How Microsoft found a potential new battery material using AI
The Verge
|
Contributed by: Reid Stephan
Summary
AI and cloud computing expedite search for new battery materials. Microsoft and PNNL's collaboration yields promising solid-state electrolyte for safer, less lithium-dependent batteries. Despite long road for alternatives, generative AI accelerates this work. AQE platform identified 32 million potential battery materials, whittled down to 23 high-potential candidates in 80 hours. One produced working battery, combining lithium and sodium, reducing sodium usage by about 70%. Solid-state battery potentially safer than current lithium-ion batteries, but efficiency struggling. Process simultaneously highlights need for energy-efficient computing and improved batteries. AI and high-performance computing seen as accelerators of scientific discovery.
