Direct-to-Chip Cooling Market

Overview

The global Direct-to-Chip Cooling Market was valued at USD 2.5 billion in 2025 and is projected to reach USD 13.4 billion by 2034, growing at a CAGR of 20.5% during the forecast period (2026–2034). The market is driven by the sharp rise in thermal design power of artificial intelligence accelerators and central processing units, which has pushed rack power densities beyond the practical limits of air cooling and turned liquid contact with the processor package into a technical necessity rather than an efficiency preference. The market is shifting from conventional, single rack-level retrofit installations toward fully engineered, factory-integrated cooling architectures that are co-designed with server original equipment manufacturers and silicon vendors ahead of chip launch. Single-phase water-glycol loops remain the dominant technology in volume terms, but two-phase, waterless dielectric systems are advancing quickly as operators look to remove water-related leak risk and unlock heat-reuse opportunities, while coolant distribution unit vendors are increasingly consolidating liquid and air heat rejection into single hybrid platforms to simplify deployment across mixed-density facilities. Government-backed research programs are reinforcing this shift toward advanced direct-to-chip architectures. The U.S. Department of Energy's ARPA-E COOLERCHIPS program, which added Accelsius as a technology contributor in May 2025, is targeting a reduction in data center cooling energy consumption to below five percent of total information technology load, a goal that is accelerating collaborative development of hybrid direct-to-chip and air-assisted cooling systems for next-generation high-density computing. By region, North America held the largest share of the direct-to-chip cooling market in 2025, supported by concentrated hyperscale AI data center construction across the United States. Asia-Pacific is projected to grow at the fastest CAGR during the forecast period, driven by expanding AI compute capacity in China and India and rising coolant distribution unit manufacturing investment across the region.

Market Size & Share

Market Dynamics

KEY MARKET TREND

AI-Driven Cold Plate Advancement and the Rise of Two-Phase Architectures Emerging as a Transformational Trend

• Cold plate suppliers are re-engineering microchannel and split-flow geometries to keep pace with GPU thermal design power that has moved past 1,500 watts per socket. Manufacturers are combining computational fluid dynamics optimization with multi-directional flow paths to raise heat flux tolerance without increasing pressure drop across the cooling loop.
• Two-phase, waterless direct-to-chip systems are gaining commercial traction alongside mature single-phase designs, particularly among facilities seeking to eliminate water-related leak risk near costly accelerator hardware. Dielectric refrigerant-based platforms allow operators to reach higher rack densities while supporting heat-reuse applications that conventional single-phase water loops cannot easily deliver.
• Coolant distribution unit vendors are consolidating liquid and air cooling functions into single hybrid units to simplify deployment across mixed-density data halls. This convergence reduces the footprint required for retrofit projects and shortens the qualification cycle for operators transitioning existing air-cooled facilities toward direct-to-chip infrastructure.
• CoolIT Systems demonstrated a validated single-phase coldplate design capable of capturing more than 97 percent of the heat generated by a 4,000-watt thermal test vehicle, confirming that established direct-to-chip architectures can scale to cool the next generation of AI accelerators.

KEY MARKET DRIVER

Rising Thermal Design Power of AI Accelerators is the Key Driver

• Next-generation GPU and custom AI accelerator platforms are drawing well beyond 1,000 watts per chip, a level that conventional air cooling and rear-door heat exchangers can no longer manage economically. This has pushed hyperscalers and colocation operators to specify direct-to-chip cold plates as the default thermal architecture for new AI training and inference clusters.
• Data center electricity consumption is climbing sharply as AI workloads expand, with global data center demand having reached roughly 415 terawatt-hours in 2024 and continuing to grow at a pace far faster than overall electricity use. Direct-to-chip cooling helps operators contain this growth by supporting warmer facility water temperatures and reducing mechanical chiller reliance.
• Silicon vendors and server original equipment manufacturers are co-designing cold plates and coolant distribution units with cooling specialists before chip launch, shortening qualification timelines for new accelerator generations. This close integration is turning direct-to-chip cooling into a standard line item in server bills of materials rather than an optional add-on.
• JetCool, a Flex company, announced a collaboration with Broadcom to co-develop direct-to-chip thermal architectures for next-generation AI application-specific integrated circuit platforms, aligning cold plate design with silicon packaging early in development.

