The 2010s and beyond marked a significant resurgence for direct liquid cooling. The increasing demands of high-performance computing (HPC), artificial intelligence (AI), machine learning (ML), and data centers led to higher power densities and heat loads that traditional air cooling struggled to manage efficiently. Today, direct liquid cooling is primarily implemented in two main forms: Direct-to-chip liquid cooling (DLC): This method involves attaching cold plates directly to heat-generating components like CPUs and GPUs. A liquid coolant circulates through these plates, absorbing heat and carrying it away through a closed-loop system. Immersion cooling: Single-phase immersion cooling submerges IT hardware in dielectric fluid that remains liquid throughout the cooling process, transferring heat through external heat exchangers. In contrast, two-phase immersion cooling uses fluids that boil at low temperatures to create vapor that condenses above the tank, offering higher efficiency, but requiring more complex infrastructure
The Air Cooling Era Metal oxide
The Comeback HPC, AI, and ML drive heat demands Data center power densities soar Air cooling reaches limits Liquid cooling renaissance begins
Two Paths Forward Direct Liquid Cooling (DLC) Cold plates on CPUs/GPUs Closed-loop circulation
semiconductors emerge Air-cooled rack-mounted servers dominate Liquid cooling demand diminishes Cost-effective air solutions rule
Immersion Cooling Complete hardware submersion Dielectric liquid coolant
1990s – 2000s
2010s – Present
Today’s Solutions
Photo Credit: Eabel Cooling
Photo Credit: TechTarget
9
Made with FlippingBook Ebook Creator