Waterless Liquid Cooling Market Size, Share & Growth

The global waterless liquid cooling market is valued at approximately USD 2.30 billion in 2025 and is projected to reach USD 12.02 billion by 2032, expanding at a CAGR of 26.5% between 2026 and 2032. The surge is inseparable from one force: artificial intelligence. As GPU power draw climbs past what air and even water-glycol loops were designed to handle, and as data centers face mounting pressure to stop consuming millions of liters of water a day, operators are turning to cooling that removes heat with engineered dielectric fluids instead of water. Waterless liquid cooling — spanning two-phase direct-to-chip systems, single- and two-phase immersion, and emerging in-chip microfluidics — is moving from pilot racks to production AI factories, and the market's trajectory reflects that shift.

Top 10 Key Takeaways

• North America holds the largest base, anchored by hyperscale AI-factory build-outs and the densest cluster of waterless-cooling innovators, from ZutaCore and Accelsius to LiquidStack.
• Asia Pacific is the fastest-growing region, driven by China's immersion deployments, India's data center expansion, and strict power-usage-effectiveness targets across Beijing, Singapore, and Japan.
• Two-phase direct-to-chip is the highest-momentum cooling technology, prized for its ability to cool the hottest AI accelerators without water anywhere near the servers.
• Cloud and hyperscale providers are the leading vertical, because they own the densest racks and the most acute water and energy exposure.
• The decisive technology shift is dielectric-fluid heat capture replacing evaporative water and, increasingly, water-glycol loops at the rack.
• Water-usage-effectiveness and power-usage-effectiveness mandates are the dominant regulatory force, especially in water-scarce and PUE-capped jurisdictions.
• Key companies span thermal majors and specialists — Vertiv, Schneider Electric, Carrier, and Daikin alongside ZutaCore, Accelsius, LiquidStack, Submer, GRC, and Iceotope.
• The near-term opportunity is retrofit and cooling-as-a-service, letting operators adopt waterless systems without rebuilding white space.
• The near-term risk is vendor lock-in and immature interoperability standards, which slow multi-vendor procurement.
• The strategic implication is clear: thermal management has become a gating constraint on AI scale, and waterless cooling is now a board-level infrastructure decision rather than a facilities footnote.

Why the Waterless Liquid Cooling Market Matters Now

For two decades, data centers cooled themselves with air and water. Chilled-air handlers pushed cold air across racks; evaporative cooling towers rejected heat to the atmosphere by evaporating enormous volumes of water. That model worked while a rack drew ten or fifteen kilowatts. It does not work for AI. A single rack of next-generation accelerators now draws sixty to eighty kilowatts, far beyond the roughly twenty-five to thirty kilowatts that air can practically remove. At the chip level, the arithmetic is even starker: GPU thermal design power has roughly doubled with each generation, and the industry is already planning for accelerators that draw several thousand watts apiece. Air simply cannot carry that heat away fast enough.

Water-based liquid cooling answered part of the problem, and single-phase direct-to-chip systems using water-glycol mixes remain the workhorse of the transition. But water introduces its own liabilities. It conducts electricity, so a leak near live silicon is a catastrophe. It requires treatment, filtration, and — in evaporative designs — staggering consumption at a moment when regulators, communities, and investors are scrutinizing every liter a data center draws. This is the opening the waterless liquid cooling market has stepped into. By circulating non-conductive dielectric fluids that boil or absorb heat directly at the chip, or by submerging entire servers in engineered fluids, waterless systems remove heat with no water touching the IT and, in closed-loop designs, no evaporative water at all.

The timing is not incidental. Digital transformation, the generative-AI boom, and sustainability regulation have converged. Hyperscalers are committing to water-positive operations, governments are capping power usage effectiveness, and the cost of the electricity spent on cooling has become impossible to ignore — cooling can consume nearly half of an AI facility's power budget without efficiency gains. Waterless liquid cooling sits at the intersection of all three pressures, which is why it has moved so quickly from curiosity to procurement priority. For a broader view of the parent category, see [INTERNAL LINK: Data Center Liquid Cooling Market]; for the compute demand driving it, see [INTERNAL LINK: AI Data Center Market]; and for the adjacent thermal category, see [INTERNAL LINK: Data Center Cooling Market].

