arXiv:2605.27722v1 Announce Type: new Abstract: Two-phase boiling enables heat transfer rates an order of magnitude higher than single-phase cooling, but it remains difficult to model due to the strong coupling between phase change, turbulence, and transport, as well as extreme sensitivity to fluid properties and thermodynamic conditions. Existing learning-based surrogates are either condition- or fluid-specific, limiting generalization and requiring separate models. We present NUCLEUS, a mixture-of-experts model for pool boiling that replaces collections of specialized surrogates with a singl
The increasing computational demands of AI and other advanced technologies are pushing the limits of current cooling solutions, making more efficient heat transfer methods critical.
Improved liquid cooling models could unlock higher performance in data centers and advanced computing, easing a key bottleneck for AI scaling and energy efficiency.
Current fragmented and specialized cooling models could be replaced by more generalized and efficient unified systems, accelerating R&D and deployment of advanced cooling.
- · AI compute infrastructure providers
- · Data center operators
- · High-performance computing companies
- · Advanced cooling technology developers
- · Companies reliant on less efficient traditional cooling methods
More efficient cooling enables denser and more powerful compute clusters, facilitating AI research and deployment.
Reduced energy consumption for cooling could alleviate pressure on energy grids supporting large-scale AI operations.
The ability to run more powerful AI models could accelerate breakthroughs in other scientific and engineering fields.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG