Jordan’s AI ambitions just got attached to a very large gas meter. Mindstream Energy has unveiled a plan for an AI-ready data-center campus at Al-Risha in Mafraq Governorate, built around dedicated natural-gas generation rather than a conventional dependence on the utility grid. The headline number is 400 megawatts of eventual capacity. The number that matters first is 70 MW, targeted for the second quarter of 2027.
I like that the company is at least starting with a defined first phase instead of waving its arms at some distant, foggy mega-campus. Still, this is a proposal with serious moving parts, not a finished infrastructure asset. Investors and prospective customers ought to keep both thoughts in their heads. AI compute demand is real; getting power, cooling, capital, equipment and paying customers lined up on schedule is where these grand digital-infrastructure stories usually earn their scars.
Power Is the Strongest Part of the Pitch
Al-Risha sits on a 250-acre site at Jordan’s natural-gas field, and Mindstream plans to use locally available gas for on-site generation. That is the core commercial proposition. Data centers chasing high-density GPU workloads need dependable electricity, and grid constraints have become a painfully familiar bottleneck in many markets. A dedicated, modular and off-grid setup could give a customer more certainty than waiting for a crowded transmission system to catch up.
That does not make fuel supply, generation equipment, operating economics or approvals disappear. It simply puts them at the center of the deal, where they belong. Mindstream cites a Natural Gas Supply Agreement with Jordan’s National Petroleum Company, alongside investment and tax incentives. Those are meaningful building blocks. They are not the same thing as a fully financed campus with installed turbines, commissioned data halls and contracted recurring revenue. There is a canyon between those milestones, and nobody should pretend otherwise.
The 70 MW Deadline Will Separate Substance From Sizzle
Q2 2027 is close enough to be useful as a yardstick. By then, watchers should expect evidence that the first deployment is progressing: capital arrangements, engineering and procurement progress, a clear path through construction, and ideally named customer commitments or credible capacity agreements. Mindstream says phased expansion will track demand, which is sensible. Building hundreds of megawatts on speculation has wrecked plenty of balance sheets over the years.
The preliminary estimate for initial infrastructure investment runs from roughly $150 million to $400 million. That is a huge range, and it tells you the design and commercial scope are still taking shape. Longer term, the company says cumulative infrastructure investment could exceed $1 billion, including solar photovoltaic components, subject to financing, demand and phased development. Compute investment could reach into the billions. Those are possibilities, not booked economics. The qualifiers matter more than the glittering totals.
Jordan Has a Real Opening, Not a Free Pass
There is a legitimate strategic angle here. Jordan wants a larger role in the digital economy, and a location serving Jordan, the broader MENA region and potentially Europe could attract attention if the project delivers reliable capacity at workable economics. The company also says it has relationships with 50 organizations across power, engineering, modular infrastructure, cooling, hardware, fiber, capital, risk management and possible offtake. That breadth is encouraging because no single developer can muscle through a project like this alone.
But relationships are relationships. They are not all binding contracts, and the announcement does not identify specific hyperscalers, neoclouds or enterprise customers that have committed to take capacity. For an AI campus, that distinction is not nitpicking; it is the whole ballgame. A gorgeous power plan without durable offtake can become an expensive monument to optimism.
“The site-and-power story is compelling, but the investment case will be written by execution and customer demand.”
What Deserves Close Attention From Here
- Whether the initial 70 MW remains on track for Q2 2027.
- How the broad $150 million to $400 million first-phase estimate narrows as design and financing mature.
- Whether prospective offtake turns into disclosed, credible customer commitments.
- How natural-gas supply, generation costs and cooling requirements translate into competitive delivered compute capacity.
- Whether expansion toward 400 MW follows actual demand rather than promotional momentum.
Mindstream is pointing at a problem the AI industry cannot wish away: advanced computing needs power, fast. Al-Risha could become a useful answer for Jordan and regional customers if the company turns its ecosystem, gas supply and modular plan into operating capacity. For the moment, though, this belongs in the promising-but-unproven file. The 70 MW launch is the first hard checkpoint, and hard checkpoints are where infrastructure dreams either grow teeth or lose them.