What makes this moment distinct from earlier cooling transitions is that the decision has climbed the organizational ladder. A decade ago, cooling was a facilities matter, negotiated between site engineers and equipment vendors and largely invisible to the executives setting corporate strategy. Today, because thermal capacity directly limits how much AI compute a company can stand up, cooling architecture is discussed in the same conversations as chip supply, power procurement, and site selection. Chief technology officers, heads of infrastructure, and sustainability leaders now weigh in on choices that were once left to the mechanical team. This elevation matters commercially: it lengthens sales cycles, raises the stakes of vendor selection, and rewards suppliers who can speak to business outcomes — capacity unlocked, water saved, energy reclaimed — rather than component specifications alone. The waterless liquid cooling market is therefore not only growing in size but maturing in the way it is bought and sold, and that maturation is itself a signal of how central the technology has become to the AI build-out.

Market Trends

The most visible trend is the migration away from water in the cooling loop entirely. For years the debate was air versus liquid; now, within liquid, the debate is water versus waterless. Vendors that built their reputations on water-glycol direct-to-chip are being challenged by two-phase specialists arguing that dielectric fluids scale better as chip power climbs. ZutaCore, whose HyperCool technology uses a dual-phase pool-boiling approach with no water, has positioned itself explicitly around cooling the hottest AI GPUs while eliminating water risk. Accelsius makes a parallel case with its NeuCool two-phase system, in which a low-pressure dielectric refrigerant boils on contact with the chip and no water enters the rack.

A second trend is the pursuit of ever-higher per-chip cooling capacity through phase change. Single-phase systems have publicly demonstrated cooling in the range of one thousand watts per chip; two-phase advocates argue that only the latent heat of vaporization can keep pace with accelerators headed toward several thousand watts. This is why the frontier of the market is defined less by incremental efficiency and more by thermal headroom — the ability to cool a chip that does not yet ship.

A third trend is standardization. Immersion and two-phase systems have historically been proprietary, with tanks, fluids, and coolant distribution units that do not interchange across vendors. That is beginning to shift as fluid suppliers publish qualified specifications and as reference architectures from chip vendors and hyperscalers push the ecosystem toward common interfaces. Shell's launch of its dielectric coolant designed for both direct-to-chip and immersion use is one signal that the fluid layer is maturing into a supply chain rather than a set of one-off formulations. See [INTERNAL LINK: Immersion Cooling Market] for the deployment models this enables.

Sustainability design targets form a fourth trend. The "zero-water" claim has become a competitive differentiator, particularly in water-stressed regions, and the ability to pair waterless cooling with heat reuse — capturing rejected heat for district heating or industrial processes — is increasingly part of the sales conversation. In Europe especially, where regulators and municipalities scrutinize the environmental footprint of new data center capacity, the promise of exporting rejected heat to nearby buildings or industrial users can be the difference between a permit granted and a project stalled. Waterless systems, because they concentrate heat in a fluid loop rather than dissipating it into evaporating water, lend themselves naturally to this kind of recovery, and vendors are beginning to design for it from the outset rather than bolting it on afterward.

A fifth trend is the emergence of cooling-as-a-service and managed models. Many operators, particularly enterprises and smaller colocation providers, are wary of taking on the operational burden of an unfamiliar cooling technology. Service-led offerings — in which a vendor or integrator owns the cooling system and guarantees a thermal outcome — lower that barrier and convert a daunting capital project into a predictable operating expense. This shift mirrors the broader servitization of infrastructure and is likely to accelerate adoption among buyers who want the performance of waterless cooling without the learning curve.

Finally, supply chain localization is rising up the agenda as tariffs and trade friction push manufacturers toward regional production of cooling hardware. The 2025 US tariff environment and broader geopolitical tension have made operators and vendors alike more sensitive to where coolant distribution units, tanks, and fluids are made. Localization reduces exposure to trade disruption and shortens lead times at a moment when every week of delay can strand expensive, idle GPUs, and it is prompting cooling suppliers to build or partner for manufacturing closer to their largest demand centers.

Market Drivers

The primary driver is thermal necessity. AI accelerators have crossed the threshold where air cooling is not merely inefficient but physically inadequate, and each new GPU generation raises the ceiling. When a chip draws well over a kilowatt and racks exceed sixty kilowatts, operators do not choose liquid cooling for elegance — they adopt it because the hardware will not run otherwise. Waterless approaches extend that logic by promising headroom for the chips still on the roadmap, which is why hyperscalers and neocloud operators are qualifying two-phase and immersion systems now rather than later.

Water scarcity and regulation form the second driver. A large data center using evaporative cooling can consume enormous volumes of water, and that consumption has become a flashpoint in communities from the American Southwest to the Middle East. Microsoft has publicly committed to becoming water-positive, replenishing more water than it consumes, and is piloting zero-water cooling designs in Arizona and Wisconsin. Where water-usage-effectiveness reporting is required and increasingly enforced, waterless cooling is often the only high-density option that complies. This regulatory tailwind is strongest precisely where the largest new capacity is being built.

