Immersion Cooling Market Size and Forecast 2035
The global immersion cooling market was valued at approximately US$351.3 million in 2025 and is projected to reach US$1.34 billion by 2035, growing at a CAGR of 14.3% during 2026–2035.
Immersion cooling involves placing servers, GPUs, CPUs, ASICs and other IT hardware directly into electrically non-conductive dielectric fluid. Heat is transferred from electronic components into the fluid and subsequently rejected through heat exchangers, coolant distribution systems or facility-water loops. The technology is increasingly evaluated for AI, high-performance computing, cloud infrastructure and other high-density workloads.
AI infrastructure is becoming a major adoption catalyst because accelerated-computing clusters can create rack densities that are increasingly difficult to manage using conventional air cooling. Open Compute Project guidance specifically addresses both single-phase and two-phase immersion systems and is developing specifications covering fluids, hardware compatibility, tanks, CDUs, maintenance and interoperability.
Immersion Cooling Market Scope
| Metrics | Details |
| Market Size in 2025 | US$351.3 Million |
| Forecast Value in 2035 | US$1.34 Billion |
| CAGR | 14.30% |
| Revenue Unit | US$ Million/Billion |
| By Cooling Architecture | Single-Phase Immersion and Two-Phase Immersion |
| By System Component | Immersion Tanks, CDUs, Heat Exchangers, Pumps, Controls, Fluid Handling and Integrated Systems |
| By Cooling Fluid | Hydrocarbon/Synthetic Fluids, Mineral Oils, Engineered Dielectric Fluids and Other Compatible Fluids |
| By Data Center Type | Hyperscale, Colocation, Enterprise, Edge, HPC and AI Infrastructure |
| By Application | AI/ML, HPC, Cloud Computing, Cryptocurrency Mining, Telecom, Edge and Research Computing |
| By Deployment | New-Build Data Centers and Retrofit Installations |
| By Region | North America, Europe, Asia-Pacific, Latin America, Middle East and Africa |
| Major Growth Market | North America |
| High-Potential Region | Asia-Pacific |
| Report Coverage | Capacity, Rack Density, Fluids, TCO, Energy, Water, Standards, Vendors, Deployment Pipeline and Competitive Strategy |
Request for Exclusive Sample on Immersion Cooling Market
Immersion Cooling Market Key Takeaways
- AI and high-density computing are accelerating immersion cooling adoption as rack heat loads exceed conventional air-cooling capabilities.
- Single-phase immersion is gaining traction for AI, HPC, enterprise, and edge deployments due to simpler fluid management.
- North America remains a major growth market, while Asia-Pacific offers significant long-term expansion potential.
- Energy and water efficiency are strengthening the business case for immersion cooling in next-generation data centers.
- Standardization through OCP initiatives is improving interoperability, fluid qualification, hardware compatibility, and deployment scalability.
- Direct-to-chip cooling remains a key competitor, pushing immersion providers to differentiate through density, efficiency, serviceability, and total cost of ownership.
- Integrated solutions combining tanks, dielectric fluids, CDUs, monitoring, and AI-ready infrastructure will create the strongest competitive opportunities.
Immersion Cooling Market Definition
The immersion cooling market includes hardware, tanks, fluid-management systems, dielectric fluids and directly attributable thermal-management components used to operate servers and compute equipment while submerged in a dielectric cooling medium.
It includes single-phase and two-phase systems, integrated immersion platforms, compatible cooling fluids, pumps, heat exchangers, control systems and immersion-specific CDUs. General data center chillers, conventional air cooling, rear-door heat exchangers and direct-to-chip cold plates are excluded except when benchmarked as competing cooling architectures.
The market should also distinguish immersion cooling from the broader liquid-cooling market. Direct-to-chip cooling transports liquid through cold plates attached to processors, while immersion cooling places the IT hardware itself within a dielectric fluid bath.
