AI power infrastructure newcomer SB Energy secretly files for IPO: SoftBank, OpenAI invest with deep ties to "Stargate"

On May 20th, SB Energy Corp., a digital infrastructure company owned by SoftBank Group, announced that it had confidentially submitted a draft registration statement Form S-1 to the U.S. Securities and Exchange Commission (SEC) to initiate the listing process. Based in Redwood City, California, this company evolved from CECEP Solar Energy and energy storage developers into a full-stack AI infrastructure platform spanning natural gas power generation, data center construction, and operation. Behind it is a market worth hundreds of billions of dollars driven by the global demand for AI computing power, as well as a "compute-electricity synergy" competition reshaping the global energy and digital infrastructure landscape. $1.8 billion in funding: SoftBank, OpenAI, and Ares join forces Part of SB Energy's IPO confidence comes from the massive funding it has completed in the past year. In January 2026, OpenAI and SoftBank Group each invested $500 million in SB Energy, injecting a total of $1 billion in equity capital and officially incorporating SB Energy into the "Stargate" plan. Prior to this, SB Energy had received $800 million in redeemable preferred stock financing from Ares Infrastructure Opportunity Fund in 2025, totaling more than $1.8 billion in funding. According to CB Insights data, SB Energy has raised a total of $2.4 billion in four rounds of financing, with investors including SoftBank, OpenAI, and Ares Management. Collaboration with OpenAI goes beyond just capital. OpenAI has selected SB Energy to build and operate its 1.2 gigawatt data center campus in Milam County, Texas, which is expected to be operational starting in 2026. OpenAI co-founder and president Greg Brockman said that this partnership "combines SB Energy's strengths in data center infrastructure and energy development with OpenAI's deep expertise in data center engineering," enabling "rapid, reliable compute expansion." SB Energy co-CEO Rich Hossfeld also noted that this collaboration will "accelerate our ability to deliver advanced AI data center campuses and related energy infrastructure at the scale required, drive the implementation of the Stargate plan, and ensure the future development of the U.S. AI industry." It is worth noting that SB Energy is not just a recipient of investment. As part of the deal, OpenAI, SoftBank, and SB Energy have established a non-exclusive preferred partnership, exploring a new model that integrates OpenAI's data center design capabilities with SB Energy's construction speed, cost control, and energy integration expertise. SB Energy further strengthened its end-to-end capabilities from development to operation by acquiring data center construction management company Studio 151. Ohio Super Project: 9.2 gigawatt natural gas power generation + 10 gigawatt data center What truly propels SB Energy to the forefront of AI infrastructure is its mega project in Piketon, Ohio. In March 2026, the U.S. Department of Energy announced a public-private partnership with SoftBank and SB Energy to build a 10 gigawatt data center and up to 10 gigawatts of accompanying power generation facilities at the former Portsmouth Gaseous Diffusion Plant in Piketon, now renamed as the "PORTS Technology Park," including at least 9.2 gigawatts of natural gas power generation units. According to information disclosed by the U.S. Department of Energy, SB Energy is partnering with AEP Ohio to invest $4.2 billion in grid upgrades and new transmission lines, with a clear commitment not to pass on costs to local residents' electricity bills. In addition, SB Energy has pledged to pay for accelerated environmental cleanup and restoration costs at the site, and has signed a $40 million community benefit agreement. The total investment for the project, $33.3 billion, including the natural gas power generation part, comes from Japan and is part of the U.S.-Japan Strategic Trade and Investment Agreement framework under the Trump administration. One gigawatt of power generation capacity can supply electricity to approximately 750,000 households at the same time, while the 10 gigawatt data center in the PORTS park consumes electricity equivalent to the power needs of about 7.5 million households. Energy Secretary Chris Wright said at the project launch ceremony that the project will "increase generation capacity, create jobs, and ensure that the United States wins the AI race"; Commerce Secretary Howard Lutnick positioned it as part of the "reindustrialization" of America through large-scale energy and infrastructure projects. However, this mega project is not without controversy. Before and after the launch ceremony on