Energy cost and efficiency constitute the third driver. Cooling can consume a substantial share of a data center's total power, and every point of power-usage-effectiveness improvement translates into real operating savings at hyperscale. Waterless two-phase systems have demonstrated facility PUE figures approaching the theoretical floor, and vendors point to double-digit reductions in cooling energy against conventional setups. In an era of soaring electricity prices and grid-connection bottlenecks, cooling efficiency is no longer a sustainability nicety — it is a capacity enabler, because power saved on cooling is power available for compute.

The fourth driver is the sheer scale of capital flowing into AI infrastructure. Hyperscale and AI-factory build-outs are proceeding at unprecedented pace, and cooling is a line item in every one of them. Strategic investors have taken notice: Carrier's venture arm has invested in and then expanded its stake in ZutaCore, and ZutaCore itself raised significant new funding to scale deployments. Capital is validating waterless cooling as core infrastructure, not experimental technology. The involvement of climate and thermal majors as investors — rather than pure financial backers — is especially telling, because it signals that incumbents view waterless cooling as strategically important enough to embed in their own product roadmaps. See [INTERNAL LINK: Direct-to-Chip Cooling Market] for the fastest-scaling segment of this build-out.

A fifth, quieter driver is the growing recognition among operators that cooling choice affects hardware longevity and reliability. High and fluctuating chip temperatures accelerate component wear, and by holding silicon at more stable, lower temperatures, waterless systems can extend the useful life of extremely expensive accelerators and reduce failure rates. When a single AI server can cost as much as a luxury car, the reliability argument reinforces the thermal and efficiency arguments, and it resonates with finance teams evaluating total cost of ownership rather than upfront price alone. This convergence — thermal necessity, water and regulatory pressure, energy economics, capital availability, and hardware protection all pointing the same direction — is what gives the market its unusual momentum. Rarely do so many independent forces align behind a single infrastructure shift.

Market Challenges and Restraints

The most cited restraint is cost. Waterless systems — particularly immersion and two-phase — carry higher upfront capital than air cooling, and the dielectric fluid itself can be expensive, especially the fluorocarbon fluids used in some two-phase designs. For greenfield hyperscale AI facilities the payback is compelling at high rack densities, but for enterprises running moderate loads the economics are less obvious, and the capital hurdle slows adoption in the mid-market.

Retrofit complexity compounds the cost challenge. Full-room immersion can require structural reinforcement and fluid containment, and even where retrofit-ready two-phase direct-to-chip systems exist, integrating them into facilities built for air demands careful planning. Operators weigh the disruption of converting live white space against the risk of building capacity they cannot cool.

Fluid supply and regulatory scrutiny form a third headwind. Some engineered dielectric fluids face environmental questions, and shifting regulation around per- and polyfluoroalkyl substances has already reshaped the supplier landscape, prompting at least one major chemical producer to exit certain fluorinated fluids. Buyers must now weigh long-term fluid availability and compliance alongside thermal performance.

Two operational challenges round out the picture. First, vendor lock-in: tanks, fluids, and distribution units are frequently not interchangeable, so choosing a vendor can mean committing infrastructure for years, and the immersion segment remains concentrated among a handful of suppliers. This concentration cuts both ways — it limits pricing pressure and raises switching costs, but it also means the financial stability of a chosen vendor becomes an operational risk, because a supplier's failure could strand assets that cannot easily be serviced by anyone else. Procurement teams increasingly diligence vendor balance sheets alongside thermal performance, and the lack of common standards remains one of the clearest brakes on faster adoption. Second, fire safety and serviceability: immersion tanks containing combustible dielectrics raise fire-code considerations, and servers bonded into fluid systems can be harder to service than socketed, air-cooled parts. Fire codes may demand enhanced suppression, greater separation distances, or limits on fluid quantities per zone, all of which add cost and design complexity. These are solvable problems, but they are real friction in the buying process.

A further, more subtle challenge is organizational rather than technical: the shortage of talent and institutional knowledge around waterless cooling. Facilities teams trained for decades on air and chilled-water systems must learn new operating procedures, new failure modes, and new maintenance routines. The skills gap slows deployment, raises the perceived risk of adoption, and reinforces the appeal of managed-service models. Until waterless cooling is as familiar to operations staff as a computer-room air handler, a meaningful share of the market will hesitate, and vendors that invest in training, documentation, and support will hold an advantage disproportionate to their hardware alone.