KEY MARKET OPPORTUNITY

Expansion of Waterless and Heat-Reuse Cooling Architectures Creating New Revenue Streams

• Waterless, dielectric-based direct-to-chip systems are opening opportunities beyond traditional hyperscale campuses, including telecommunications facilities and edge sites where water access or floor space is constrained. Vendors are packaging these systems with heat-recovery options that let operators redirect captured heat toward adjacent office or district heating loads.
• Investors are backing specialist direct-to-chip cooling companies as AI infrastructure spending accelerates, providing capital for manufacturing capacity expansion and faster product development cycles. This funding is enabling smaller technology providers to scale production and compete for large multi-megawatt deployment contracts previously reserved for established infrastructure vendors.
• Retrofit opportunities are growing as operators of existing air-cooled facilities seek to add direct-to-chip capacity without full-facility rebuilds, favoring modular coolant distribution units that integrate with legacy chilled-water plants. This is creating demand for hybrid and hot-swappable equipment designed specifically for phased infrastructure upgrades.
• Carrier Global Corporation's venture arm led an investment and technology partnership with ZutaCore to advance waterless, two-phase direct-to-chip cooling development, reflecting growing strategic investor interest in next-generation thermal management for data centers. 

Direct-to-Chip Cooling Market Size, 2025-2034 (USD Billion)

Segmentation Analysis

Analysis by Cooling Technology

Single-phase direct-to-chip cooling held the largest market share in 2025 because it is the most mature and widely deployed liquid cooling technology, already implemented across millions of AI accelerators and CPUs in production data centers. Its use of water-glycol coolant circulated through microchannel cold plates offers proven reliability, lower component cost, and straightforward integration with existing chilled-water infrastructure. Server original equipment manufacturers have standardized cold plate designs for major processor sockets, reducing qualification time for new deployments. Continued engineering advances, including split-flow and multi-channel geometries validated for thermal loads above 4,000 watts, are extending the technology's runway well into the next generation of high-density AI accelerator platforms.

Two-phase direct-to-chip cooling is projected to grow at the fastest CAGR during the forecast period as operators seek waterless architectures that eliminate leak-related corrosion and condensation risk near expensive AI hardware. The technology uses a dielectric refrigerant that boils on contact with the chip surface, absorbing substantially more heat per unit of fluid than single-phase water loops and enabling higher rack densities within the same footprint. Growing vendor investment, including new row-based and end-of-row coolant distribution units launched through 2025 and 2026, is expanding two-phase system availability. Heat-reuse compatibility and reduced water consumption are further strengthening its appeal for telecommunications and sustainability-focused deployments.

Cooling Technology categories include

• Single-Phase Direct-to-Chip Cooling (Dominating Segment)
• Two-Phase Direct-to-Chip Cooling (Highest CAGR Segment)

Analysis by Component

Cold plates held the largest share of the direct-to-chip cooling market in 2025, reflecting their role as the essential interface component that every direct liquid cooling deployment requires regardless of coolant distribution unit design or facility architecture. Demand has been reinforced by the precision engineering required to match cold plate geometry to each new generation of CPU and GPU packages, with manufacturers producing millions of units annually for hyperscale customers. Advanced manufacturing techniques, including microchannel etching and split-flow structures, allow cold plates to manage thermal fluxes exceeding 300 watts per square centimeter, supporting the highest-power accelerators entering production today.

Coolant distribution units are projected to grow at the fastest CAGR during the forecast period as operators scale from single-rack pilots to megawatt-class, facility-wide liquid cooling deployments. Vendors have introduced increasingly powerful modular and row-based units, with several manufacturers now offering multi-megawatt platforms capable of serving dozens of racks from a single system. This scalability reduces the number of discrete units operators must install, commission, and maintain across large AI data halls. Growing integration of hybrid liquid-to-air functionality within coolant distribution units is further widening their applicability across both new-build and retrofit facility types.

