The scale of technological ambition announced by SoftBank this week in Ohio carries implications that extend well beyond American energy markets. A Japanese technology conglomerate is working to build a massive AI data centre on federally owned land in Ohio that it plans to power with roughly $33 billion worth of natural gas-fired electricity to be installed by the end of the decade. The project, situated at a former uranium enrichment site in Piketon, represents not merely a corporate infrastructure decision but a strategic geopolitical bet on who will dominate artificial intelligence capacity in the coming decade.
SB Energy, a SoftBank Group company, is planning to build a $33 billion natural gas plant with 10 gigawatts of new power generation—including 9.2 GW of natural gas generation—that will connect to the local grid and provide power to a new 10 GW data centre development built by SB Energy on site. For context, a single gigawatt of capacity can power roughly 750,000 homes at any given moment. The computing infrastructure itself will cost $30 billion to $40 billion.
What often goes unmentioned in coverage of such megaprojects is the strategic calculus here. The facility sits within a broader geopolitical framework: as part of a larger US-Japan trade deal, Japanese capital is flowing into American energy and data infrastructure at precisely the moment when access to computing capacity has become a matter of national competition. The competing forces reveal themselves when one examines who needs this power and why. Global artificial intelligence capacity is constrained not by algorithmic sophistication but by raw computational throughput, and throughput depends on reliable electrical supply. By anchoring Japanese investment in US soil with long-term energy commitments, SoftBank is simultaneously securing its own AI ambitions while deepening Japanese-American technological integration.
The natural gas choice, though predictable, deserves scrutiny. As of 2024, natural gas supplied over 40% of electricity for U.S. data centres, with renewables such as wind and solar supplying about 24% of electricity at data centres, while nuclear power supplied around 20%. The Piketon facility represents a deliberate commitment to dispatchable, controllable power rather than intermittent renewable sources. The company has sourced turbines for the gas project, with the first expected to be delivered within a year and the rest coming online by the end of the decade, with turbines capable of generating 9.2 gigawatts in total installed across the region rather than at a single complex. This represents industrial engineering at scale that matches the ambitions of the data centre itself.
The Ohio location is not incidental. Ohio ranks third nationally in data centre capacity with 1.6 GW currently and 2.4 GW planned, with more than 60% of the state's data centres in Columbus. The state sits within electricity grids whose operators already face mounting pressure. Residential electricity bills are expected to rise by $18 a month in western Maryland and $16 a month in Ohio as data centres reshape local energy economics. Yet from a fiscal perspective, the project commits significant private capital to infrastructure development that historically might have fallen to the state, and it creates immediate employment during construction and ongoing operational jobs.
The tension embedded in this decision is genuine. By 2030, data centre electricity consumption is projected to grow by 133% to 426 TWh. Meeting that demand through natural gas remains the most straightforward path, yet it locks in fossil fuel dependency at a moment when energy policy increasingly emphasises decarbonisation. Data centres emitted 105 million metric tons of carbon emissions in 2024, up from 31.5 million tons in 2018, representing a 300% increase as the number of data centres grew from 418 to 5,381 facilities between 2018 and 2024.
The reasonable counter-argument carries weight: attempting to power artificial intelligence development through renewables alone, given current battery storage limitations and transmission infrastructure constraints, would slow deployment and cede computing advantage to rivals less constrained by environmental commitments. The US has explicit national interest in maintaining technological parity with China, whose data centres already consume vastly more energy than American equivalents. From this perspective, the SoftBank project represents pragmatic infrastructure development in service of genuine national competition.
What is often overlooked in the public discourse is that this facility will join other gigawatt-scale projects progressing quickly in the US, with Amazon launching the first phase of its own mega-data centre in Indiana claiming to be the world's largest for now. The proliferation of such facilities raises legitimate questions about whether electricity grids, transmission infrastructure, and supply chains can support the deployment timeline these companies envision. The IEA warns that unless significant investments are made into transmission infrastructure, up to 20 percent of planned data centre projects could be at risk of delays, and generation equipment is likely to face supply chain issues.
The Piketon facility represents neither simple progress nor simple folly. It embodies the genuine complexity of industrial policy in an era when artificial intelligence infrastructure has become strategically equivalent to energy or military capacity. The commitment of Japanese capital to American electricity infrastructure, though denominated in private investment, carries strategic weight that extends beyond quarterly earnings reports. The choice of fossil fuel dispatchability over renewable intermittency reflects real engineering constraints, even if it offends climate sensibilities. The risk to local electricity markets is material and measurable. Yet the risk of American computational disadvantage relative to Chinese competitors is equally material.
Reasonable people can disagree on whether the trade-off serves long-term national interest. What cannot be reasonably disputed is that the stakes have become too large to treat as a purely commercial transaction. The electricity grid, once a utility matter of parochial concern, has become infrastructure for geopolitical competition. SoftBank's Ohio investment reveals that plainly.