Immersion Cooling Market Segmentation Analysis
By Cooling Architecture
Single-Phase Immersion Cooling
Single-phase immersion cooling keeps the dielectric fluid in liquid form during operation. Heat generated by servers is absorbed by the fluid and transported to a heat exchanger, usually through natural convection or pumped circulation, before being rejected into a secondary cooling loop.
The architecture benefits from comparatively simple fluid management, limited vapor-control requirements and compatibility with a growing range of engineered hydrocarbon fluids. Open Compute Project documentation formally includes single-phase systems within its immersion standardization and interoperability work.
Single-phase systems are increasingly positioned for high-density AI, HPC and enterprise deployments where customers require predictable thermal control and straightforward servicing. Fluid lifetime, material compatibility, pumping efficiency and server warranty remain important purchasing considerations.
Two-Phase Immersion Cooling
Two-phase immersion cooling uses a dielectric fluid that boils at controlled component temperatures. Heat from processors causes the fluid to vaporize, after which the vapor condenses and returns to the bath, creating a passive phase-change cooling cycle.
The approach can provide very high heat-transfer capability but introduces additional complexity around fluid management, sealing, vapor containment, environmental considerations and operating economics. OCP has continued industry work around both single- and two-phase immersion technologies.
Its long-term share will depend on fluid availability, environmental regulation, cost, serviceability and whether operators require thermal performance beyond what single-phase immersion and direct-to-chip systems can provide.
By System Component
Immersion Tanks and Enclosures
Immersion tanks represent the core infrastructure in which servers, accelerators and other IT equipment are submerged. Designs range from compact modular systems for edge deployments to large tanks serving dense HPC and AI clusters.
Commercial differentiation increasingly depends on usable compute density, serviceability, server orientation, fluid circulation, maintenance access and integration with facility cooling. Modular architectures are becoming more relevant as operators seek incremental capacity expansion rather than full data hall redesign.
Coolant Distribution Units and Heat Rejection
CDUs, heat exchangers and secondary cooling loops transfer heat from the immersion bath to facility water or external heat-rejection systems. Their design affects system efficiency, redundancy, pumping power and the ability to reuse waste heat.
OCP’s broader Cooling Environments program treats CDUs and immersion as interconnected parts of liquid-cooled data center architecture. Buyers increasingly evaluate the full thermal chain rather than purchasing immersion tanks as isolated equipment.
Monitoring and Control Systems
Monitoring platforms track fluid temperature, flow, pressure, server condition and heat-rejection performance. Control software is becoming more important as AI clusters require stable thermal performance across rapidly changing compute loads.
Vendors that combine hardware with telemetry and automation can differentiate through predictive maintenance, remote monitoring and operational optimization. The value of controls is expected to increase as deployments move from small HPC clusters toward larger commercial AI data centers.
By Cooling Fluid
Synthetic Hydrocarbon and Engineered Dielectric Fluids
Synthetic hydrocarbon fluids are increasingly used in single-phase immersion systems because of their dielectric properties, thermal stability and material compatibility. Commercial selection depends on viscosity, flash point, oxidation stability, heat capacity and interaction with server materials.
OCP-recognized products include engineered hydrocarbon-based fluids designed specifically for immersion cooling, illustrating the shift from generic oils toward purpose-formulated data center fluids.
The fluid market will increasingly compete on lifecycle cost rather than purchase price alone. Operators must evaluate degradation, replacement intervals, filtration, environmental profile and compatibility with seals, plastics, circuit boards and cables.
Mineral Oil
Mineral oil has historically been used because of its low cost and dielectric characteristics. The supplied market page emphasizes its ability to absorb server heat and reduce dependence on conventional fan-based cooling.
However, untreated or unsuitable mineral oils can create material-compatibility, oxidation and contamination concerns. Enterprise and hyperscale deployments increasingly require fluids supported by validated material testing and long-duration operating data.