Component categories include

• Cold Plates (Dominating Segment)
• Coolant Distribution Units (Highest CAGR Segment)
• Manifolds and Quick Disconnects
• Pumps
• Others (Heat Exchangers and Sensors)

Analysis by Coolant Type

Water-based coolants, typically deionized water blended with glycol and corrosion inhibitors, held the largest share of the direct-to-chip cooling market in 2025 due to their high heat capacity, low cost, and long operating history across enterprise and hyperscale computing environments. Facility teams are already familiar with water treatment, monitoring, and filtration practices, which lowers the operational barrier to adoption compared with newer fluid chemistries. Coolant distribution units isolate the water-based technology cooling system loop from the facility water system, containing purity requirements to a closed secondary circuit and reducing the risk of contamination reaching sensitive processor cold plates.

Dielectric fluids are projected to grow at the fastest CAGR during the forecast period as two-phase and waterless single-phase systems gain adoption among operators prioritizing leak safety near high-value AI hardware. Because dielectric fluids are electrically non-conductive, a leak does not risk short-circuiting server components, which is reshaping risk assessments for liquid cooling near dense GPU clusters. Suppliers are expanding dielectric fluid portfolios with lower global-warming-potential formulations to meet emerging sustainability requirements. This combination of safety and environmental positioning is accelerating dielectric fluid adoption across new two-phase coolant distribution unit deployments launched through 2025 and 2026.

Coolant Type categories include

• Water-Based Coolants (Dominating Segment)
• Dielectric Fluids (Highest CAGR Segment)
• Refrigerants
• Others

Analysis by End-Use Industry

Hyperscale data centers held the largest share of the direct-to-chip cooling market in 2025, driven by the concentration of large-scale AI training and inference infrastructure among a small number of cloud and technology companies. These operators deploy direct-to-chip cooling at gigawatt scale, working directly with cold plate and coolant distribution unit manufacturers to co-engineer systems for specific accelerator platforms before general availability. Their purchasing volume gives hyperscalers significant influence over component standardization, pricing, and supply chain prioritization, reinforcing their position as the anchor customer segment for direct-to-chip cooling equipment manufacturers worldwide.

Colocation data centers are projected to grow at the fastest CAGR during the forecast period as multi-tenant operators race to offer direct-to-chip-ready suites for AI-focused enterprise and neocloud customers who lack the capital to build dedicated facilities. Colocation providers are retrofitting existing halls with coolant distribution units and facility water upgrades to support high-density racks without requiring tenants to manage liquid cooling infrastructure themselves. This shift is turning liquid cooling readiness into a competitive differentiator across the colocation industry, prompting accelerated capital investment in direct-to-chip-compatible power and cooling infrastructure across major markets.

End-Use Industry categories include

• Hyperscale Data Centers (Dominating Segment)
• Colocation Data Centers (Highest CAGR Segment)
• High-Performance Computing Centers
• Enterprise Data Centers
• Telecommunications Infrastructure

By Region

Direct-to-Chip Cooling Market Share 2025, (CAGR)

North America

38%

South America

xx%

Europe

xx%

Middle East Africa

xx%

Asia Pacific

30%

North America held the largest share of the direct-to-chip cooling market in 2025, accounting for approximately 38% of global revenue, anchored by the United States' concentration of hyperscale AI data center construction. The United States alone represented the single largest national market, supported by continued build-out from major cloud providers, growing colocation liquid cooling retrofits, and government-backed research initiatives such as the Department of Energy's ARPA-E COOLERCHIPS program advancing next-generation hybrid cooling architectures. Leading cold plate and coolant distribution unit manufacturers, including CoolIT Systems, Vertiv, and LiquidStack, maintain significant manufacturing and engineering operations across the region. Canada is contributing through growing AI compute investment, while Mexico is emerging as a secondary manufacturing and data center location supported by nearshoring trends benefiting the broader North American technology supply chain.

Asia-Pacific is projected to grow at the fastest CAGR during the forecast period, accounting for approximately 30% of global revenue in 2025, driven by rapid AI data center expansion across China, India, Japan, and South Korea. China's large domestic cloud and hyperscale operators are accelerating coolant distribution unit procurement, with domestic manufacturers such as Huawei Digital Power and Envicool scaling production to meet surging demand, including reported supply discussions with major international cloud customers in 2026. India is emerging as a fast-growing country market within the region as domestic and international operators expand AI-ready data center capacity. Japan and South Korea continue to contribute through precision component manufacturing and adoption among semiconductor and electronics companies deploying high-density computing infrastructure.

