Summary
- Ecolab has closed its approximately $4.75bn acquisition of direct liquid cooling supplier CoolIT Systems.
- The company plans to combine CoolIT cold plates and CDUs with fluids, monitoring, and 3D TRASAR optimisation for AI data centres.
- The deal shows liquid cooling moving from specialist equipment supply into larger water, power, and operations platforms.
Ecolab has completed its approximately $4.75bn acquisition of CoolIT Systems, bringing a major direct liquid cooling supplier into its high-tech water, process, and digital optimisation business.
The transaction adds CoolIT’s cold plates, cooling distribution units, and direct liquid cooling expertise to Ecolab’s existing work across ultra-pure water, industrial water systems, cooling fluids, and monitoring.
Ecolab said CoolIT’s year-to-date sales had grown by more than 100%, driven by demand for liquid cooling in AI data centres. The company now plans to launch an end-to-end 3D TRASAR cooling platform at Supercomputing in Chicago in November 2026.
The planned platform will combine CoolIT CDUs and cold plates with Ecolab digital optimisation and advanced cooling fluids. Ecolab said the system is designed to give operators real-time visibility into cooling performance, reduce cooling power demand, improve power efficiency, and support closed-loop approaches that reduce water use.
The deal also changes Ecolab’s commercial scale in high-tech infrastructure. Its Global High-Tech business generated about $150m in annual sales in 2021 and is now approaching $1.5bn in 2026 annualised sales after the additions of Ovivo and CoolIT. Ecolab is targeting $4bn of annual sales in the business by 2030.
Liquid cooling becomes a platform
Direct liquid cooling has moved from engineering option to capacity enabler for AI infrastructure. The question is no longer whether the technology can remove heat from dense chips. The harder work sits in deployment, serviceability, fluid quality, leak control, monitoring, and integration with existing plant.
CoolIT gives Ecolab hardware at the point where chip heat enters the facility system. Cold plates, manifolds, and CDUs connect directly to water treatment, cooling fluids, pumps, heat rejection, digital monitoring, and maintenance routines. That makes the acquisition more than a bolt-on product deal.
AI facilities are changing the relationship between IT hardware and building services. GPU platforms, liquid loops, high-capacity electrical systems, and cooling controls must be planned together because a thermal issue in a dense AI hall can escalate quickly and interrupt expensive compute utilisation.
Closed-loop cooling also changes the sustainability argument. Water use is under rising scrutiny in European markets, especially where data centres are proposed in water-stressed or politically sensitive locations. Direct liquid cooling can reduce evaporative dependence in some designs, but the final environmental case depends on heat rejection, pump energy, fluid management, water treatment, and heat reuse options.
Supplier boundaries narrow
The acquisition arrives as cooling suppliers, electrical equipment manufacturers, chip companies, and operators converge around AI-ready data centres. Customers want fewer gaps between IT hardware, facility systems, and operational support. That favours suppliers that can provide integrated packages, global service coverage, and engineering support through design, commissioning, and operation.
Platform claims still have to survive site-level complexity. Liquid cooling systems must fit rack layouts, redundancy choices, manifold routes, sensor strategies, commissioning procedures, and staff capabilities. Retrofitting existing sites adds another layer because operators may have to introduce liquid systems into buildings designed around air cooling.
For European operators, the deal strengthens one of the supplier routes into AI-ready cooling. Power density, water reporting, local permitting, and the risk of stranded capacity are already pushing cooling decisions up the investment agenda.
Ecolab has bought a position at the boundary between silicon heat, data centre plant, water risk, and AI capacity delivery. That boundary is becoming one of the most contested layers of the facility stack.