Mineral oil is therefore expected to remain relevant in cost-sensitive applications but face stronger competition from engineered synthetic fluids where equipment reliability, service life and warranty support carry greater commercial value.
By Application
Artificial Intelligence and Machine Learning
AI is becoming the strongest strategic growth opportunity for immersion cooling because GPU and accelerator clusters create extreme localized heat loads. Training systems also concentrate substantial electrical capacity within relatively small floor areas, increasing pressure on air-based cooling architecture.
OCP has highlighted 100 kW-plus rack-density considerations in discussions of liquid cooling for AI data centers. Immersion systems can support higher compute density by eliminating conventional server fans and transferring heat directly into liquid.
The opportunity is strongest where operators prioritize compute density, constrained floor space, energy efficiency and reduced water consumption. Competition from direct-to-chip cooling will remain significant.
High-Performance Computing
HPC was one of the earliest commercial use cases for immersion cooling because research and engineering clusters often operate processors at sustained high utilization. Universities, laboratories and technical computing organizations therefore provide an established adoption base.
HPC buyers evaluate thermal stability, performance consistency, infrastructure footprint and the ability to operate processors at higher sustained loads. The segment also acts as a proving ground for technologies later deployed in hyperscale AI infrastructure.
Hyperscale and Cloud Data Centers
Hyperscale operators represent a major long-term opportunity but typically require stronger interoperability, vendor qualification and operational standardization than small specialized installations. Large fleet deployments also make maintenance procedures and hardware warranties critical.
OCP’s immersion program is specifically working on reference designs, component support and interoperability to reduce barriers to broader deployment. Standardized immersion-ready hardware could materially accelerate adoption during the forecast period.
Cryptocurrency Mining
Cryptocurrency mining helped establish early commercial demand because ASIC miners operate continuously and generate substantial heat. Immersion can reduce fan requirements, improve thermal stability and allow higher-density miner deployment.
The segment remains sensitive to cryptocurrency prices, electricity economics and regional regulation. As a result, it should not be used as the principal basis for long-term immersion cooling forecasts despite its historical contribution.
Edge and Telecom Computing
Edge computing creates smaller but technically attractive use cases where equipment operates in space-constrained or environmentally difficult locations. Immersion can protect electronics from dust and other contaminants while reducing dependence on traditional computer-room cooling.
Modular immersion platforms are increasingly being positioned for these deployments. Asperitas demonstrated its DFCX1 system for edge and on-site high-performance computing during Cisco Live 2026.
Deployment Analysis
New-Build Data Centers
New facilities offer the strongest opportunity to optimize electrical, mechanical and thermal infrastructure around immersion from the design stage. Operators can plan appropriate floor layouts, heat rejection, fluid handling and high-density power distribution without adapting legacy infrastructure.
AI-focused data centers are particularly relevant because cooling architecture is increasingly selected alongside GPU, networking and power design. This allows immersion economics to be evaluated against direct-to-chip cooling before construction begins.
Retrofit Deployments
Retrofit projects allow operators to increase compute density without constructing an entirely new facility, but installation can be more complex. Existing floor loads, plumbing, electrical distribution and server compatibility must be assessed before deployment.
Retrofits are attractive where data halls are power constrained or unable to support additional air-cooled racks. Vendors that offer modular tanks and prevalidated immersion-ready servers can reduce integration risk.
Immersion Cooling Market Dynamics
AI Compute Density Accelerates Liquid Cooling Adoption
AI training and inference infrastructure is increasing processor power and rack density, creating thermal conditions that make conventional air cooling progressively less attractive. This is shifting data center design toward direct liquid cooling and immersion.
The commercial opportunity is not determined only by cooling efficiency. Buyers compare rack density, CapEx, OpEx, floor-space savings, power availability, maintenance procedures and compatibility with next-generation accelerators.