Countries and Regions Covered

North America (Dominating Region)

• United States (Largest Country Market)
• Canada
• Mexico

Asia-Pacific (Fastest Growing Region)

• China (Largest Country Market)
• India (Fastest-Growing Country Market)
• Japan
• South Korea
• Rest of Asia-Pacific

Europe

• Germany (Largest Country Market)
• France
• United Kingdom
• Italy
• Rest of Europe

Latin America

• Brazil (Largest Country Market)
• Chile (Fastest-Growing Country Market)
• Rest of Latin America

Middle East & Africa

• Saudi Arabia (Largest Country Market)
• United Arab Emirates (Fastest-Growing Country Market)
• Rest of Middle East & Africa

Market Share

The direct-to-chip cooling market is consolidated, with a group of established thermal management and data center infrastructure companies, including Vertiv, Schneider Electric, Eaton, and Delta Electronics, competing alongside specialized liquid cooling pure-plays such as CoolIT Systems, LiquidStack, ZutaCore, and Accelsius. Competitive intensity is increasing as diversified infrastructure vendors acquire specialist cooling technology firms to build end-to-end portfolios, illustrated by Schneider Electric's integration of Motivair and Daikin Applied's acquisition of Chilldyne. Key success factors include early co-design relationships with silicon vendors and server manufacturers, manufacturing scale to meet hyperscale volume commitments, and global service coverage for mission-critical deployments. Leading companies are prioritizing capacity expansion, two-phase and hybrid product development, and strategic partnerships with GPU and server original equipment manufacturers to secure long-term supply agreements ahead of next-generation accelerator launches.

Key Players

• CoolIT Systems Inc. (Canada)
• Vertiv Holdings Co (US)
• Schneider Electric SE – Motivair (France)
• Eaton Corporation plc – Boyd Thermal (Ireland)
• Jet Cool Technologies (US)
• LiquidStack (US)
• Iceotope Technologies Limited (UK)
• nVent Electric plc (UK)
• Zuta CoreLtd. (US)
• Delta Electronics Inc. (Taiwan)
• Accelsius LLC (US)
• STULZ GmbH (Germany)
• Rittal GmbH & Co. KG (Germany)
• Chilldyne, Inc. – a Daikin Applied company (US)
• Huawei Digital Power (China)
• Shenzhen Envicool Technology Co., Ltd. (China)

Recent Market Developments

• In January 2026, LiquidStack secured a 300-megawatt Coolant Distribution Unit order from a major U.S.-based data center operator for its CDU-1MW platform, supporting large-scale AI-ready deployments and underscoring accelerating direct-to-chip demand across hyperscale customers.
• In June 2025, LiquidStack unveiled the GigaModular CDU, described as the industry's first modular, scalable coolant distribution unit, offering up to 10MW of cooling capacity through a pay-as-you-grow installation approach, showcased at the Datacloud Global Congress in Cannes.
• In October 2025, Accelsius announced general availability of the NeuCool MR250, its first row-based two-phase coolant distribution unit, delivering up to 250kW of liquid cooling capacity per rack for AI and high-performance computing workloads.
• In November 2025, Daikin Applied Americas acquired Chilldyne, adding negative-pressure direct-to-chip liquid cooling technology to its data center cooling portfolio and expanding its end-to-end thermal management offering for hyperscale and AI facilities.

Frequently Asked Questions

Key Questions Answered

Request a Sample

1

What is Direct-to-Chip Cooling?

2

What is the CAGR of the Direct-to-Chip Cooling Market?

3

Which component leads the Direct-to-Chip Cooling Market?

4

Which end-use industry dominates the Direct-to-Chip Cooling Market?

5

Which coolant type has the highest market share?

6

What are the latest trends in the Direct-to-Chip Cooling Market?

7

Who are the end users of Direct-to-Chip Cooling?

Why Choose IG Transformation

Speak to Analyst

Strong Industry Focus

Extensive Product Offerings

Customer Research Services

Robust Research Methodology

Comprehensive Reports

Latest Technological Developments

Value Chain Analysis

Potential Market Opportunities

Growth Dynamics

Quality Assurance

Post-sales Support

Regular Report Updates