Summary
- AirJoule Prime is planned for deployment at a commercial data centre operated by a Net Zero Innovation Hub member during 2026.
- The system is intended to use low-grade waste heat from cooling systems to produce usable water.
- The trial tests an alternative heat-reuse route for facilities where district heating networks are unavailable or impractical.
A European data centre is planned to host a trial that will use low-grade waste heat from cooling systems to produce usable water, testing an alternative to the district-heating schemes normally associated with heat reuse.
AirJoule Technologies has agreed a strategic collaboration with the Net Zero Innovation Hub for Data Centers to deploy and validate its AirJoule Prime system at a commercial facility operated by one of the hub’s members.
The deployment is planned during 2026. The companies have not yet publicly identified the participating data centre.
AirJoule’s system is designed to use thermal energy to extract water from air. In a data centre application, the company says low-grade heat rejected by air- or liquid-cooling systems can provide an energy input for the process, creating distilled water rather than simply discharging the heat to the atmosphere.
The project sits within the hub’s wider Heat-to-X work, which is looking at ways to turn data centre waste heat into useful outputs where conventional heat networks are not available.
Heat reuse has become an increasingly prominent part of European data centre policy and planning, but its practicality varies sharply by location. A facility can produce large volumes of heat while still lacking an adjacent user, suitably configured district network, or stable year-round demand capable of taking it.
Temperature is another constraint. Data centre cooling systems often reject heat at relatively low temperatures, meaning additional equipment such as heat pumps may be required before that energy can be fed economically into some district-heating networks.
Turning part of that thermal output into water represents a different engineering proposition. Rather than depending on a nearby external heat customer, the process could potentially produce a utility that is useful on or around a site.
The attraction is particularly clear as cooling technology changes. Higher-density AI hardware is accelerating deployment of direct-to-chip liquid cooling and other thermal systems that move heat more efficiently away from processors. That can make the heat easier to capture at a defined point in the cooling loop, although its final temperature and usefulness depend on the facility design.
AirJoule says its technology can work with low-grade waste heat from both air-cooled and liquid-cooled data centres. The planned trial is intended to test that proposition inside an operating commercial environment rather than a standalone demonstration.
The water angle also intersects with a separate pressure on the sector. Data centres are facing growing scrutiny over water consumption, particularly where evaporative cooling systems operate in areas exposed to drought or competing municipal demand.
A heat-to-water system would not automatically make a facility water neutral. The volume of water recovered, parasitic energy use, equipment cost, maintenance requirements, local humidity, cooling design, and quality of the produced water all affect the real operating case.
Those are precisely the questions a commercial deployment needs to answer. Technologies that look efficient at component level can impose additional pumps, fans, heat exchangers, controls, or electrical loads once integrated into a live data centre.
The Net Zero Innovation Hub was established to test technologies under real operating conditions and help move systems from demonstration towards deployment. AirJoule previously participated in its start-up programme before moving to the planned site trial.
The project also broadens the definition of heat reuse. European planning discussions often frame the issue around exporting heat to homes, offices, or industrial users, but not every facility sits close enough to a suitable network for that model to work.
A process that can create another useful resource on site could provide an option for locations where heat export would otherwise require expensive new pipework or would face weak seasonal demand.
The planned trial is intended to show how much water can be produced from a given amount of rejected heat, what additional energy the process consumes, and whether the economics remain credible alongside the primary job of maintaining safe temperatures for IT equipment.
