Published:  05, Sep 2026

United States AI Data Center Liquid Cooling Market

United States AI Data Center Liquid Cooling Market Size, Share and Analysis By Cooling Technology (Direct-to-Chip Cooling, Immersion Cooling, Rear Door Heat Exchangers, Hybrid Cooling Systems), By Component (Hardware Solutions, Services), By Data Center Type (Hyperscale Data Centers, Colocation Data Centers, Enterprise Data Centers, Edge Data Centers), By Coolant Type (Water-based Coolants, Dielectric Fluids, Refrigerant-based Coolants), By End User (Cloud Service Providers, Colocation Providers, Enterprise Data Centers, Government and Defense), and Regional Forecast Till 2034

Download Free PDF
banner icon
Market Size (2025):

USD 1.3 Billion

banner icon
CAGR (2026–2034):

27.4%

banner icon
Report Pages:

160-170

banner icon
Market Tables:

50-60

Overview

The United States AI Data Center Liquid Cooling Market was valued at USD 1.3 billion in 2025 and is projected to reach USD 11.8 billion by 2034, growing at a CAGR of 27.4% during the forecast period (2026-2034). The market is driven by the rapid deployment of high-thermal-density GPU accelerators, federal permitting reform for AI data center infrastructure, and expanding hyperscale campus construction across the country. Demand for liquid-cooled thermal management has moved from a niche high-performance computing requirement to a mainstream design standard as GPU thermal design power routinely exceeds 700 watts per chip, a threshold beyond which conventional air handling cannot reliably maintain safe operating temperatures at scale. The market is shifting from conventional, air-assisted and hybrid retrofit cooling toward purpose-built, factory-integrated direct-to-chip and immersion architectures designed alongside next-generation GPU platforms such as NVIDIA's Blackwell and Vera Rubin families. Operators are moving away from chilled-water-only designs toward warm-water loops operating near 45 degrees Celsius, reducing chiller dependence and water withdrawal. Government initiatives such as Executive Order 14318, Accelerating Federal Permitting of Data Center Infrastructure, signed on July 23, 2025, are streamlining environmental review and expediting construction of AI data center projects exceeding 100 megawatts, while the Department of Energy's July 2025 selection of four federal sites, including Idaho National Laboratory and the Savannah River Site, opens federal land to co-located AI data center and power generation development. The accompanying America's AI Action Plan, released the same day, directs the Environmental Protection Agency, Department of Commerce, and other federal agencies to expand categorical exclusions under the National Environmental Policy Act, while H.R. 5332, the Liquid Cooling for AI Act of 2025, signals emerging congressional attention to thermal-efficiency standards for AI-supporting facilities. By region, the South held the largest share of the market in 2025, anchored by Northern Virginia's concentrated hyperscale corridor and large-scale Texas AI campuses. The West is projected to be the fastest-growing region during the forecast period, propelled by new hyperscale AI construction across Arizona, Nevada, and Oregon, where available power capacity and land are attracting gigawatt-scale liquid-cooled data center investment.

Market Size & Share

CAGR (2026–2034):

Market Snapshot

Study Period: 2021-2034
Market Size in 2025: USD 1.3 Billion
Market Size in 2026: USD 1.7 Billion
Market Size by 2034: USD 11.8 Billion
Unit Value: USD Billion
Projected CAGR: 27.4% (2026-2034)
Largest Region: South
Fastest-Growing Region: West
Fastest-Growing Cooling Technology: Immersion Cooling

Market Dynamics

KEY MARKET TREND

Rise of Two-Phase and Warm-Water Direct-to-Chip Cooling Reshaping Thermal Architecture

  • Server manufacturers are re-engineering coolant loops to operate at warmer inlet temperatures instead of relying on continuously chilled water. Supermicro's DLC-2 platform now supports facility water entering at up to 45 degrees Celsius, cutting chiller reliance and reducing data center water consumption by as much as 40% compared with conventional chilled installations.
  • Two-phase dielectric cooling is moving from pilot projects into rack-scale commercial deployment as chip thermal design power climbs past 1,000 watts. Accelsius' NeuCool platform has demonstrated cooling of 4,500 watts per GPU socket using 40 degrees Celsius facility water, removing the need for chilled-water infrastructure across an entire 250-kilowatt AI rack.
  • Engineering standards bodies are formalizing guidance to support consistent liquid-cooled facility design nationwide. The Telecommunications Industry Association published TIA-942-C with a dedicated annex addressing liquid immersion cooling, aligning it with ASHRAE thermal guidelines and giving operators a shared reference framework for auditing and qualifying high-density liquid-cooled deployments.
  • Representative Jay Obernolte introduced H.R. 5332, the Liquid Cooling for AI Act of 2025, in the House of Representatives on September 11, 2025. The bill directs the Comptroller General to assess liquid-cooling technology adoption and develop federal best-practice guidance for AI and high-performance computing facilities.