Industry and Application Growth

Growth is not evenly distributed across verticals — it concentrates wherever rack density and resource exposure are highest. Cloud and hyperscale providers lead by a wide margin, because they operate the densest AI clusters and carry the most acute water and energy scrutiny. For these operators, waterless cooling is a strategic capability tied directly to how much compute they can deploy per site.

IT and telecommunications is a fast-rising vertical as carriers bring megawatt-scale compute into their networks for AI workloads; ZutaCore has explicitly courted telecom providers under pressure to cut energy use while adding AI capacity. Telecom operators occupy an interesting position, because they must place compute close to users — at the edge, in central offices, and in facilities never designed for high density — where closed-loop waterless systems that need no external water infrastructure are especially valuable. As networks evolve to support AI inference at the edge, the pressure on these constrained sites will only grow, and waterless cooling is one of the few options that fits.

Research, academia, and high-performance computing form another natural stronghold, as supercomputing centers have long pushed thermal limits and are early adopters of immersion and two-phase systems. These institutions often run the densest sustained workloads in existence and have both the technical sophistication and the mandate to experiment, which makes them proving grounds for approaches that later diffuse into commercial data centers. BFSI and healthcare and life sciences are adopting waterless cooling as their AI and analytics workloads densify — the former for risk modeling, fraud detection, and algorithmic trading, the latter for genomics, drug discovery, and medical imaging AI — and both value the operational resilience that stable thermals provide. Government and defense value the reliability and reduced water dependency in constrained and sometimes austere environments, where access to water cannot be assumed and uptime is non-negotiable. Media and entertainment, driven by rendering and AI-generated content, and energy and manufacturing, driven by edge and industrial compute, round out the demand base. Across all of them, the common thread is the same: more compute per rack, more heat, less tolerance for water.

Segment Insights

Waterless Liquid Cooling Market, By Cooling Technology

Two-phase direct-to-chip leads the technology conversation and commands the strongest strategic mindshare. By boiling a dielectric fluid directly at the chip, it captures heat where it is generated and scales with rising chip power in a way single-phase approaches struggle to match, because the latent heat absorbed during the liquid-to-vapor phase change carries away far more energy per unit of fluid than simply warming a liquid can. Its appeal to AI operators — no water in the rack, high thermal headroom, and retrofit-friendly deployment that avoids a white-space rebuild — has made it the reference point against which other waterless approaches are measured, and its momentum has been reinforced by high-profile qualification against leading GPU platforms.

Immersion cooling, both single-phase and two-phase, retains a durable position, especially for operators who value the simplicity of submerging entire servers and eliminating fans altogether. Single-phase immersion, using synthetic or hydrocarbon fluids that circulate without boiling, appeals to buyers seeking lower fluid cost and regulatory simplicity, while two-phase immersion offers the highest heat-transfer performance for the most extreme densities. The trade-offs among these approaches — serviceability, fluid cost, retrofit feasibility, and peak thermal capacity — are exactly what buyers weigh, and no single technology has yet won every use case.

The fastest-growing technology is in-chip microfluidics, albeit from a nascent base. By etching coolant channels into the silicon itself, microfluidics removes heat before it ever reaches a cold plate, and early research has shown dramatically better heat removal than conventional cold plates. It is not yet a production technology, and formidable obstacles remain around manufacturability, chip-packaging compatibility, and the risk of clogging in channels finer than a human hair. But its trajectory is steep because it addresses the exact constraint — chip-level thermal density — that will define the next decade of accelerator design, and its progress from research to roadmap has been strikingly fast.

Waterless Liquid Cooling Market, By Offering

Within offerings, solutions lead decisively, and within solutions, coolant distribution units and the dielectric fluids themselves anchor the value. The CDU is the heart of a waterless deployment, and vendors are scaling these units to megawatt capacities and to centralized end-of-row configurations that serve multiple racks from a single unit while preserving rack-level monitoring; the fluid, meanwhile, is both a consumable and a performance-defining component, which keeps it central to the sale.

Services are the faster-growing slice as the installed base expands. Design and consulting matter most at the point of adoption, when operators are deciding among technologies and integrating cooling into a broader facility plan, but maintenance, support, and the emerging cooling-as-a-service model grow fastest over the forecast. As more systems move from pilot to production, the recurring work of monitoring fluids, servicing components, and guaranteeing uptime accumulates, and operators increasingly want the outcome — guaranteed thermal performance — rather than the burden of running an unfamiliar system themselves. This is where vendors can build durable, high-margin relationships that outlast any single hardware sale, and it is a major reason the majors are so eager to enter the category.

