When Masayoshi Son makes a move, he does so with the kind of audacity that leaves even seasoned energy executives reaching for their calculators. SoftBank Group, the Japanese conglomerate known for its outsized bets on technology, has announced plans to spend approximately $33 billion constructing a massive natural gas power plant in the United States — a project that would rank among the largest single energy infrastructure investments the country has ever seen.
The announcement, first reported by TechCrunch, signals a dramatic escalation in the race to secure reliable electricity generation capacity as artificial intelligence data centers continue to consume power at rates that have caught utilities and grid operators off guard. SoftBank’s move is not merely an energy play — it is a vertically integrated strategy designed to ensure the company’s sprawling AI ambitions are never bottlenecked by something as fundamental as access to electricity.
The Scale of Son’s Energy Ambitions Dwarfs Industry Norms
To appreciate the magnitude of this investment, consider that the average cost of building a new combined-cycle natural gas power plant in the United States runs between $1 billion and $2 billion for a facility generating roughly 1,000 megawatts. SoftBank’s $33 billion price tag suggests a project of extraordinary scale — potentially capable of generating tens of thousands of megawatts, enough to power millions of homes or, more to the point, enough to feed the insatiable appetite of next-generation AI computing infrastructure.
The project fits within SoftBank’s broader commitment, announced during a high-profile January 2025 appearance alongside then-incoming President Donald Trump, to invest $100 billion in the United States over the coming years. That pledge, made under the banner of the Stargate AI infrastructure initiative — a joint venture with OpenAI and Oracle — was already one of the most ambitious foreign direct investment commitments in recent memory. The gas power plant represents roughly a third of that total commitment, underscoring just how central energy security has become to SoftBank’s technology strategy.
Why Natural Gas, and Why Now?
The choice of natural gas over renewable energy sources is both pragmatic and politically shrewd. While solar and wind generation have seen dramatic cost declines over the past decade, they remain intermittent — a fundamental limitation when the customer is a data center that demands 24/7 uptime with near-perfect reliability. Natural gas plants can be dispatched on demand and ramped up or down to match load, making them the preferred baseload and peaking power source for facilities where even brief outages can cost millions.
The decision also aligns with the current political environment in Washington. The Trump administration has been vocal in its support for domestic fossil fuel production and infrastructure development, rolling back permitting restrictions and signaling that natural gas will remain a cornerstone of American energy policy for decades to come. SoftBank’s investment effectively places the company on the favorable side of the administration’s energy agenda, potentially smoothing the path for the regulatory approvals and environmental permits such a massive project will inevitably require.
AI’s Voracious Power Appetite Is Reshaping the Energy Sector
The backdrop for SoftBank’s announcement is an energy demand crisis that few predicted even five years ago. According to the U.S. Energy Information Administration, electricity consumption from data centers is expected to more than double by 2030, driven almost entirely by the computational demands of training and running large AI models. Goldman Sachs has estimated that a single ChatGPT query consumes roughly ten times the electricity of a standard Google search, and as AI models grow larger and more complex, that ratio is only expected to widen.
Major technology companies have been scrambling to secure power. Microsoft has signed deals to restart a reactor at the Three Mile Island nuclear plant in Pennsylvania. Amazon Web Services has invested in nuclear energy projects and purchased a data center campus adjacent to a nuclear facility in Pennsylvania. Google has signed agreements with nuclear startup Kairos Power for small modular reactors. Meta has explored geothermal energy. Yet despite these efforts, the gap between projected AI power demand and available generation capacity continues to grow, creating what industry analysts have described as the most significant constraint on AI development outside of chip supply.
SoftBank’s Vertical Integration Play
What distinguishes SoftBank’s approach from that of its Big Tech rivals is the sheer verticality of the strategy. Rather than negotiating power purchase agreements with existing utilities or investing in third-party generation projects, Son appears intent on building and owning the generation capacity outright. This gives SoftBank direct control over one of the most critical inputs to its AI operations — a move that mirrors the logic behind Tesla’s decision to build its own battery factories or Apple’s long-running campaign to design its own semiconductors.