 

KEY MARKET DRIVER

Escalating AI Accelerator Power Density and Federal Infrastructure Policy Driving Market Growth

  • Modern AI accelerators such as NVIDIA's Blackwell and Vera Rubin platforms generate thermal loads exceeding 1,000 watts per chip, well beyond what computer room air handlers can dissipate. Facility operators increasingly treat liquid cooling as a baseline design requirement rather than an optional upgrade for any rack exceeding roughly 80 kilowatts.
  • US data center electricity consumption rose from approximately 76 terawatt-hours in 2018 to roughly 200 terawatt-hours in 2025, according to Department of Energy and Lawrence Berkeley National Laboratory analysis. Cooling systems account for an estimated 25% to 40% of that total electricity draw, making thermal efficiency a direct lever on operating cost.
  • President Trump's Executive Order 14318, signed July 23, 2025, directs federal agencies to streamline environmental review and expand categorical exclusions for AI data center projects exceeding 100 megawatts of new load. The accompanying America's AI Action Plan further calls for a nationwide permitting pathway to accelerate qualifying data center construction.
  • The Department of Energy selected four federal sites, Idaho National Laboratory, Oak Ridge Reservation, the Paducah Gaseous Diffusion Plant, and Savannah River Site, on July 24, 2025, to host AI data center and co-located power generation development. The announcement directly supports liquid-cooling infrastructure demand tied to new federally sited AI compute capacity.

 

KEY MARKET OPPORTUNITY

Expansion of Colocation Retrofits and Cooling-as-a-Service Models Creating New Revenue Streams

  • Colocation operators are retrofitting existing air-cooled halls with direct-to-chip and rear-door liquid loops to capture AI tenant demand without building new campuses. Digital Realty reported that liquid-cooling deployments across its North American portfolio grew 320% year-over-year in 2025, pointing to substantial retrofit opportunity for CDU and manifold suppliers.
  • Subscription-style cooling-as-a-service offerings are emerging to lower the capital barrier for mid-market and enterprise operators adopting liquid cooling for the first time. Vendors bundling tanks, coolant, installation, and ongoing maintenance for a recurring fee are shifting liquid cooling from a capital expenditure into an operating expense.
  • Heat-reuse integration is opening a secondary revenue path for liquid-cooled facilities located near district heating networks, greenhouses, or industrial users. Because liquid coolant carries waste heat at higher, more usable temperatures than exhaust air, operators can now monetize a byproduct that air-cooled facilities simply vent.
  • Colovore secured a USD 925 million debt financing round in 2025 to expand its liquid-cooled colocation capacity in Santa Clara, California, one of the largest raises dedicated specifically to liquid-cooled colocation infrastructure. The deal signals institutional capital's growing confidence in liquid-cooled colocation as a durable asset class.
United States AI Data Center Liquid Cooling Market Size, 2025-2034 (USD Billion)

Segmentation Analysis

Analysis by Cooling Technology

Direct-to-chip cooling held the largest market share in 2025 because it delivers targeted thermal management directly to the GPU and CPU die while preserving the familiar rack-and-server operating model that most US hyperscale and colocation operators already use. Cold plates mounted on Blackwell and Hopper-generation accelerators handle thermal design power well above 700 watts per chip, and nearly every major server OEM, including Dell, HPE, and Supermicro, now ships direct-to-chip-ready platforms as standard configurations. Because retrofitting an existing air-cooled hall for direct-to-chip cooling requires less structural change than installing immersion tanks, operators have adopted it as the default pathway for scaling AI racks quickly across new and existing US facilities.

 

Immersion cooling is projected to grow at the fastest CAGR during the forecast period as US hyperscalers and colocation providers pursue higher rack densities than cold plates alone can economically support. Vendors such as GRC and Submer are expanding tank-based platforms capable of cooling up to 100 kilowatts per rack while removing nearly all mechanical airflow components, cutting both energy use and noise. Partnerships such as CleanSpark's selection of Submer for North American AI data center expansion illustrate how immersion cooling is moving from niche high-performance computing use into mainstream AI infrastructure build-outs, particularly at greenfield sites where facility design can be optimized around tank-based cooling loops from the outset.