Energy and Water Efficiency Strengthen the Business Case
Immersion cooling removes or reduces server fans and can operate with higher coolant temperatures than many traditional cooling systems. This can reduce cooling-related electrical consumption and create opportunities for dry cooling or heat reuse.
Actual savings vary by site, climate, workload and thermal architecture. Buyers increasingly require transparent TCO models rather than generalized efficiency claims.
Standardization Is Improving
The older market page describes lack of standardization as a major constraint. That remains relevant, but the market has progressed substantially through OCP specifications, immersion requirements, fluid specifications, warranty guidance and component-compatibility work.
The remaining challenge is translating open specifications into widespread server warranties, multi-vendor interoperability and repeatable deployment practices. Standardization progress should therefore be treated as an adoption catalyst as well as an unresolved issue.
Direct-to-Chip Cooling Creates Competitive Pressure
Immersion does not compete only with air cooling. Direct-to-chip systems are becoming an important alternative for high-density AI infrastructure and can allow conventional server servicing while removing heat directly from GPUs and CPUs.
The winning architecture will vary by workload, rack density, facility design and operating model. Many data center operators are likely to deploy multiple liquid-cooling architectures rather than standardizing globally on one technology.
Technology Comparison
| Metric | Single-Phase Immersion | Two-Phase Immersion | Direct-to-Chip |
| Cooling Medium | Dielectric Liquid | Phase-Change Dielectric Fluid | Water/Glycol or Other Coolant |
| IT Equipment Submerged | Yes | Yes | No |
| Thermal Density Potential | High | Very High | High |
| Service Model Change | Significant | Significant | Moderate |
| Fluid Complexity | Moderate | High | Lower |
| Heat-Reuse Potential | High | High | High |
| Primary Opportunity | AI, HPC, Edge, Dense Compute | Extreme-Density Compute | AI and Hyperscale Racks |
Immersion Cooling Market Regional Analysis
North America
North America represents one of the most important commercial markets because of hyperscale data centers, AI infrastructure investment, semiconductor innovation and large HPC deployments. The United States is particularly attractive for high-density GPU clusters and specialized AI cloud infrastructure.
The market also benefits from a strong supplier ecosystem covering immersion tanks, fluids, servers, heat rejection and facility integration. The February 2026 agreement for Trane Technologies to acquire LiquidStack illustrates growing strategic interest from larger thermal-management companies.
North American growth will depend on hyperscale qualification, hardware warranties and competition from direct-to-chip architectures. Large customers increasingly expect integrated cooling solutions rather than standalone tanks.
Europe
Europe has an established immersion cooling ecosystem supported by dense colocation markets, HPC activity and strong energy-efficiency requirements. The supplied page identifies the region as an important market but relies on data center counts and cryptocurrency discussions that are several years old.
The Netherlands, Germany, France, the United Kingdom and Nordic countries are important deployment markets. Europe is also home to suppliers such as Asperitas and Submer, while sustainability requirements can strengthen interest in lower-energy cooling and waste-heat reuse.
The regional opportunity is increasingly linked to AI and high-density compute rather than cryptocurrency mining. Regulatory treatment of energy consumption, water and cooling fluids will influence technology selection.
Asia-Pacific
Asia-Pacific offers significant long-term growth through rapid data center expansion in China, Japan, South Korea, Singapore, India and Southeast Asia. Increasing AI adoption and cloud infrastructure development are driving demand for thermal architectures capable of supporting high-density accelerators.
Japan has historical immersion-cooling experience through technology suppliers such as Fujitsu, while Southeast Asia is becoming more important for new data center capacity. The region also benefits from proximity to server, electronics and component manufacturing ecosystems.
Commercial adoption remains uneven because energy markets, data center standards, fluid availability and hyperscale requirements differ substantially between countries.
Middle East
The Middle East is becoming a high-potential market as Gulf countries invest in AI infrastructure, hyperscale cloud platforms and sovereign computing capacity. High ambient temperatures make cooling performance and water efficiency particularly important.
