Economy March 20, 2026

SoftBank Announces Plans for Large-Scale AI Campus in Ohio Fueled by Natural Gas

Project aims to deliver roughly 10 GW of computing demand with $33 billion of gas-fired generation and $30-$40 billion in data center hardware

By Ajmal Hussain
SoftBank Announces Plans for Large-Scale AI Campus in Ohio Fueled by Natural Gas

SoftBank Group is preparing to develop a major AI computing campus on federally owned land in Ohio, to be powered by an estimated $33 billion of natural gas-fired generating capacity and supported by $30-$40 billion in computing equipment. The planned facility is designed to draw about 10 gigawatts of power; turbines totaling 9.2 GW have been sourced and will be placed across the region, with the first arrivals expected within a year and full deployment by the end of the decade.

Key Points

  • SoftBank plans a major AI data center on federally owned land in Ohio at a former uranium enrichment complex; the facility is designed to draw about 10 gigawatts of power - sectors impacted: technology infrastructure and regional development.
  • The power component involves roughly $33 billion of natural gas-fired generation, with turbines totaling 9.2 GW sourced and to be installed across the region - sectors impacted: energy and utilities markets.
  • Project hardware costs for the data center, including chips and equipment, are estimated between $30 billion and $40 billion, indicating material implications for data center suppliers and capital markets - sectors impacted: enterprise IT hardware and cloud infrastructure.

SoftBank Group Corp. has laid out plans to build a substantial AI computing complex on federally owned land in Ohio, with the power backbone supplied by roughly $33 billion of natural gas-fired electricity scheduled to be installed by the end of the decade.

The company intends to locate the project at a former uranium enrichment complex under the ownership of the US Department of Energy. The proposed data center campus is engineered to draw approximately 10 gigawatts of power - a scale that the company has highlighted in public descriptions of the plan. For context supplied by the project brief, a single gigawatt of capacity can service about 750,000 homes at any one time.

SoftBank anticipates that the broader data center initiative - inclusive of the computing chips and associated equipment - will require capital in the range of $30 billion to $40 billion.

While US administration officials have previously referenced SoftBank’s $33 billion gas generation component as part of a wider $550 billion US-Japan trade initiative, this announcement provides the first explicit detail of the AI data center scheme and its intended technical footprint.

Rich Hossfeld, co-chief executive officer of SB Energy, the SoftBank-backed energy arm, said the company has sourced turbines for the gas project. According to the timeline provided, the initial turbine units are expected to be delivered within a year, and the remainder are slated to come online by the end of the decade.

The set of turbines procured is capable of producing a combined 9.2 gigawatts of generation capacity. Importantly, SB Energy said those turbines will be installed across the surrounding region rather than concentrated at a single central complex.

In addition to the turbines and regional generation build-out, SB Energy is planning an incremental 800 megawatts of new capacity specifically for the data center installation.


Taken together, the plans describe a multi-billion dollar effort that pairs large-scale gas-fired generation with a significant deployment of computing hardware. The information released so far outlines the broad financial and technical parameters - site, power draw, estimated equipment costs, turbine capacity, and staged delivery - while details beyond those elements have not been specified in the material provided.

Risks

  • Timeline uncertainty: turbine deliveries are described as starting within a year and completing by the end of the decade, creating execution risk for the project timeline - relevant to construction and energy sectors.
  • Cost range and scale: the data center equipment costs are provided as a broad $30 billion to $40 billion estimate, leaving financial scope and budget exposure uncertain - relevant to investors and suppliers in hardware and infrastructure.
  • Plan disclosure limitations: this is the first time detailed plans for the AI campus have been published, suggesting that further specifications or changes could follow as implementation progresses - relevant to regional planners and regulatory stakeholders.

More from Economy

What Would Unlock a Recovery in Home Improvement Demand? Mar 21, 2026 Powell lauds Volcker’s resolve as political pressure intensifies Mar 21, 2026 Power, Not Just Chips: How China Could Close the AI Compute Gap Mar 21, 2026 United Says Higher Oil Linked to Iran Conflict Will Strain Airlines, Plans Capacity Pullbacks Mar 21, 2026 Energy-Driven Cost Pressure Is Making Iron Ore Prices More Responsive to Oil Moves Mar 21, 2026