Waterless Liquid Cooling Market, By Dielectric Fluid Type

Engineered fluorocarbon-based fluids lead in two-phase applications, where their precise boiling behavior is essential to phase-change cooling. Their thermal properties make them the default for the highest-performance systems, even as their cost and regulatory profile draw scrutiny.

Synthetic and hydrocarbon-based fluids are growing fastest, particularly in single-phase immersion, because they offer a lower-cost, lower-regulatory-risk path as buyers weigh the long-term availability of fluorinated alternatives. With regulatory scrutiny of certain fluorinated chemistries intensifying and at least one major producer having stepped back from parts of that market, operators increasingly favor fluids whose supply and compliance outlook they can count on over a ten- or fifteen-year facility life. The fluid layer is quietly becoming one of the most consequential decisions in a waterless deployment — it determines not only thermal performance but also cost trajectory, environmental profile, and exposure to future regulation, and it is one of the hardest choices to reverse once a facility is filled.

Waterless Liquid Cooling Market, By Data Center Type

Hyperscale data centers lead adoption, because they combine the densest racks, the deepest capital, and the greatest exposure to water and energy scrutiny. When a hyperscaler standardizes on a waterless architecture, it moves the whole market.

Colocation is the fastest-growing type, as operators race to offer high-density, AI-ready space and increasingly market waterless cooling as a differentiator to tenants who demand both performance and sustainability. For colocation providers, the ability to advertise a low water footprint and readiness for the latest accelerators is becoming a commercial necessity rather than a nice-to-have, because enterprise and AI-native tenants now screen for exactly these attributes when choosing where to place workloads. Enterprise and on-premise facilities adopt more cautiously, constrained by capital and legacy design, though managed-service models are lowering that barrier. Edge deployments follow, where closed-loop waterless systems suit sites with no access to water infrastructure and where the compact, self-contained nature of dielectric cooling is a genuine advantage in locations that were never built to be data centers at all.

Key segmentation conclusions:

• Two-phase direct-to-chip is the strategic center of gravity, while in-chip microfluidics is the steepest emerging curve.
• Solutions dominate revenue, but services and cooling-as-a-service grow fastest as the installed base matures.
• Fluid choice is a first-order decision, with fluorocarbons leading performance and hydrocarbons gaining on cost and compliance.
• Hyperscale sets the direction; colocation drives the fastest incremental growth.
• Across every axis, the demand signal traces back to AI rack density and water/energy exposure.

Regional Analysis

North America

North America holds the largest share of the waterless liquid cooling market, valued at roughly USD 0.87 billion in 2025 and projected to reach about USD 4.39 billion by 2032 at a CAGR of 26.0%. The United States is the engine, home to the hyperscalers building AI factories at unprecedented scale and to the densest cluster of waterless-cooling innovators — ZutaCore, Accelsius, and LiquidStack among them. Water scarcity in build-out hotspots such as Arizona and Texas sharpens the case for zero-water designs, and Microsoft's pilots of zero-water cooling in Arizona and Wisconsin signal how the largest operators are responding. Canada's cool climate and clean-power availability make it attractive for high-density campuses, while Mexico is emerging as a nearshoring destination for both compute and cooling-equipment manufacturing. Leading demand comes from cloud, telecom, and HPC operators racing to deploy accelerators that air cooling cannot support.

Europe

Europe's waterless liquid cooling market is valued at approximately USD 0.51 billion in 2025 and is expected to reach around USD 2.30 billion by 2032 at a CAGR of 24.0%. Growth here is distinctly regulation-driven. Germany, with its dense data center corridor around Frankfurt and its energy-efficiency rules, leads adoption, followed by the United Kingdom's London cluster and France's nuclear-powered, comparatively low-carbon grid. Italy and Spain are expanding as southern-European hubs where water stress reinforces the waterless argument, and the Nordics leverage cold climates and abundant renewable power for efficient, sustainable capacity. EU energy-efficiency directives and tightening reporting requirements make waterless cooling a compliance asset as much as a performance one, and European operators are among the most focused on pairing it with heat reuse.

Asia Pacific

Asia Pacific is the fastest-growing region, valued at roughly USD 0.69 billion in 2025 and projected to reach about USD 4.10 billion by 2032 at a CAGR of 29.0%. China dominates regional volume, with documented immersion deployments serving domestic technology giants and PUE targets in Beijing and Tianjin that are achievable only with advanced cooling. Japan's high electricity costs and earthquake-resilience requirements favor immersion's reduced mechanical footprint, while South Korea and Australia expand high-density capacity for AI and cloud. India is a standout growth story as its data center build-out accelerates, and Singapore — long constrained on land, power, and water — is pioneering closed-loop, no-evaporative-water designs suited to hot, humid climates. Government digitalization initiatives across the region add a policy tailwind that few other geographies match.

