Liquid Cooling for SSDs

Overview

Liquid cooling for SSDs is an emerging thermal management approach made necessary by the power trajectory of PCIe interface generations: Gen 5 at ~25W already strains air cooling limits for common SSD form factors, and Gen 7 at 60-70W will make air cooling infeasible. Unlike CPUs and GPUs where liquid cooling is now standard, SSDs present three unique engineering challenges that required dedicated solutions.

The three challenges are: (1) hot-swap serviceability — cold plates must make and break thermal contact with every drive insertion and removal without stress or damage; (2) dual-sided PCB heat sources — modern SSDs have SOC, NAND, DRAM, and other heat-generating components on both sides of the PCB, requiring a cold plate design that can cool both sides; (3) existing SSD specifications written for air-cooled environments with no guidance on thermal interface flatness tolerances, cold plate attachment surfaces, or mechanical insertion geometry.

Air cooling fails for SSDs above approximately 35-40W TDP because SSD form factors (E1.S, E3.S) have no active thermal solution — no internal fans or pumps — and depend entirely on system-level airflow. Liquid cooling provides approximately 100x better cooling energy efficiency: 1000x lower fluid flow rate for equivalent cooling, with only 10x higher impedance, yielding a net 100x energy advantage. As of October 2025, Solidigm has addressed all three challenges and submitted a standardization draft through SNIA to bring the broader industry to deployable liquid-cooled SSD infrastructure.