Data Center Liquid Cooling Market Projected to Reach $38.4B by 2033 at 28.7% CAGR

The data center liquid cooling market is entering a period of rapid expansion according to a report published February 17, 2026 by GlobeNewswire. The market is valued at approximately $6.6 billion in 2026 and is projected to reach $38.4 billion by 2033, representing a compound annual growth rate of 28.7%. The primary driver is straightforward: the GPU clusters powering AI training and inference workloads are generating heat that conventional air cooling cannot remove cost-effectively at scale.

AI training clusters now routinely exceed 30 to 50 kW per rack, far beyond the 8 to 15 kW range that traditional raised-floor air cooling was designed for. Next-generation GPU platforms push the envelope further—NVIDIA's Vera Rubin NVL144 architecture operates above 300 kW per rack in dense configurations. At those power densities, air cooling requires either massive over-provisioning of CRAC units and hot-aisle containment or simply becomes impractical. Liquid cooling solves the thermal problem at the source by moving heat into water or refrigerant that can be transported and rejected far more efficiently than air.

Technology Breakdown

The report segments the market by cooling technique. Direct-to-chip cooling—cold plates bonded to CPU and GPU heat spreaders, carrying chilled water through the server—currently leads adoption due to its compatibility with existing data center infrastructure. Customers can retrofit direct-to-chip solutions into air-cooled facilities without rebuilding the building envelope, which lowers the barrier to adoption. Immersion cooling, where servers are submerged in dielectric fluid, is identified as the fastest-growing segment, particularly for hyperscale AI deployments where entire racks or pods are purpose-built for the thermal solution. Rear-door heat exchangers serve as a transitional technology for operators upgrading existing air-cooled rows without committing to full liquid infrastructure. Two-phase immersion is emerging in specialized high-performance computing environments.

North America commands the largest share of the current market, hosting the world's highest concentration of AI training clusters. The report notes that hyperscalers including Google, Meta, Microsoft, and Amazon are collectively planning hundreds of billions of dollars in AI infrastructure investment, with the aggregate figure cited at $650 billion, creating sustained demand for liquid cooling vendors across the hardware stack.

Sustainability as a Co-Driver

Energy efficiency and carbon reduction mandates are reinforcing the AI-driven demand. Liquid-cooled systems typically operate at a Power Usage Effectiveness of 1.03 to 1.1, compared to 1.4 to 1.6 for air-cooled equivalents, because far less energy is spent moving air through the building. Regulatory pressure in the European Union and increasingly in North American jurisdictions is pushing operators toward measurable PUE improvements. The combination of thermal necessity from AI workloads and regulatory pressure on efficiency makes liquid cooling investment a near-certainty for new large-scale builds, regardless of whether operators would have chosen it purely on economics.