Rest of World

The Rest of World market is valued at approximately USD 0.23 billion in 2025 and is forecast to reach around USD 1.23 billion by 2032 at a CAGR of 27.0%. The Middle East is the standout: Saudi Arabia and the UAE are building AI capacity aggressively, and because these are among the most water-scarce places on earth, near-zero-water cooling is effectively mandatory for new facilities, pushing growth well above the region's small base would suggest. In Latin America, Brazil anchors demand as its cloud and colocation sector matures. In Africa, South Africa leads a nascent but expanding market. Across Rest of World, infrastructure development and policy shifts toward digitalization are opening demand pockets faster than the modest base implies.

Regional outlook — key conclusions:

• North America leads on installed base, capital intensity, and innovator density.
• Asia Pacific grows fastest, powered by China's scale, India's build-out, and strict PUE regimes.
• Europe's growth is regulation-led, with heat reuse a distinctive local priority.
• The Middle East punches above its base as water scarcity makes waterless cooling non-negotiable.
• Every region's demand ultimately tracks AI capacity growth and local water and energy constraints.

Country-Specific Insights

The United States is the single most important country in this market, combining hyperscale demand, a concentration of specialist vendors, and acute regional water stress that turns waterless design into a strategic requirement rather than a preference. Federal and state-level attention to data center water use, alongside corporate water-positive commitments, reinforces the trajectory. China is the second pivotal market, where national and municipal efficiency targets — particularly aggressive PUE caps in major hubs — compel advanced cooling, and where domestic vendors reduce equipment costs relative to imports, accelerating local adoption.

Germany leads Europe on the strength of its data center concentration and its energy-efficiency framework, with heat-reuse expectations increasingly written into planning. The Frankfurt corridor's density, combined with rules that push operators toward measurable efficiency and, in some cases, toward supplying rejected heat to surrounding communities, makes Germany a bellwether for how regulation can pull waterless cooling forward rather than merely permitting it. India deserves particular attention as well: its data center sector is expanding rapidly on the back of digital-public-infrastructure initiatives and surging cloud demand, and because much of that capacity is being built new, operators can design waterless cooling in from the start rather than retrofitting later.

Singapore, despite its size, exerts outsized influence: its long-standing resource constraints — limited land, power, and water — have made it a proving ground for closed-loop, water-free designs that the rest of tropical Asia is watching, and its lifting of earlier moratoriums on new data centers came with strict sustainability conditions that effectively reward advanced cooling. Saudi Arabia and the UAE, finally, are reshaping the Middle East's demand profile, coupling sovereign AI ambitions and large-scale national investment programs with the hard constraint of water scarcity to make waterless cooling the default for new capacity. In these markets the question is rarely whether to go waterless but which waterless approach best suits the climate, the power profile, and the compute roadmap.

Country-level conclusions:

• The US sets global direction through hyperscale demand and vendor concentration.
• China's efficiency mandates make advanced cooling a regulatory necessity.
• Germany anchors European adoption with efficiency rules and heat reuse.
• Singapore is the tropical-climate proving ground for zero-water design.
• Saudi Arabia and the UAE turn water scarcity into a waterless-cooling mandate.

Key Company Insights

The competitive landscape blends established thermal and infrastructure majors with focused waterless specialists. The leading players include Vertiv, Schneider Electric, ZutaCore, Accelsius, LiquidStack, Submer, Green Revolution Cooling, Iceotope, Asperitas, Boyd, Rittal, Fujitsu, Carrier, Daikin, and Shell.

• Vertiv Holdings Co.
• Schneider Electric SE
• ZutaCore Ltd.
• Accelsius
• LiquidStack Holding B.V.
• Submer Technologies
• Green Revolution Cooling (GRC)
• Iceotope Technologies
• Asperitas
• Boyd Corporation
• Rittal GmbH & Co. KG
• Fujitsu Limited
• Carrier Global Corporation
• Daikin Applied
• Shell plc