 

Cooling Technology categories include

                 ·           Direct-to-Chip Cooling (Dominating Segment)

                 ·           Immersion Cooling (Highest CAGR Segment)

                 ·           Rear Door Heat Exchangers

                 ·           Hybrid Cooling Systems

 

Analysis by Component

Hardware solutions, including cold plates, coolant distribution units, manifolds, heat exchangers, and pumps, held the largest market share in 2025 because every liquid-cooled rack requires this physical equipment regardless of the cooling architecture chosen. US data center operators are purchasing complete hardware packages, such as Supermicro's DLC-2 system and Vertiv's coolant distribution units extending to roughly 2.3 megawatts, to shorten deployment timelines from months to weeks. As GPU thermal design power continues to climb, operators are also replacing first-generation hardware with higher-capacity coolant distribution units and larger-diameter manifolds, sustaining continued hardware-segment revenue even inside already-built facilities.

 

Services, spanning installation, integration, monitoring, and maintenance, are projected to grow at the fastest CAGR during the forecast period as the installed base of liquid-cooled US data centers expands and requires ongoing technical support. Operators lacking in-house liquid cooling expertise increasingly rely on vendor-delivered commissioning, leak-detection monitoring, and coolant maintenance contracts to manage operational risk in facilities that previously relied only on air handling. The emergence of subscription-style cooling-as-a-service contracts, which bundle equipment, fluid, and maintenance into a single recurring fee, is further accelerating services-segment growth among mid-market and colocation operators new to liquid cooling.

 

Component categories include

                 ·           Hardware Solutions (Dominating Segment)

                 ·           Services (Highest CAGR Segment)

 

Analysis by Data Center Type

Hyperscale data centers held the largest market share in 2025 because cloud platform operators such as Microsoft, Amazon, and Google are deploying the largest volume of GPU-dense AI training and inference clusters in the country. These operators commit to multi-year, gigawatt-scale liquid cooling procurement directly with vendors, giving hyperscale facilities outsized influence over cooling technology standards adopted across the broader US market. Microsoft's Build 2025 commitment to zero-waste water cooling for all new data center designs illustrates how hyperscale operators are setting design precedents that influence coolant loop architecture across the wider supplier ecosystem.

 

Colocation data centers are projected to grow at the fastest CAGR during the forecast period as enterprises without the capital or scale to build private AI campuses turn to colocation providers for liquid-cooled capacity. Digital Realty reported that liquid-cooling deployments across its North American portfolio grew 320% year-over-year in 2025, reflecting surging tenant demand for direct-to-chip and rear-door retrofit capacity. Colocation operators are increasingly pre-building liquid-ready shells and coolant distribution plants speculatively, competing for AI tenants that require rack densities their existing enterprise customers never approached.

 

Data Center Type categories include

                 ·           Hyperscale Data Centers (Dominating Segment)

                 ·           Colocation Data Centers (Highest CAGR Segment)

                 ·           Enterprise Data Centers

                 ·           Edge Data Centers

 

Analysis by Coolant Type

Water-based coolants held the largest market share in 2025 because single-phase water loops remain the lowest-cost, best-understood coolant option for direct-to-chip and rear-door cooling deployments across existing US facilities. Facility engineers are broadly familiar with water treatment, corrosion inhibition, and leak-management practices, reducing operational risk relative to newer fluid chemistries. Supermicro's warm-water DLC-2 platform, supporting inlet temperatures up to 45 degrees Celsius, demonstrates how water-based loops continue to advance in efficiency, allowing operators to reduce chiller dependence without switching to an entirely different coolant class.

 

Dielectric fluids are projected to grow at the fastest CAGR during the forecast period as two-phase and immersion cooling architectures scale to support next-generation GPU thermal loads. Accelsius' NeuCool platform has demonstrated cooling of 4,500 watts per GPU socket using engineered dielectric fluid, a density that single-phase water loops cannot economically match. As chip vendors continue raising thermal design power for successive GPU generations, US operators are increasingly evaluating dielectric-fluid systems for the highest-density racks in new AI-optimized facilities.