Saudi Arabia and the UAE provide attractive conditions for advanced cooling because new campuses can be designed around high-density computing from the beginning. However, long-term fluid performance and heat-rejection design must be optimized for extreme climates.
Latin America
Latin America remains an emerging immersion cooling market led by Brazil, Mexico and selected hyperscale and colocation projects. Growing cloud usage and AI adoption are increasing data center power density, although conventional cooling remains dominant.
Adoption is likely to begin with specialized HPC, enterprise and cryptocurrency applications before expanding into broader AI infrastructure. Import costs and limited local immersion ecosystems may restrict near-term growth.
Regional Commercial Comparison
| Region | Commercial Position | Main Demand Driver | Main Constraint |
| North America | Major Growth Market | AI and Hyperscale Compute | Architecture Competition |
| Europe | Established Immersion Ecosystem | Efficiency and HPC | Regulation and Retrofit Complexity |
| Asia-Pacific | High-Growth Opportunity | Cloud and AI Expansion | Uneven Standards and Access |
| Middle East | Emerging Premium Market | Sovereign AI and New Data Centers | Extreme Climate Design |
| Latin America | Early-Stage Market | Cloud and HPC Growth | Limited Local Ecosystem |
Immersion Cooling Market Competitive Landscape
The competitive landscape should be organized by immersion-platform specialists, integrated thermal-management companies, fluid suppliers and server/infrastructure partners. The supplied page lists Fujitsu, GRC, Submer, Asperitas, LiquidStack, Midas, Iceotope, LiquidCool Solutions and DCX among major participants.
Submer
Submer is one of the major immersion cooling specialists serving data center, AI and high-performance computing deployments. Its positioning combines immersion infrastructure with broader thermal-management and data center optimization capabilities.
The company benefits from strong participation in the Open Compute Project ecosystem and immersion-standardization initiatives. Its competitive position is strengthened where customers prioritize modular infrastructure, high rack density and open hardware compatibility.
Future growth depends on hyperscale qualification, ecosystem partnerships and the ability to compete against both immersion specialists and direct-to-chip cooling providers.
Green Revolution Cooling
Green Revolution Cooling, or GRC, is an established immersion cooling specialist with experience across enterprise, HPC and data center environments. The company has historically focused on single-phase liquid immersion architectures.
Its competitive strengths include operational experience, system-level expertise and a long history of immersion deployment. Future differentiation will increasingly depend on AI-ready density, serviceability and integration with modern accelerator platforms.
Asperitas
Asperitas develops modular immersion cooling systems for enterprise, cloud, HPC and edge environments. Its current strategy emphasizes adaptive systems and integration with immersion-ready compute infrastructure.
In February 2026, Asperitas and UNICOM Engineering announced a commercial agreement allowing customers to procure immersion systems and immersion-ready servers under a single contract. This reduces procurement complexity and addresses one of the sector’s major integration barriers.
The company is also advancing its DFCX architecture for higher-density and edge computing deployments. Its growth depends on channel partnerships and broader OEM validation.
LiquidStack
LiquidStack provides immersion and other liquid-cooling systems for AI, hyperscale and high-performance data centers. Its portfolio increasingly spans immersion and direct-to-chip technologies rather than relying on one cooling architecture.
In February 2026, Trane Technologies announced an agreement to acquire LiquidStack, strengthening LiquidStack’s access to global thermal-management infrastructure, heat rejection and customer channels.
The transaction illustrates industry consolidation as large HVAC and data center infrastructure companies seek broader exposure to AI thermal management.
Fujitsu
Fujitsu has historical experience with liquid immersion through its PRIMERGY immersion cooling system. The supplied page describes the system as submerging IT hardware in cooling fluid and reducing cooling-related infrastructure requirements.