Strategic activity clusters around three moves. The first is investment and consolidation: Carrier's venture arm invested in ZutaCore and then expanded that stake, while Daikin acquired direct-to-chip specialist Chilldyne and Vertiv absorbed liquid-cooling assets in China. The majors are buying their way into waterless capability rather than building it slowly, and their motivation is straightforward — the incumbents own the customer relationships, the manufacturing scale, and the service networks, but the specialists own the technology, so acquisition and investment close the gap faster than internal development could. For the specialists, in turn, the backing of a thermal major brings the credibility, distribution, and balance-sheet reassurance that risk-averse hyperscale buyers demand. The second is product scale-up: ZutaCore launched a waterless end-of-row coolant distribution unit family reaching multi-megawatt capacity, and Accelsius has pushed per-socket cooling capacity to levels aimed squarely at future accelerators. The race here is to demonstrate headroom for chips that have not yet shipped, because operators buying today are provisioning for the accelerators of the next several years. The third is ecosystem partnership: chip and server qualification matters enormously, and waterless systems that pass hyperscaler and chip-vendor certification gain decisive credibility. A system that has been validated against a leading GPU platform carries a trust signal that no marketing claim can replicate, which is why vendors invest heavily in certification. Fluid suppliers such as Shell are meanwhile industrializing the coolant layer, turning bespoke formulations into supply chains — a shift that lowers cost, improves availability, and gradually erodes the vendor lock-in that has held some buyers back. The companies that will lead the market are those that combine strong technology with the ecosystem relationships, service depth, and financial stability that convert a promising pilot into a multi-site standard.

Key company strategies — conclusions:

• Thermal majors are acquiring or investing to enter waterless cooling quickly.
• Specialists differentiate on per-chip thermal headroom and zero-water credentials.
• Chip- and hyperscaler-level qualification is the key trust signal.
• Fluid suppliers are becoming strategic actors as the coolant layer matures.
• Partnerships across chip, server, and cooling vendors define the winning ecosystems.

Recent Developments

• In June 2025, Shell launched its DLC Fluid S3, a dielectric coolant designed for both direct-to-chip and immersion cooling applications.
• In September 2025, Microsoft published research on in-chip microfluidic cooling, developed with Swiss startup Corintis, showing up to three times better heat removal than conventional cold plates.
• In November 2025, Daikin Applied acquired Chilldyne to add direct-to-chip liquid cooling to its data center portfolio.
• In November 2025, ZutaCore launched its Waterless End-of-Row Coolant Distribution Unit family, featuring 1.2 MW and 2 MW systems for AI and HPC environments.
• In April 2026, Carrier Ventures expanded its investment in ZutaCore to scale waterless liquid cooling for high-density AI data centers.
• In June 2026, ZutaCore raised USD 100 million to accelerate deployment of its waterless cooling technology.

Real-World Use Cases

In September 2025, Microsoft unveiled a working in-chip microfluidic cooling system developed in collaboration with Swiss startup Corintis, addressing the thermal bottleneck constraining increasingly dense AI accelerators. Rather than sitting a cold plate atop the chip, Microsoft etched microscopic channels — roughly the width of a human hair — directly into the silicon and used AI to optimize a bio-inspired, leaf-vein channel pattern so coolant flows precisely where heat concentrates. The company validated the design on a server running simulated Microsoft Teams meetings and reported up to three times better heat removal than conventional cold plates, with peak temperature rise cut by as much as 65%. Microsoft framed the work as part of a broader systems strategy to reduce both water and power in data center cooling, consistent with its goal of becoming water-positive by 2030.

Also in 2025, a rack-scale waterless two-phase direct-to-chip solution built on ZutaCore's HyperCool technology was integrated with NVIDIA H200 GPUs and passed NVIDIA's NVQual certification, demonstrating stability and compatibility at the rack level while achieving a partial power-usage-effectiveness figure of 1.03 — a benchmark in cooling efficiency. Building on that momentum, server maker Compal unveiled an AI server integrating HyperCool at CloudFest 2025, using the two-phase system to raise rack density and energy efficiency. The deployments show waterless two-phase cooling meeting the qualification bar set by the most demanding AI hardware, not merely in the lab but in shipping server platforms.

Market Segmentation

The waterless liquid cooling market is best understood across several interlocking dimensions. By cooling technology, it spans two-phase direct-to-chip, single-phase and two-phase immersion, waterless closed-loop direct-to-chip, and the emerging frontier of in-chip microfluidics — a spectrum that runs from proven, deployable systems to research reshaping the next generation of silicon. By offering, it divides into solutions — coolant distribution units, cold plates and evaporators, immersion tanks, dielectric fluids, heat rejection units, and control software — and the services that design, install, and maintain them, with cooling-as-a-service emerging as a distinct model. By dielectric fluid type, it separates engineered fluorocarbon fluids, favored in two-phase systems, from synthetic and hydrocarbon fluids gaining ground on cost and regulatory grounds.

