Thermonat: A machine learning tool for semiconductor modeling @ibmresearch
Thermonat: A machine learning tool for semiconductor modeling  @ibmresearch
Uploaded January 2026 | Updated November 2026, 1 day ago
When chips shrink, heat becomes the limiting factor

IBM Research shared new results from DARPA’s Thermonat project, introducing atom-accurate thermal modeling for next-generation semiconductors. With 0.002% prediction accuracy and speeds 50,000× faster than current tools, Thermonat helps designers better account for heat as chips keep shrinking.

Dive into the blog for more → ibm.co/60438LiXL

#thermonat #machinelearning #ai #semiconductors
Thermonat: A machine learning tool for semiconductor modelingWhat is AI governance?The Short: IBM Quantum goes to college, Intro to AI inferencing, Modernizing code with watsonxHow can AI help modernize code?Using AI models to fight climate change in KenyaHow quantum computers are being used for drug discoveryWhy do AI models need to be safe?Integrating quantum and high-performance computing: Co-design platform for quantum error correctionIBM Quantum Webinar Series: Quantum-Centric Supercomputing Algorithms and Use CasesWhats next in AI automation for enterprise?IBM Granite tops the SQL charts, faster chatbots with speculative decoding, & IBM AI at WimbledonSimulating Quarks on Quantum Computers
IBM Research |

Thermonat: A machine learning tool for semiconductor modeling

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER