WOBURN, Mass. — SES AI has released an updated version of its Molecular Universe platform that includes a new AI-powered research feature designed to drastically shorten the time required for advanced battery development.
The new capability, called Deep Space, is part of Molecular Universe 0.5 (MU-0.5) and leverages a multi-agent large language model to perform what the company describes as senior scientist-level battery research. According to SES, Deep Space can deliver insights and recommendations within 10 to 20 minutes—research that might otherwise take human experts years to complete.
Deep Space builds on the platform’s existing “Ask” feature, which provides quick, AI-generated answers to battery-related queries. While Ask returns results in under a minute, Deep Space is intended for more intensive queries, such as designing commercial battery formulations based on performance, cost, or novelty criteria.
SES CEO Qichao Hu said the platform has already seen strong adoption among battery manufacturers, automotive companies, and research institutions. “The excitement the Molecular Universe has generated among customers is due to its speed in delivering useful research insights and reducing the costs of traditional trial-and-error methods,” Hu said.
The company sees Deep Space as particularly relevant for battery makers, as well as producers of cathode and electrolyte materials and automakers pursuing next-generation battery designs. The platform is positioned as a tool to support faster go-to-market strategies by cutting down on lab testing, hiring, and patent filings.
Additional updates in MU-0.5 include a new molecule sketch pad for finding structurally or functionally similar compounds, and expanded language support in Mandarin and Korean alongside English.
SES AI develops and manufactures lithium-metal and lithium-ion batteries. The launch of Deep Space reinforces its strategy to offer R&D as a service, combining AI tools with materials science expertise to remain competitive in the battery technology space.
