Mindstream Energy is developing a data center campus in Jordan that could scale to 400 megawatts, using dedicated natural gas-fired generation and closed-loop cooling to reduce its reliance on the country's power grid and water resources, the company said on Tuesday.
The 245-acre Sovereign AI & Digital Infrastructure Campus is being developed at Jordan's Al-Risha natural gas field, where Mindstream has a long-term gas contract.
The first phase is planned to provide 70 MW of commercial capacity, with operations targeted for March 2027.
The project comes as surging demand for AI computing puts pressure on electricity grids, particularly in the US and Europe, where developers face limited transmission capacity, rising power demand, water constraints, and lengthy permitting processes.
Mindstream said its model is to locate computing infrastructure near existing energy resources rather than add large loads to already constrained electricity networks.
Power for the Al-Risha campus will be generated onsite, allowing the facility to scale without drawing hundreds of megawatts from Jordan's consumer electricity grid, according to the company.
The project will also use closed-loop cooling systems intended to limit its dependence on water, Mindstream said.
Mindstream said it has secured land, contracted energy, on-site generation capabilities and government-approved investment and tax incentives as part of the project's development.
The company said it plans to deploy capacity as customers require it rather than build the full 400 MW in advance.
The modular approach is intended to allow Mindstream to tailor infrastructure to customers' computing requirements, including chip density, cooling systems and operating specifications.
Mindstream is also working with Orange Jordan to establish fiber connectivity linking the campus to regional and international networks, with routes toward Europe and India.
The company said that Jordan could use its domestic natural gas resources to support higher-value digital exports by converting energy into computing capacity.