{"id":11799,"date":"2026-06-05T06:39:49","date_gmt":"2026-06-05T06:39:49","guid":{"rendered":"https:\/\/www.modelon.com\/ja\/blog\/why-two-phase-direct-to-chip-cooling-is-reaching-a-tipping-point\/"},"modified":"2026-06-05T06:39:45","modified_gmt":"2026-06-05T06:39:45","slug":"why-two-phase-direct-to-chip-cooling-is-reaching-a-tipping-point","status":"publish","type":"post","link":"https:\/\/www.modelon.com\/ja\/blog\/why-two-phase-direct-to-chip-cooling-is-reaching-a-tipping-point\/","title":{"rendered":"Why Two\u2011Phase Direct\u2011to\u2011Chip Cooling is Reaching a Tipping Point"},"content":{"rendered":"\n

Modelon & University of Maryland Researchers to Present Findings at Upcoming Conference<\/strong><\/h3>\n\n\n\n

Artificial intelligence and high-performance computing (HPC) are fundamentally reshaping the way data centers<\/strong> are designed. Power densities continue to rise, thermal margins are tightening, and workloads are becoming more dynamic and less predictable. HGX H100 racks introduced in 2022 consumed around 40\u201360 kW, while Blackwell racks introduced in 2024 consume around 120\u2013140 kW, roughly three times more than previous AI racks. Next-generation racks expected in 2026 could reach 200\u2013240 kW. For engineering teams, this shift requires a new approach to rack cooling, one that accounts for higher heat loads, faster transients, and system-level behavior earlier in the design process.<\/p>\n\n\n\n

To better understand what comes next, we spoke with Dr. Lingnan Lin, Assistant Professor of Mechanical Engineering<\/strong> at the University of Maryland, College Park, and Kagan Sears, a PhD student <\/strong>in Prof. Lin\u2019s lab. Dr. Lin\u2019s research focuses on phase\u2011change heat and mass transfer, with applications spanning HVAC systems and electronics cooling. Together, their academic\u2013industry collaboration with Modelon explores two\u2011phase direct\u2011to\u2011chip liquid cooling from both a physics and system\u2011level perspective. Their assessment cuts across boiling physics, transient system behavior, and practical implementation offering a clear view into why this technology matters now, what engineers are learning today, and where the field is headed.<\/p>\n\n\n\n

Why two\u2011phase direct\u2011to\u2011chip cooling matters now<\/h3>\n\n\n\n

As Prof. Lin described during the interview, the core driver is simple, but profound: the thermal problem has changed faster than cooling architectures have.<\/p>\n\n\n\n

\u201cIf you look at where AI accelerators and next\u2011generation processors are going, the heat fluxes are no longer something that traditional approaches were designed for,\u201d<\/em> Lin explained.<\/p>\n\n\n\n

Two\u2011phase direct\u2011to\u2011chip cooling leverages boiling directly at the chip interface, absorbing heat through latent heat rather than sensible heat alone. For AI and HPC workloads with sharp power ramps, localized hotspots, and sustained high utilization, this provides a path to higher heat removal in a smaller footprint.<\/p>\n\n\n\n

But neither Lin nor Sears framed two\u2011phase cooling as a silver bullet. Instead, they emphasized that its promise only becomes real when it is treated as part of a system<\/em>, not just a superior cold plate.<\/p>\n\n\n\n

Why transient system simulation is essential\u2014not optional<\/h3>\n\n\n\n

A recurring message throughout the conversation was that steady\u2011state thinking breaks down for two\u2011phase systems.<\/p>\n\n\n\n

\u201cWe\u2019re not just interested in whether boiling occurs,\u201d<\/em> Sears noted. \u201cWe want to understand how the system behaves dynamically\u2014as loads change, as flow conditions vary, and as the loop responds over time.\u201d<\/em><\/p>\n\n\n\n

Boiling initiation, vapor generation, flow instabilities, startup behavior, and control actions all evolve on different timescales. These dynamics directly affect performance, stability, and long\u2011term reliability. As a result, transient system simulation becomes a design requirement rather than a nice\u2011to\u2011have.<\/p>\n\n\n\n

Through Modelon\u2019s system\u2011level simulation capabilities, the team can model<\/p>\n\n\n\n