March 20th, 2026, some rural residents of Ohio have launched petitions, attempting to write "ban super-scale data centers" into the state's constitution for a referendum. This reflects the increasingly strong resistance to the excessive consumption of water and electricity resources by data centers across the United States. U.S. electricity dilemma: the era where data centers "cannot be built even with money" The reason why SB Energy's IPO story is worth a deep analysis lies in its business model, which hits the biggest structural pain point in global AI infrastructure development: the fact that power supply is becoming the actual limit to the expansion of computing power. In the United States, the situation of power supply is severe. At a Silicon Valley industry summit, several leaders in AI cloud computing and energy companies pointed out that the scale of AI computing clusters is experiencing exponential growth, while traditional public grid expansion speed has seriously lagged behind. The mismatch between supply and demand is becoming increasingly apparent, and the industry is accelerating its shift from relying on public grids to "behind-the-meter self-generation" and distributed energy models. In regions where the demand for computing power is concentrated, there is a significant gap in grid capacity, and data centers have queue times for grid connections of up to several years. "Some top tech companies have to adopt the 'on-site self-generated power' model, relying on diesel generators, gas gensets, and even second-hand power generation equipment to ensure continuous operation of computing power, and the power supply situation is severe." Data from the International Energy Agency (IEA) shows that global data center electricity consumption increased by 17% to 485 terawatt-hours in 2025, and is expected to double to around 950 terawatt-hours by 2030. The growth in energy consumption for training large AI models is even more astonishing - the Stanford AI Index report shows that from GPT-2 (2019) to GPT-4 (2023), the energy consumption for training has increased by about 500,000 times. At the same time, electricity costs already account for over 60% of AI data center operating costs, and the intense fluctuations in AI workloads present unprecedented challenges for energy quality management. SB Energy's unique narrative: a full-stack "compute-electricity integration" platform In this macro background, SB Energy's differentiated value becomes clear: it is not just a data center developer, but a full-stack platform integrating natural gas power generation, grid access, data center construction, and operation - In the new era where even with money, computing power cannot be built, this full-stack capability has extremely strong competitive barriers. SB Energy is not an exception. Meta Platforms recently disclosed that it has signed agreements to ultimately obtain over 6 gigawatts of nuclear power supply. Giant cloud providers like Microsoft, Google, Amazon, and others are accelerating their efforts to build their own or lock in dedicated power infrastructure. Behind this trend is a structural change in capital models: "A combination of distributed on-site power generation and large-scale real estate platforms is becoming a feasible path to bridge the supply-demand gap" (Data Center World 2026 conference conclusions). SB Energy's IPO is a landmark event where this trend transitions from industry practice to the capital market. From a financial perspective, SB Energy's IPO also faces indirect pressure from shareholders. SoftBank Group is seeking up to $40 billion in bridge loans to support its cumulative $64.6 billion equity investments in OpenAI (holding about 13%), which would be SoftBank's largest ever pure-dollar borrowing. Standard & Poor's downgraded SoftBank's credit outlook in March of this year, expressing concerns that its massive investments in OpenAI might harm liquidity and asset credit quality. Analysts point out that SoftBank's loan-to-value ratio is approaching S&P's 35% warning line, and market concerns about the continuous rise in its leverage are intensifying. Against this backdrop, pushing for an independent listing and financing of SB Energy will help inject new capital into SoftBank's AI infrastructure layout without further increasing its own debt. The timing of SB Energy's IPO listing still depends on market conditions and the SEC review process, and the company has not disclosed specific issuance size and valuation range. However, regardless of the final valuation, this company, which has transformed from a developer of CECEP Solar Energy into a full-stack AI infrastructure platform, has already become a core barometer for observing the "compute-electricity synergy" global race.