Microsoft’s advanced cooling design could open the way for smaller, faster and more efficient AI data centers


  • Microsoft Lab tests microfluidic cooling that triples efficiency compared to cold plates
  • Researchers say that the new system reduces the 65% GPU heat increase in tests
  • Microfluidic cooling promises more dense, faster data centers while reducing the energy costs of IA workloads

Microsoft revealed that he was testing a new chip cooling approach called microfluidic, aimed at resolving one of the biggest challenges for artificial intelligence equipment.

As the AI ​​chips become more powerful, they generate more heat than previous generations, pushing traditional cooling methods near their limits.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top