Key highlights
- 1The Asia Pacific data center liquid cooling market is expected to reach USD 17.5 billion by 2036, at a CAGR of 23.6% from 2026 to 2036, driven by AI workloads, rising rack density, and efficiency regulation, from a growing base.
- 2AI rack density forces liquid cooling. GPU rack power has risen sharply, with NVIDIA's GB200 NVL72 rack drawing 120 to 130 kW and next-generation racks unveiled at OCP 2025 demanding up to 1 MW, densities that air cooling cannot manage.
- 3China leads and regulates. China's National Development and Reform Commission requires new data centers to meet a power usage effectiveness below about 1.25, driving the shift to liquid cooling, and domestic supplier Envicool, whose customers include NVIDIA, Alibaba and Tencent, reported 2025 revenue of about RMB 6.07 billion, up 32%.
- 4Direct-to-chip leads, immersion grows. Direct-to-chip cold-plate cooling handles densities around 100 to 175 kW and is the mainstream choice, while immersion cooling, popular in Asia Pacific, supports the highest densities.
- 5Singapore set a tropical standard. Singapore launched Standard SS 715:2025, described as the world's first liquid cooling standard for tropical data centers, enabling equipment to operate at up to 35 degrees Celsius and reducing mechanical cooling power.
- 6Japan is piloting at scale. Preferred Networks, Internet Initiative Japan and JAIST began a liquid cooling pilot for AI servers, and operators such as MC Digital Realty expanded AI-ready capacity.
- 7Key companies include Vertiv Holdings Co., Schneider Electric SE, Envicool, Delta Electronics, Inc., and CoolIT Systems.
Report Overview
The Asia Pacific data center liquid cooling market covers liquid cooling systems for data centers, spanning direct-to-chip cold-plate cooling, single-phase and two-phase immersion cooling, and rear-door heat exchangers, together with cooling distribution units, coolants, heat exchangers and services. Conventional air cooling is outside the scope except as context. The market is driven by artificial intelligence, which pushes rack power densities beyond air cooling, and is scaling rapidly across the region's hyperscale, colocation and enterprise data centers. Demand is shaped by AI and high-performance computing, rack density, energy-efficiency regulation, and the data center buildout. This report examines the size, drivers, cooling types, data center types, applications, components, countries, pricing, competition, recent developments, and outlook of the market, and provides recommendations. Sizing is built bottom-up from liquid cooling systems, components and services by cooling type, data center type and country, and reflects liquid cooling deployed in Asia Pacific data centers.
Market dynamics
Drivers
- 01AI and high-performance computing are the primary driver as training and inference clusters use dense GPU servers that generate heat far beyond what air cooling can remove, and the buildout of AI data centers across China, Japan, South Korea, India and Singapore drives demand for liquid cooling.
- 02Rising rack power density is the direct technical driver as GPU rack power has risen from hundreds of watts per chip to racks drawing 120 to 130 kW, such as NVIDIA's GB200 NVL72, and next-generation racks approaching 1 MW, densities at which liquid cooling becomes necessary.
- 03Energy-efficiency regulation is a driver as China's National Development and Reform Commission requires new data centers to meet a power usage effectiveness below about 1.25, and Singapore's tropical liquid cooling standard enables warmer operation, both favouring liquid cooling over energy-intensive mechanical chillers.
Opportunities
- 01Direct-to-chip cooling for GPU clusters is the leading opportunity because it handles the densities around 100 to 175 kW typical of AI racks and is the mainstream choice for GPU deployments, with strong demand across hyperscale and colocation.
- 02Immersion cooling for the highest densities is an opportunity as immersion supports ultra-high-density racks and is comparatively popular in Asia Pacific, and may become the default for some new hyperscale projects.
- 03Retrofits and cooling distribution units are an opportunity as existing data centers add liquid cooling and require cooling distribution units, coolants and integration.
Trends
- 01The mainstreaming of liquid cooling is the defining trend as AI rack densities move liquid cooling from niche to a standard part of new data center design.
- 02The lead of direct-to-chip with immersion growing is a technology trend, with cold-plate cooling the mainstream choice and immersion expanding for the highest densities.
- 03Efficiency regulation and warm-water operation are a regional trend with China's power usage effectiveness limits and Singapore's tropical standard driving adoption.
- 04Standardisation and waste-heat recovery are trends with OCP and ASHRAE standards and reuse of data center heat.
Report Summary
| Base Year | 2025 |
|---|---|
| Forecast Period | 2026-2036 |
| Market Size (2025) | USD 1.7 billion |
| Market Size (2026) | USD 2.1 billion |
| Market Size (2036) | USD 17.5 billion |
| CAGR (Value) | 23.6% (2026-2036) |
| Format | PDF & Excel |
| Segments Covered | By Cooling Type: Direct-to-Chip, Immersion (Single-Phase, Two-Phase), Rear-Door Heat Exchangers. By Data Center Type: Hyperscale, Colocation, Enterprise, Edge. By Application; By Country. |
| Geographies Covered | China, Japan, South Korea, India, Singapore, Australia, and Rest of Asia Pacific |
| Key Companies | Vertiv Holdings Co., Schneider Electric SE, Envicool, Delta Electronics, Inc., CoolIT Systems, LiquidStack, Green Revolution Cooling (GRC), Submer, Boyd Corporation, Shenling Environmental Systems, Other Companies |
Segmental analysis
By Cooling Type
- Direct-to-chip cold-plate cooling holds the largest share at about 55% of the market in 2026, with the remaining share divided across immersion cooling and rear-door heat exchangers.
- Direct-to-chip cooling which circulates coolant through cold plates on the chips, is the largest type because it handles the densities around 100 to 175 kW typical of AI GPU racks, integrates with existing server and rack designs more readily than immersion, and is the mainstream choice for GPU clusters.
- Immersion cooling in single-phase and two-phase forms, submerges servers in dielectric fluid and supports the highest densities, and is comparatively popular in Asia Pacific, though it requires custom servers and adds maintenance complexity.
- Rear-door heat exchangers cool at the rack door and serve lower densities and retrofits.
The dominance of direct-to-chip reflects its fit with AI rack densities and existing infrastructure.
By Data Center Type
- Hyperscale holds the largest share at about 50% of the market in 2026, with the remaining share divided across colocation, enterprise and edge.
- Hyperscale is the leading type because cloud and AI companies build the largest, highest-density clusters that most require liquid cooling, and the region's AI buildout is concentrated in hyperscale facilities.
- Colocation is a large and growing type as providers build AI-ready, liquid-cooled capacity for tenants.
- Enterprise data centers adopt liquid cooling for AI and high-performance computing.
- Edge data centers use liquid cooling in compact high-density deployments.
The dominance of hyperscale reflects the concentration of AI compute and the scale of hyperscale liquid cooling deployments.
By Application
- Artificial intelligence and high-performance computing hold the largest share, at about 60% of the market in 2026, with the remaining share divided across cloud and enterprise information technology.
- AI and high-performance computing are the leading application because dense GPU clusters for training and inference generate the heat that makes liquid cooling necessary, and the region's investment in AI drives most liquid cooling demand.
- Cloud applications use liquid cooling for high-density general compute.
- Enterprise information technology adopts liquid cooling for AI and demanding workloads.
The dominance of AI and high-performance computing reflects that liquid cooling is driven above all by the density of AI hardware.
By Component
- Solutions and hardware including cold plates, cooling distribution units, immersion tanks, heat exchangers, manifolds and coolants, hold the largest share, as the physical cooling systems represent most of the spending in a rapidly building market.
- Services including design, integration, installation, maintenance and cooling-as-a-service, are a growing component, important to operators adopting and maintaining liquid cooling.
The balance of hardware during rapid buildout and growing services over the life of systems shapes the component mix, with hardware dominant during the current expansion.
Geographic analysis
China Data Center Liquid Cooling Market
China is the largest market in the region driven by its vast AI and data center buildout and by regulation, with the National Development and Reform Commission requiring new data centers to meet a power usage effectiveness below about 1.25, effectively pushing operators toward liquid cooling. The country has a strong domestic supply base, including Envicool, whose customers include NVIDIA, Alibaba and Tencent and which reported 2025 revenue of about RMB 6.07 billion, up 32%, and Shenling, Sugon and others. China's scale, regulation and domestic manufacturing make it the dominant and fastest-scaling market.
Japan and South Korea
Japan is scaling AI-ready liquid-cooled capacity, with Preferred Networks, Internet Initiative Japan and JAIST piloting liquid cooling for AI servers and operators such as MC Digital Realty expanding capacity, supported by high energy costs that favour efficient cooling. South Korea builds AI and high-performance computing clusters with liquid cooling, backed by its technology industry. Both markets combine strong AI investment with demand for energy-efficient cooling, and are significant regional markets.
India, Singapore, Australia, and Rest of Asia Pacific
India is a rapidly growing data center market building AI-ready liquid-cooled capacity. Singapore, a major data center hub constrained by power, launched Standard SS 715:2025, the world's first liquid cooling standard for tropical data centers, enabling warmer operation at up to 35 degrees Celsius to cut cooling power. Australia builds AI and cloud capacity with liquid cooling. The rest of Asia Pacific, including Southeast Asia, is an emerging market as data center investment spreads. These markets add fast-growing demand across the region.
Pricing Analysis
Pricing in data center liquid cooling reflects cooling type, scale, and integration. Direct-to-chip systems are priced on cold plates, cooling distribution units, manifolds and integration, and immersion systems on tanks, dielectric fluid and custom servers, with immersion carrying higher fluid and server-adaptation cost but supporting higher densities. The economics are weighed against the cost and limits of air cooling, the energy savings from lower power usage effectiveness, and the ability to support dense AI hardware that air cooling cannot. Several factors set price. Cooling type is central, as direct-to-chip and immersion have different cost structures, with immersion adding dielectric fluid and server adaptation. Scale and density raise cost, with larger and denser deployments requiring more cooling capacity. Coolants and fluids are a cost, particularly dielectric fluids for immersion. Retrofit versus greenfield affects cost, as adding liquid cooling to existing facilities is more complex. Services and integration add cost. Energy savings and regulation support value, as lower power usage effectiveness cuts operating cost and meets efficiency limits.
The trajectory of pricing depends on scale, competition, and technology, and the market combines high capital cost with strong value from enabling AI and improving efficiency.
Competitive landscape
The market is served by global cooling and infrastructure vendors, specialised liquid cooling companies, and strong Chinese suppliers. Vertiv Holdings Co. and Schneider Electric SE, the latter through Motivair, offer end-to-end liquid cooling portfolios including cold plates, cooling distribution units and integration. Delta Electronics provides cooling and infrastructure. Specialised companies CoolIT Systems, LiquidStack, Green Revolution Cooling, Submer, Boyd, Iceotope and Asetek provide direct-to-chip and immersion cooling. In China, Envicool, whose customers include NVIDIA, Alibaba and Tencent, Shenling and Sugon are strong domestic suppliers serving the region's largest market. Other vendors including Alfa Laval, Rittal and Stulz supply components and systems.
Competition turns on cooling performance and density capability, integration and reliability, coolant and system design, and cost and service, and the region combines global vendors with strong Chinese suppliers benefiting from the domestic buildout and regulation. AI demand, rising density and efficiency regulation favour vendors that can deliver reliable, high-density direct-to-chip and immersion systems at scale, and Vertiv, Schneider, Envicool and the specialists hold strong positions. The field is competitive and fast-growing, with rapid capacity expansion, standardisation and new entrants, and will continue to intensify as AI infrastructure scales across the region.
Voice of Customer
AI rack densities have made liquid cooling non-negotiable; our GPU racks draw well over 100 kW, and air cooling simply cannot remove that heat. We deploy direct-to-chip cooling as standard for GPU clusters and evaluate immersion for the highest densities, and efficiency regulation and power constraints make the lower power usage effectiveness of liquid cooling essential.
Infrastructure lead, hyperscale operator (Asia Pacific):
Meeting the power usage effectiveness requirement for new data centers means liquid cooling, and our tenants building AI clusters demand it. We work with domestic and global vendors on direct-to-chip and immersion, and the challenge is scaling capacity and integrating cooling distribution units reliably as we build out quickly.
Design manager, colocation provider (China):
We piloted liquid cooling for our AI servers because the density and our high energy costs require it, and the efficiency gains are significant. Immersion supports very high density but adds maintenance complexity with the fluid and custom servers, so we weigh direct-to-chip against immersion by deployment, and reliability is paramount.
Facilities director, AI company (Japan):
Analyst perspective
Asia Pacific's data center liquid cooling market is one of the fastest-growing infrastructure markets, driven directly by artificial intelligence. The technical reality is decisive: GPU rack power has risen from hundreds of watts per chip to racks drawing 120 to 130 kW, such as NVIDIA's GB200 NVL72, with next-generation racks approaching 1 MW, and air cooling cannot remove this heat, so liquid cooling has moved from niche to necessary for AI infrastructure. Direct-to-chip cooling leads for GPU clusters, immersion cooling, comparatively popular in Asia Pacific, supports the highest densities, and the region combines the world's largest data center buildout in China with strong growth in Japan, South Korea, India and Singapore. Regulation reinforces the shift, with China's power usage effectiveness limits and Singapore's tropical liquid cooling standard.
The honest considerations are the AI capital cycle, cost and complexity, and execution. The market's very high growth rests on continued heavy investment in AI infrastructure, which is subject to cycles and expectations, so a slowdown in AI capital spending would slow liquid cooling demand, and the figures should be read as tied to the AI buildout. Liquid cooling is capital-intensive, retrofits are complex, and immersion adds dielectric-fluid cost, custom servers and maintenance considerations, with environmental questions around some two-phase fluids. Rapid scaling strains supply chains, with specialised manufacturing taking time to build. The market should be assessed on the pace of AI infrastructure investment, cooling technology and cost, and regulation rather than on the density imperative alone, and the region's AI buildout, scale and regulation support very strong growth, with the AI capital cycle and execution the key determinants of the pace.
Strategic Recommendations
For cooling vendors
The priority is to scale reliable direct-to-chip and immersion systems and to build capacity and supply chains, because AI-driven demand is surging and the constraint is delivering reliable, high-density cooling at scale. Companies should expand direct-to-chip portfolios for GPU clusters, develop immersion for the highest densities, provide cooling distribution units, coolants and integration, and build manufacturing capacity, while meeting efficiency and standardisation requirements. Serving China's regulated, large-scale market, partnering with hyperscale and colocation operators, and offering services and cooling-as-a-service strengthen the position.
For data center operators
The recommendation is to adopt liquid cooling as standard for AI and high-density deployments, choosing direct-to-chip for GPU clusters and immersion for the highest densities, and to plan for efficiency regulation and power constraints, prioritising reliability and integration. For chip and server makers, designing for liquid cooling is essential. For policymakers, efficiency standards, power planning and support for data center infrastructure shape the market. For investors, this is one of the fastest-growing infrastructure markets, to evaluate on the pace of AI investment, cooling technology and cost, and regulation rather than on the density imperative alone, recognising that the region's AI buildout, scale and regulation support very strong growth while the AI capital cycle and execution remain the key determinants.
Sustainability impact
Energy Efficiency and Power Usage Effectiveness
Liquid cooling reduces the energy needed to cool data centers, lowering power usage effectiveness compared with air cooling.Liquid cooling supports energy efficiency.
By removing heat far more efficiently than air, liquid cooling lowers the power used for cooling and reduces power usage effectiveness, helping data centers meet efficiency limits such as China's and cut the large energy footprint of computing.
Enabling Dense, Efficient Compute
Liquid cooling enables dense AI hardware to run efficiently, supporting more computing per unit of infrastructure. Liquid cooling supports efficient computing.
By cooling dense GPU racks that air cannot handle, liquid cooling enables high-performance computing to run within efficient, compact infrastructure, supporting the AI workloads that drive the digital economy while managing their energy intensity.
Waste-Heat Recovery
Liquid cooling captures data center heat at higher temperatures, enabling reuse for district and industrial heating. Liquid cooling supports waste-heat recovery.
By capturing heat in liquid at useful temperatures, liquid cooling makes it possible to recover and reuse data center waste heat for district or industrial heating, turning a waste stream into a resource and improving overall energy use.
Water, Fluids, and Electricity Dependence
The sustainability of liquid cooling depends on water use, coolant choice, and the carbon intensity of electricity. Liquid cooling requires responsible design.
By using water and dielectric fluids and running on electricity, liquid cooling must manage water consumption, select low-impact coolants and avoid harmful fluids, and rely on low-carbon power for its full benefit, so responsible design is essential to its sustainability.
Table of contents
15 chapters · 167 pages · click to expandFrequently asked questions
The Asia Pacific data center liquid cooling market was valued at USD 1.7 billion in 2025 and is projected to reach USD 17.5 billion by 2036, growing from USD 2.1 billion in 2026, at a CAGR of 23.6% from 2026 to 2036, driven by AI workloads, rising rack density, and efficiency regulation, with growth tied to AI investment.