direct exposure to a company whose business is essentially selling power and space to AI compute operators.
Beyond the Nasdaq listing, SB Energy is also reportedly seeking to raise $500 million separately from Japanese investors, according to Bloomberg. That capital is earmarked specifically for data center and power infrastructure development, adding another funding stream on top of the US offering and Nvidia’s equity commitment.
The reason Nvidia is willing to write such large checks becomes clear once you look at what SB Energy is actually building. Its centerpiece initiative, the PORTS-Pike Technology Campus located in Ohio, is engineered to provide an initial AI computing capacity of 4.25 gigawatts, with room to grow up to 8 gigawatts.
That scale puts PORTS-Pike among the largest dedicated AI computing sites currently under development in the country. The campus is built specifically to handle the power demands of large-scale AI training and inference, the kind of workload that has driven surging electricity needs across the data center industry.
The Ohio campus is primarily leased to OpenAI under a long-term agreement, and that lease mandates the exclusive use of Nvidia chips. In practice, that means every watt of computing power flowing through PORTS-Pike is tied directly to Nvidia’s GPU business, giving the chipmaker a built-in commercial return on its equity stake.
Nvidia’s involvement goes beyond equity. In connection with the Ohio project, the company has additionally pledged residual value guarantees reaching as high as $105 billion. According to Invezz, that guarantee sits within a broader $108.5 billion maximum gross exposure Nvidia disclosed in its latest filing, covering AI infrastructure commitments including $3.5 billion tied to certain AI-cloud partners. Invezz reported that the OpenAI-linked guarantee does not become fully effective all at once; obligations are expected to increase as nine data centers enter service, beginning around fiscal 2029, before declining as OpenAI makes lease payments. The guarantee is also triggered only under specified circumstances, such as tenant default or insolvency, which limits its immediate financial impact without removing the exposure entirely.
This is where the story shifts from a straightforward equity purchase to something more structural. Nvidia is no longer simply selling GPUs into demand created elsewhere. By backing the financing behind projects like PORTS-Pike, the company is helping ensure the infrastructure exists in the first place, and positioning itself to benefit commercially once it does, through chip-exclusivity clauses baked directly into the lease.
SB Energy’s own numbers show a company growing fast but still losing money. During the first half of 2026, revenue climbed 66.4% compared to the prior year, driven by soaring demand for AI-grade data center capacity, while net losses grew larger over that same span.
The combination of rapid revenue growth alongside wider losses is not unusual for a capital-intensive infrastructure company scaling up ahead of an IPO, but it does put a spotlight on how SB Energy’s business model performs once construction costs and financing obligations are weighed against incoming lease revenue.
For investors watching the offering, the revenue trajectory offers reassurance that demand for AI computing capacity remains strong. The widening losses, on the other hand, are a reminder that building gigawatt-scale campuses is expensive long before the leases fully pay off. Nvidia’s willingness to keep increasing its stake, and to back the project with residual value guarantees, suggests the chipmaker sees the long-term payoff as worth the near-term financial exposure.
Nvidia is purchasing an additional $1.5 billion in SB Energy shares, bringing its total equity commitment to $3 billion.
Nvidia is acquiring newly issued Class N non-voting shares via private placement and a prepaid forward contract.
SB Energy aims to raise between $5 billion and $7 billion through its IPO on Nasdaq under the ticker SBE, targeting a valuation of roughly $50 billion.
It is an AI computing facility designed for an initial load of 4.25 gigawatts expandable to 8 gigawatts, leased long-term to OpenAI under terms requiring exclusive use of Nvidia chips.
{"@context":"","@type":"FAQPage","mainEntity":[{"@type":"Question","name":"How much is Nvidia investing in SB Energy ahead of the IPO?","acceptedAnswer":{"@type":"Answer","text":"Nvidia is purchasing an additional $1.5 billion in SB Energy shares, bringing its total equity commitment to $3 billion."}},{"@type":"Question","name":"What type of shares is Nvidia acquiring in the latest investment tranche?","acceptedAnswer":{"@type":"Answer","text":"Nvidia is acquiring newly issued Class N non-voting shares via private placement and a prepaid forward contract."}},{"@type":"Question","name":"What are SB Energy's IPO goals and planned listing exchange?","acceptedAnswer":{"@type":"Answer","text":"SB Energy aims to raise between $5 billion and $7 billion through its IPO on Nasdaq under the ticker SBE, targeting a valuation of roughly $50 billion."}},{"@type":"Question","name":"What is the significance of SB Energy's Ohio PORTS-Pike Technology Campus?","acceptedAnswer":{"@type":"Answer","text":"It is an AI computing facility designed for an initial load of 4.25 gigawatts expandable to 8 gigawatts, leased long-term to OpenAI under terms requiring exclusive use of Nvidia chips."}}]}
Article produced with the assistance of artificial intelligence and reviewed by the editorial team.