By data center type, the market covers hyperscale, colocation, enterprise, and edge facilities, each with distinct density and infrastructure profiles. By end-user vertical, it reaches cloud and hyperscale providers, IT and telecommunications, BFSI, healthcare and life sciences, government and defense, media and entertainment, research and HPC, and industrial users. And by geography, it splits across North America, Europe, Asia Pacific, and Rest of World, with the strategically decisive countries being the United States, China, Germany, Singapore, and the Gulf states. Taken together, these axes show a market defined not by a single product but by an ecosystem responding to one relentless pressure: the heat of AI compute.

Segmentation summary:

• Cooling technology ranges from deployable two-phase direct-to-chip to emerging microfluidics.
• Offerings split into fluid-and-hardware solutions and a fast-growing services layer.
• Fluid type is a strategic axis balancing performance against cost and compliance.
• Data center type spans hyperscale leadership to edge closed-loop niches.
• Verticals and regions both trace demand back to AI density and resource constraints.

Conclusion and Future Outlook

Through 2032, the waterless liquid cooling market will be shaped by the same forces propelling it today, only intensified. AI accelerators will keep getting hotter, and each generation will widen the gap between what air and water-glycol can cool and what the silicon demands. Automation and AI will increasingly design the cooling itself — as Microsoft's AI-optimized microfluidic channels already demonstrate — and predictive control software will squeeze more efficiency from every deployment. Sustainability regulation will tighten rather than relax, and in the most water-stressed and PUE-capped jurisdictions, waterless cooling will shift from advantage to requirement.

The technology mix will also keep evolving. Two-phase direct-to-chip is likely to remain the workhorse of high-density AI deployment through much of the forecast, but in-chip microfluidics could reshape the landscape if its manufacturing and reliability challenges are solved, and immersion will hold its ground wherever whole-server simplicity and extreme density matter most. Expect the boundaries between these approaches to blur as hybrid architectures emerge, combining, for example, direct-to-chip cooling for the hottest components with complementary systems for the rest of the rack. The fluid layer will mature into a genuine supply chain, standards and reference designs will gradually reduce lock-in, and cooling-as-a-service will bring waterless technology to buyers who would never operate it themselves.

For businesses weighing this market, the strategic importance is hard to overstate. Thermal management has become a gating constraint on AI capacity: the ability to cool determines the ability to compute, and the ability to compute increasingly determines competitive position. Operators that master waterless cooling will deploy denser, more efficient, more sustainable infrastructure; those that do not will find themselves capacity-constrained and exposed to water and energy risk. Investors, meanwhile, are treating the category as core infrastructure, and the flow of strategic capital into leading specialists suggests the market's growth is underpinned by conviction rather than hype. The market's strong double-digit growth reflects a simple reality — waterless liquid cooling has moved from experimental edge to strategic core, and the companies building, buying, and financing it are positioning for the defining infrastructure decision of the AI era.

Frequently Asked Questions

How big is the waterless liquid cooling market?

The global waterless liquid cooling market is valued at approximately USD 2.30 billion in 2025 and is projected to reach USD 12.02 billion by 2032. That growth reflects the rapid shift from air and water-based cooling toward dielectric-fluid systems as AI workloads push data center heat densities to new extremes.

What is the waterless liquid cooling market growth rate?

The market is expected to grow at a CAGR of 26.5% between 2026 and 2032. The pace is driven by AI-accelerator power growth, tightening water and energy regulation, and heavy capital investment in hyperscale and AI-factory infrastructure.

Which segment leads the waterless liquid cooling market?

By cooling technology, two-phase direct-to-chip holds the strongest strategic position because it cools the hottest AI chips without any water in the rack and scales with rising chip power. By end user, cloud and hyperscale providers lead, as they operate the densest racks and face the greatest water and energy exposure.

Who are the key players in the waterless liquid cooling market?

Leading players include Vertiv, Schneider Electric, ZutaCore, Accelsius, LiquidStack, Submer, Green Revolution Cooling, Iceotope, Asperitas, Boyd, Rittal, Fujitsu, Carrier, Daikin, and Shell. The landscape blends established thermal and infrastructure majors with focused waterless-cooling specialists.

What are the factors driving the waterless liquid cooling market?

The main drivers are AI and HPC rack densities exceeding the limits of air cooling, water-scarcity pressure and tightening water-usage-effectiveness mandates, rising energy costs that reward cooling efficiency, and the enormous capital flowing into hyperscale AI infrastructure. Together these make waterless cooling a strategic necessity rather than an option.