Fujitsu’s advantage lies in combining server expertise with infrastructure integration rather than operating solely as a cooling-system specialist. Its competitive relevance should be evaluated against the current availability and commercial scale of its immersion portfolio.
Iceotope
Iceotope focuses primarily on precision liquid cooling and sealed cooling architectures for high-power computing systems. Its technology is relevant to the broader competitive landscape because data center operators frequently compare sealed liquid cooling, direct-to-chip and immersion approaches when planning AI infrastructure.
The company competes on thermal efficiency, serviceability and reduced cooling-resource consumption. It should therefore be positioned as an adjacent liquid-cooling competitor rather than automatically classified as a pure full-server immersion provider.
Fluid Suppliers
Cooling-fluid companies are increasingly strategic participants because dielectric fluid performance affects reliability, materials, servicing and long-term system economics. The market includes specialized products from established lubricant and energy companies as well as engineered-fluid suppliers.
OCP currently recognizes immersion fluids designed around properties such as dielectric strength, oxidation stability, thermal performance and component compatibility. Fluid qualification is becoming a core part of vendor selection.
Competitive Positioning Matrix
| Company | Primary Position | Immersion Focus | AI/HPC Exposure | Strategic Differentiator |
| Submer | Immersion Infrastructure | High | High | OCP Ecosystem and Integrated Platforms |
| GRC | Single-Phase Immersion | High | High | Deployment Experience |
| Asperitas | Modular Immersion | High | High | Integrated Server Partnerships |
| LiquidStack | Multi-Architecture Liquid Cooling | High | Very High | Immersion + Direct-to-Chip Portfolio |
| Fujitsu | Servers and Infrastructure | Selective | High | Server-System Integration |
| Iceotope | Precision Liquid Cooling | Adjacent | High | Sealed High-Density Cooling |
| Fluid Suppliers | Dielectric Fluids | Enabling | High | Fluid Chemistry and Compatibility |
Competitor Profile Writing Structure
Each competitor profile should begin with the company’s current immersion or liquid-cooling platform and the compute environments it serves. The first paragraph should explain whether the company is focused on single-phase immersion, two-phase immersion, direct-to-chip cooling or a combined architecture.
The second paragraph should describe its competitive strengths, including rack density, modularity, server compatibility, fluid ecosystem, deployment model, servicing and geographic reach. Statements about efficiency should be tied to verified product data rather than generic cooling claims.
The final paragraph should cover current partnerships, acquisitions, new product launches, AI infrastructure exposure and expected competitive position through 2035. Generic corporate-history paragraphs should be removed unless they directly explain the company’s cooling strategy.
Recent Developments – Immersion Cooling Market
1. July 2026 – Iceotope Advanced Liquid Cooling for Next-Generation AI Systems
Iceotope highlighted its precision liquid cooling architecture for increasingly dense AI infrastructure, emphasizing sealed cooling systems designed to handle high-power GPUs while reducing cooling energy and water requirements.
2. June 2026 – Asperitas Advanced High-Density Immersion Cooling Architecture
Asperitas continued advancing immersion cooling for high-density computing, emphasizing precise thermal management for AI and HPC workloads and the need to optimize cooling at the component level rather than only at the system level.
3. May 2026 – Iceotope Raised US$26 Million for Liquid Cooling Expansion
Iceotope secured US$26 million in Series B funding to accelerate product and engineering development, expand its patent portfolio, and strengthen ecosystem partnerships for next-generation liquid-cooled infrastructure.
4. February 2026 – Asperitas Partnered with UNICOM Engineering
Asperitas and UNICOM Engineering entered a commercial agreement allowing customers to procure immersion cooling systems and immersion-ready servers through a single contract, supporting easier deployment of high-density computing infrastructure.
5. January 2026 – Asperitas Launched Direct Forced Convection Immersion Cooling
Asperitas introduced its DFCX1 Direct Forced Convection immersion cooling product line, targeting higher compute densities and modular deployment across high-performance computing environments.