 

Coolant Type categories include

                 ·           Water-based Coolants (Dominating Segment)

                 ·           Dielectric Fluids (Highest CAGR Segment)

                 ·           Refrigerant-based Coolants

 

Analysis by End User

Cloud service providers held the largest market share in 2025, reflecting their position as the primary buyers of GPU-dense AI training and inference infrastructure across the United States. These operators negotiate direct supply agreements with cooling vendors, exemplified by Vertiv's July 2025 partnership with Oklo to co-develop integrated power and cooling solutions for US hyperscale operators facing grid constraints. Because cloud service providers deploy the largest and most standardized fleets of liquid-cooled racks, their procurement decisions continue to shape product roadmaps across the broader liquid cooling supplier base.

 

Government and defense end users are projected to grow at the fastest CAGR during the forecast period as federal agencies expand AI computing capacity under the Department of Energy's July 2025 site-selection program covering federal locations such as Idaho National Laboratory and the Savannah River Site. Defense modernization programs incorporating AI-enabled systems are similarly driving demand for liquid-cooled compute at secure government facilities. Expedited federal permitting under Executive Order 14318 is expected to accelerate construction timelines for these federally sited, liquid-cooled AI data center projects through the forecast period.

 

End User categories include

                 ·           Cloud Service Providers (Dominating Segment)

                 ·           Government and Defense (Highest CAGR Segment)

                 ·           Colocation Providers

                 ·           Enterprise Data Centers

By Region

United States AI Data Center Liquid Cooling Market Regional Analysis

United States AI Data Center Liquid Cooling Market Share 2025, (CAGR)
world map
location map

North America

XX%

location map

South America

XX%

location map

Europe

XX%

location map

Middle East Africa

XX%

location map

Asia Pacific

XX%

Regional Analysis

The South held the largest share of the United States AI Data Center Liquid Cooling Market in 2025, driven by Northern Virginia's Data Center Alley, the most concentrated hyperscale corridor in the country, alongside major Texas AI campuses in Abilene and the Dallas-Fort Worth area. Virginia's dense fiber connectivity and established utility infrastructure continue to attract new hyperscale liquid-cooled capacity, even as local utilities manage rising interconnection queues. Texas has emerged as a parallel hub, supported by competitive power costs, available land, and large-scale AI campus commitments from major cloud and AI infrastructure developers. Georgia and the Carolinas are adding secondary capacity as operators diversify away from the most congested Virginia submarkets. Cooling vendors including Vertiv, Schneider Electric, and CoolIT Systems maintain concentrated field-service and manufacturing support across the region to serve this cluster of hyperscale and colocation demand.

 

The West is projected to grow at the fastest CAGR during the forecast period, driven by new hyperscale AI campus construction across Arizona, Nevada, and Oregon, where available land, favorable climate for economization, and expanding power capacity are attracting gigawatt-scale investment. Phoenix and its surrounding suburbs have become a focal point for hyperscale liquid-cooled construction, supported by utility commitments to expand transmission capacity for large AI loads. Oregon's established hyperscale presence, combined with cooler ambient temperatures that support hybrid economized liquid cooling designs, continues to draw incremental capacity. California retains a concentration of liquid cooling technology developers and corporate headquarters, even as new large-scale construction increasingly shifts to lower-cost neighboring states within the region.

 

Regions Covered

                 ·           South (Dominating Region)

                 ·           West (Fastest-Growing Region)

                 ·           Midwest

                 ·           Northeast

Market Share

The United States AI Data Center Liquid Cooling Market is consolidated, with a small group of established thermal-management companies, Vertiv and Schneider Electric, holding significant share alongside a growing set of specialized direct-to-chip and immersion challengers such as ZutaCore, Accelsius, and Submer. Competitive intensity has increased through 2025 and 2026 as strategic acquirers, including Ecolab, Trane Technologies, Carrier Global Corporation, and Flex, purchased specialist liquid cooling companies to add thermal management capability to broader industrial and electronics portfolios. Key success factors include validated compatibility with leading GPU server platforms, US-based manufacturing and field-service capacity, and the ability to deliver complete rack-to-facility cooling packages rather than individual components. Leading companies are prioritizing warm-water and two-phase cooling innovation, domestic production capacity expansion, and strategic partnerships with server OEMs and hyperscale operators to secure long-term supply positions.

