Data centers are becoming one of the fastest-growing water consumers in the world. AI workloads, high-density chips, and rising digital demand are pushing cooling systems far beyond traditional design limits. Most facilities still depend on evaporative cooling, a process that can use millions of gallons of water each year and create significant environmental pressure in water-stressed regions.
A new generation of startups is working to change this. By advancing liquid cooling, microfluidics, modular design, and closed-loop thermal systems, these innovators are dramatically reducing water use while meeting the performance needs of modern compute. Their solutions address both sides of the challenge: how to extract heat efficiently without evaporating water, and how to support the next wave of AI infrastructure without expanding the industry’s water footprint.
As AI workloads surge, these technologies are becoming essential for sustainable data center expansion. This article highlights startups innovating to reduce data center water consumption through next-generation cooling technologies.
Corintis (Switzerland)
Corintis advances chip-level cooling for data centers and high-performance computing. Founded in 2022 by Remco van Erp and Sam Harrison, the company develops silicon-based microfluidic systems that remove heat at unmatched efficiency. Its technology embeds microscopic liquid channels inside the chip, enabling heat extraction up to ten times higher than leading alternatives and doing so with more than fifty times greater energy efficiency.
This approach addresses a core constraint in semiconductor progress. Rising heat density now limits chip performance and drives excessive energy use across data centers. Corintis integrates cooling directly into chip architecture, allowing reliable operation at lower temperatures while reducing data center water consumption.
The company also builds advanced cold plates through additive manufacturing. These components use channels as narrow as 70 micrometers to target hot spots with high precision. Corintis enhances performance through bio-inspired designs created with its proprietary simulation tools. In trials with Microsoft, the system delivered three-times-higher heat-removal efficiency than existing solutions and reduced chip temperatures by more than 80 percent.
Corintis raised $24 million in Series A funding in September 2025.
Flexnode (USA)
Flexnode builds prefabricated, liquid-cooled modular data centers designed for high-density workloads and rapid deployment. The company minimizes water use by integrating advanced cooling systems, including a manifold microchannel heatsink, a hybrid immersion architecture, and an additive-manufactured dry-cooling heat exchanger. These systems allow Flexnode to operate in extreme environments without relying on traditional water-intensive cooling.
Flexnode delivers a deeply integrated hardware stack that combines power, cooling, compute, and container design through vertically integrated manufacturing. Its solution reduces cost and time to deployment. Digital-twin capabilities and big-data analytics support real-time monitoring, predictive maintenance, and performance optimization.
Flexnode is collaborating with the U.S. Department of Energy’s Advanced Research Projects Agency-Energy’s COOLERCHIPS program to advance next-generation cooling technologies. The company raised $9 million in Seed funding in February 2024.
Submer (Spain)
Submer, founded in 2015 in Barcelona by Daniel Pope and Pol Valls, develops immersion-cooling systems that submerge servers in a non-conductive liquid. This eliminates traditional air-cooling infrastructure, cuts energy use, and supports higher-density compute. The approach also reduces water consumption and enables compact, modular deployments for AI, high performance computing, and edge environments.
The company positions immersion cooling as a path to sustainable, future-ready digital infrastructure. Its technology reduces heat transfer losses, improves efficiency, and lowers data center water consumption. Submer has become a leading player in the immersion-cooling segment through rapid scaling and continuous R&D. Submer raised $20 million in debt financing in December 2024.
Crusoe (USA)
Crusoe is building large-scale AI data centers using an energy-first strategy that pairs high-density compute with ultra-efficient liquid cooling and low-carbon power. The company uses closed-loop, direct-to-chip (DTC) cooling to keep GPUs operating at peak efficiency, even under extreme temperature swings. This approach avoids the water loss associated with evaporative cooling towers and aligns with rising environmental and regulatory pressures.
At Crusoe’s Abilene campus in West Texas, the cooling system used in the 1.2GW data center is fully closed-loop and non-evaporative, recirculating the same water rather than consuming it. Heat is rejected through air-cooled chillers, allowing water to recirculate rather than being consumed. Annual water use is expected to be just 12,625 gallons per building, roughly 10% of the water footprint of a typical U.S. household. The site is engineered for high-energy-density AI hardware and integrates renewable wind power to reduce emissions.
Operationally, Crusoe selected closed-loop DTC cooling over immersion systems due to its superior serviceability, lower water footprint, and compatibility with modular maintenance at hyperscale. The design supports rapid hardware refresh cycles and reduces downtime—critical for large AI training clusters.
Crusoe raised $1.4 billion in Series E funding in October 2025 to accelerate its vertically integrated AI infrastructure and Crusoe Cloud platform.
Firmus Technologies (Australia)
Firmus Technologies is a vertically integrated developer and operator of AI infrastructure with a mission to build the most energy-efficient data centers. The company began designing systems for 1,000+ GPUs as early as 2019 and developed its AI Factory platform to address the energy and cooling demands of large-scale AI workloads.
A core advantage is its breakthrough in liquid-cooling efficiency. Firmus uses a synthetic, non-conductive cooling fluid, reducing waste and eliminating frequent replacement. Standard 19” rack servers are immersed directly in this fluid within purpose-built immersion racks, enabling far more efficient heat removal than air cooling and offering a practical pathway to carbon-aware computing.
Firmus builds AI Factories that combine immersion and direct-to-chip cooling with density-optimized facility design. Powered by its FactoryOS software, these facilities integrate compute, cooling, and energy management into a unified platform. The company supports both retrofits of legacy data centers and new high-density builds, while also offering managed infrastructure and GPU cloud services.
Firmus raised $327 million in Series B funding in November 2025 to scale capacity to 1.6 GW by 2028 under Project Southgate, developed in collaboration with CDC Data Centres and NVIDIA.
Scaling innovation to reduce data center water consumption
For digital infrastructure to grow sustainably, data centers must adopt cooling systems that can manage extreme heat loads without straining local water resources.
The startups featured in this article are redefining what next-generation cooling can deliver. By advancing liquid cooling, immersion systems, and heat-recovery technologies, they are reducing dependence on freshwater while supporting high-density compute.
Sustained investment and continued technical breakthroughs are essential to accelerate adoption. These measures can help the sector reduce data center water consumption while meeting the rising demand for AI and advanced computing.
Curious to explore more startups accelerating innovation toward water-efficient digital infrastructure?
Book a call to start your free trial and access our Data Centers Water Consumption Market Snapshot.
Already a customer? Log in here.
