Executive Summary
Nvidia in talks to invest $3 billion in SB Energy, the SoftBank Group-backed infrastructure developer involved in a planned Ohio data-centre project for OpenAI.
The potential investment has not been confirmed. Reuters reported the discussions on 15 August, citing a report by The Information, and said it could not independently verify the report. The development nevertheless highlights a significant shift in AI infrastructure: access to capital, power and deployable data-centre capacity is becoming increasingly interconnected with the supply of accelerators and other computing infrastructure.
What Happened
According to the report, Nvidia is discussing an investment of up to $3 billion in SB Energy in connection with the planned OpenAI data-centre development in Ohio.
The Information also reported that the proposed investment forms part of wider discussions between Nvidia, OpenAI and SB Energy concerning around $100 billion in credit support for the Ohio campus. Reuters said Nvidia and SB Energy did not immediately respond to requests for comment outside regular business hours.
The reported transaction must therefore be treated as a proposal under discussion rather than an agreed Nvidia investment.
SB Energy’s Expanding AI Infrastructure Role
SB Energy already has a confirmed relationship with OpenAI. In January 2026, OpenAI and SoftBank Group announced that they would each invest $500 million in SB Energy, providing a combined $1 billion investment to support the company’s development of data-centre campuses and associated energy infrastructure.
That agreement also included a 1.2GW OpenAI data-centre lease at a site in Milam County, Texas. SB Energy said it had several multi-gigawatt data-centre campuses under development, with initial facilities expected to enter service from 2026.
The Texas development is separate from the planned Ohio project associated with the latest reported Nvidia discussions.
Why the Nvidia Talks Matter
The strategic significance extends beyond the size of the potential equity investment. AI infrastructure increasingly depends on several constrained resources being delivered simultaneously: accelerators, networking, electrical capacity, land, cooling and financing.
Nvidia occupies a critical position in the compute layer through its GPU and networking platforms. Investment in an infrastructure developer would represent exposure further down the AI supply chain, where the availability of powered data-centre capacity can determine how quickly large accelerator clusters can actually be deployed.
SoftBank’s own infrastructure strategy illustrates the scale involved. The group says it is participating in plans for 10GW-scale power generation and a 10GW-scale AI data center at the PORTS Technology Campus in Piketon, Ohio, alongside US government bodies, Energy Global and AEP Ohio.
Power and Capital Become AI Constraints
The development reflects a broader challenge facing hyperscale AI projects. Purchasing computing hardware is insufficient if suitable electrical infrastructure cannot be delivered on a comparable timetable.
Gigawatt-scale campuses can require new generation, transmission and distribution infrastructure as well as substantial financing before compute capacity becomes operational. Developers are consequently being pushed towards closer coordination between technology suppliers, AI companies, utilities, energy developers and infrastructure investors.
For data-centre operators, this changes capacity planning. Power procurement and financing increasingly need to be considered alongside server, networking and cooling architecture rather than as downstream facilities decisions.
What Happens Next
The immediate issue is whether Nvidia and SB Energy reach a definitive investment agreement. The reported $3 billion figure and the wider credit-support discussions should remain classified as unconfirmed unless the companies formally announce a transaction.
Attention will also remain on the financing and delivery of the Ohio infrastructure. Projects of this scale require more than headline capital commitments: power generation, interconnections, construction capacity and compute deployment must ultimately converge on workable schedules.
Conclusion
Nvidia’s reported discussions with SB Energy demonstrate how the boundaries of the AI infrastructure market are changing. The accelerator supply chain can no longer be considered separately from the enormous physical infrastructure required to operate those accelerators.
If the investment proceeds, it would further connect one of the world’s most important AI computing suppliers with the development of the powered capacity required for future OpenAI infrastructure. Even before any transaction is confirmed, the talks underline an increasingly important reality for the data-centre industry: compute availability, power availability and infrastructure finance are becoming parts of the same strategic problem.
August 16, 2026 By: Joshua Anto