NVIDIA Corporation (NVDA.US) plans to invest $3 billion to enter the SB Energy market: from selling chips to securing the "power-park-computing" chain with OpenAI.

According to The Information, NVIDIA is in talks with SB Energy, a subsidiary of SoftBank, to potentially invest up to $3 billion; the latter is developing a large data center project for OpenAI in Ohio. If the deal is finalized, its significance would not merely be a financial investment, but rather NVIDIAs continued evolution from an AI chip supplier to a capital participant in AI infrastructure projects. Currently, details remain at the level of negotiations and up to $3 billion, with specifics on the investment proportion, fund usage, project scale, power supply solutions, chip procurement constraints, and whether it will be included in the Stargate system yet to be disclosed. Analysts suggest that NVIDIA could leverage its investment to engage with OpenAIs data center developer early on in a critical segment that could determine whether GPU orders will be fulfilled. This aligns with NVIDIAs recent strategies surrounding OpenAI. The cooperation between the two dates back to 2016. When OpenAI unveiled its Stargate initiative in 2025, it explicitly mentioned a deep collaboration with NVIDIA since 2016. In January 2025, Stargate announced plans to invest $500 billion in AI infrastructure in the U.S. over four years, with SoftBank, OpenAI, Oracle, and MGX as initial equity investors; SoftBank handling finances, OpenAI managing operations, and NVIDIA listed as a key technology partner alongside Arm, Microsoft, and Oracle. At that time, NVIDIA transcended being just a supplier, stepping into the core circle of computational system design, deployment, and operational coordination. The true capitalization of this relationship occurred in September 2025 with the signing of a memorandum of cooperation. NVIDIA and OpenAI announced plans to deploy at least 10 gigawatts of NVIDIA systems, corresponding to millions of GPUs; NVIDIA plans to gradually invest in OpenAI with each gigawatt system deployed, potentially reaching up to $100 billion, with the first gigawatt expected to be operational in the second half of 2026 on the Vera Rubin platform. The announcement also stated that OpenAI would prioritize NVIDIA as its strategic computing and networking partner in AI factory expansion. In other words, NVIDIA is trying to create a closed loop of investment supporting infrastructure constructioninfrastructure driving GPU deploymentmodels and software optimizing hardware routes in reverse. The potential investment in the SB Energy project can be seen as an extension of this closed loop upstream. The bottlenecks in large model training and inference competition are no longer just chip supply but also include land, grid connection, power equipment, cooling systems, park construction, and project financing. For NVIDIA, even if GPU performance continues to improve, if data centers cannot secure stable power or if construction timelines are delayed due to funding and approvals, significant orders may struggle to convert into revenue as planned. Investing in developers like SB Energy could help it lock in deployment scenarios earlier and reduce the risks of client expansion being constrained by power and park construction. SoftBank also has a unique position in this context. As the financial leader of the Stargate initiative, SoftBank needs to break down the grand commitments to computing power into specific, financeable, and buildable projects. If NVIDIA is willing to take equity in energy or park platforms, it could not only introduce industry-anchored capital into projects but also boost financial institutions' confidence in subsequent equipment purchases and long-term leases. On August 10, NVIDIA announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to establish an independent financing platform for computing power, aimed at mobilizing more than $500 billion in third-party capital. This indicates that its strategy has shifted from selling high-value hardware to promoting the entire AI factory as an investable asset. However, this model also carries risks. First, NVIDIAs investments in key customers and their infrastructure could raise market concerns about a cyclical growth wherein capital input nourishes chip orders. Second, a 10-gigawatt project is highly sensitive to grid infrastructure, transformer capacities, construction capabilities, and financing costs, which may lead to delays in timelines. Third, OpenAI is not solely dependent on NVIDIA and will continue to balance supply risks through diversified hardware and cloud collaborations. Fourth, regulators may increasingly pay attention to the growing interconnections among chip leaders, model companies, cloud platforms, and energy assets. Therefore, what makes the $3 billion rumor regarding SB Energy most noteworthy is not the amount in itself, but whether NVIDIA is pushing the boundaries of competition towards power and data center delivery capabilities.