The Stargate initiative, which SoftBank is co-leading with OpenAI and Oracle, envisions a network of AI data centers across the United States with a combined investment of up to $500 billion over the next several years. The first phase of the project, centered on a facility in Abilene, Texas, is already under construction. Having a dedicated, company-controlled power supply for these facilities would give SoftBank a significant competitive advantage, insulating it from the volatile wholesale electricity markets that have already begun to see price spikes in regions with heavy data center concentration, particularly in Northern Virginia and central Texas.
The Financial Engineering Behind the Megaproject
Financing a $33 billion power plant is no trivial matter, even for a company with SoftBank’s resources. As of its most recent earnings report, SoftBank Group held a net asset value of approximately $190 billion, but the company also carries substantial debt and has historically relied on complex financial structures to fund its largest investments. The Vision Fund, SoftBank’s flagship investment vehicle, has experienced significant volatility, posting large losses in 2022 and 2023 before recovering somewhat in 2024 and 2025.
Industry observers expect SoftBank to pursue a combination of project finance debt, government incentives, and potentially co-investment from Stargate partners to fund the power plant. The Inflation Reduction Act, despite its name, contains significant provisions for energy infrastructure investment, including tax credits for certain gas-fired generation technologies that incorporate carbon capture. Whether SoftBank’s project will include carbon capture or other emissions-reduction technologies remains unclear, but doing so could unlock billions in federal subsidies and help blunt criticism from environmental groups that are certain to oppose a fossil fuel project of this magnitude.
Environmental and Political Headwinds Loom Large
The announcement has already drawn sharp reactions from climate advocates. A natural gas plant of this size would produce millions of tons of carbon dioxide annually, running counter to the emissions reduction targets that many states and corporations have adopted. Environmental organizations are expected to challenge the project through litigation and public pressure campaigns, arguing that the same investment could fund renewable generation and battery storage sufficient to meet AI’s power needs without the carbon footprint.
SoftBank will also face logistical challenges. Building a power plant of this scale requires enormous quantities of steel, turbines, and specialized construction labor — all of which are in high demand given the current boom in energy infrastructure investment. Supply chain constraints for gas turbines have already pushed delivery times for large-frame units from General Electric and Siemens Energy to three years or more. Securing a site with adequate natural gas pipeline access, water supply for cooling, and proximity to transmission infrastructure adds further complexity.
What This Means for the Broader Energy and Technology Industries
SoftBank’s $33 billion commitment is likely to accelerate an already intense competition among technology companies to lock down energy resources. It also sends a powerful signal to utilities, independent power producers, and policymakers: the era of abundant, cheap electricity that the U.S. enjoyed for much of the past two decades is over, at least in regions where data center demand is concentrated. Wholesale power prices in key markets have already risen 30% to 50% over the past two years, according to data from the Federal Energy Regulatory Commission, and projects like SoftBank’s are both a response to and a driver of that trend.
For the natural gas industry, the announcement is an unambiguous vote of confidence. U.S. natural gas production has hovered near record levels, and the addition of a massive new source of demand could help sustain prices and investment in upstream drilling and pipeline infrastructure for years to come. For renewable energy developers, the message is more complicated: while long-term decarbonization trends remain intact, the near-term reality is that gas-fired generation is winning the race to power AI.
Masayoshi Son has built his career on bold, sometimes reckless bets — from the early investment in Alibaba to the WeWork debacle to the massive AI infrastructure commitments of the past year. The $33 billion gas power plant is perhaps his most tangible wager yet: a physical monument to the conviction that artificial intelligence will reshape the global economy, and that the companies controlling the power behind it will be the ones that ultimately prevail. Whether the bet pays off will depend not just on the trajectory of AI demand, but on SoftBank’s ability to execute a construction project of historic proportions on time and on budget — a challenge that has humbled far more experienced energy companies than a Tokyo-based technology conglomerate.