 

Key Players

           ·           Schneider Electric SE (France)

           ·           Vertiv Holdings Co. (US)

           ·           Eaton Corporation plc (Ireland)

           ·           ABB Ltd. (Switzerland)

           ·           Honeywell International Inc. (US)

           ·           Johnson Controls International plc (Ireland)

           ·           Trane Technologies plc (Ireland)

           ·           Siemens AG (Germany)

           ·           Flex Ltd. (Singapore)

           ·           Ecolab Inc. (US)

           ·           nVent Electric plc (United Kingdom)

           ·           Delta Electronics, Inc. (Taiwan)

           ·           Legrand SA (France)

           ·           Rittal GmbH & Co. KG (Germany)

           ·           Carrier Global Corporation (US)

           ·           Sunbird Software, Inc. (US)

           ·           ZutaCore, Inc. (US)

           ·           Munters Group AB (Sweden)

           ·           Submer Technologies, S.L. (Spain)

 

Recent Market Developments

  • In July 2025, Vertiv Holdings partnered with Oklo to co-develop integrated power and liquid cooling solutions for US hyperscale AI data centers, targeting grid-constrained regions where interconnection capacity has become a bottleneck to new facility construction.
  • In October 2025, Johnson Controls made a multi-million-dollar strategic investment in Accelsius, an Austin, Texas-based provider of two-phase, direct-to-chip liquid cooling technology, to accelerate deployment of its NeuCool platform across US AI data centers.
  • In October 2025, Submer was selected by CleanSpark as its first strategic liquid-cooling partner for AI data center expansion across North America, leveraging CleanSpark's three gigawatts of US power and land assets under development.
  • In November 2025, Vertiv and Caterpillar announced a collaboration integrating backup power and liquid cooling infrastructure to address power and cooling constraints affecting AI data center buildouts across North American hubs.

Frequently Asked Questions

What is the United States AI Data Center Liquid Cooling Market?

The United States AI Data Center Liquid Cooling Market covers direct-to-chip, immersion, and rear door liquid cooling equipment and services used to remove heat from GPU-dense AI infrastructure across hyperscale, colocation, enterprise, and government data centers in the United States.

What is driving the United States AI Data Center Liquid Cooling Market growth?
What is the size of the United States AI Data Center Liquid Cooling Market?
Which region dominates the United States AI Data Center Liquid Cooling Market?
Which cooling technology is growing the fastest?
Who are the main end users of AI data center liquid cooling in the United States?
Why is Executive Order 14318 significant for this market?

Key Questions Answered

Request a Sample
1

What is AI data center liquid cooling?

2

What is the CAGR of the United States AI Data Center Liquid Cooling Market?

3

Which cooling technology leads the United States AI Data Center Liquid Cooling Market?

4

Which end user dominates the United States AI Data Center Liquid Cooling Market?

5

Which coolant type has the highest market share?

6

What are the latest trends in the United States AI Data Center Liquid Cooling Market?

7

Who are the leading vendors serving US AI data center liquid cooling deployments?

Why Choose IG Transformation

Speak to Analyst
ico

Strong Industry Focus

ico

Extensive Product Offerings

ico

Customer Research Services

ico

Robust Research Methodology

ico

Comprehensive Reports

ico

Latest Technological Developments

ico

Value Chain Analysis

ico

Potential Market Opportunities

ico

Growth Dynamics

ico

Quality Assurance

ico

Post-sales Support

ico

Regular Report Updates

SINGLE USER ACCESS

$3950

  • PDF Report & Data Sheet
  • Delivered in 24-72 hrs. of purchase
  • 3-Months Analyst Support
  • One designated employee can access the report
bag ico
Buy Now

TEAM USER ACCESS

$4950

  • PDF Report & Data Sheet
  • Delivered in 24-72 hrs. of purchase
  • 3-Months Analyst Support
  • Up to 7 employees or consultants can access
bag ico
Buy Now

ENTERPRISE USER ACCESS

$5950

  • PDF Report & Data Sheet
  • Delivered in 24-72 hrs of purchase
  • 6-Months Analyst Support
  • Any employee, subsidiary, or consultant can access
bag ico
Buy Now

EXCEL SHEET ONLY

$2950

  • Full Excel Data Sheet
  • Delivered in 24-72 hrs of purchase
  • Raw data tables for independent analysis
  • Single-user access
bag ico
Buy Now

Email Subscription Management

By indicating your preferences, you give permission to send you reports, newsletters, invitations to seminars and other relevant marketing materials by email within your preferences.

Enquire Now

Empowering your business decisions through expert market research and seamless IT solutions.