SES AI Accelerates Battery Development for Robotics With AI-Powered Materials Discovery
SES AI, a Boston-based battery maker, is betting that artificial intelligence will become the primary competitive advantage in developing batteries for humanoid robots and drones. The company's Molecular Universe platform uses what CEO Qichao Hu calls "vibe research" - allowing researchers to describe what they need in plain language and letting AI agents autonomously conduct experiments to find solutions.
The shift matters because the robotics sector demands batteries with different characteristics than electric vehicles. Humanoid robots need cells that fit existing form factors while delivering higher capacity and longer operating times. SES AI's 21700-type cylindrical cells have already jumped from 5 ampere-hour capacity to 7.2 Ah, but the company sees room for further gains.
From Years to Weeks
Traditionally, battery makers spend 3 to 5 percent of annual revenue on research and development, employing hundreds of scientists to test materials through trial and error. A typical project - say, extending battery life at minus 20 degrees Celsius from 6,000 to 8,000 charge cycles - takes years and extensive experimentation.
Molecular Universe compresses that timeline. Researchers enter a natural language prompt. AI agents run simulations and experiments through connected autonomous laboratories, iterating continuously until they identify a working solution. Hu said the process can reduce development cycles "from years to weeks" while cutting research costs.
The platform's latest version, 3.0, includes an AI agent called StarSeeker that orchestrates multiple capabilities. A user can request: "Find the top 20 molecules for low-temperature sodium battery performance, generate 10 formulations for each, and rank them by performance." StarSeeker executes the full workflow and returns ranked results.
"This is a huge cost savings and time acceleration for product development," Hu said. "Instead of having a big team, you just have one product manager and this platform."
A Niche Strategy
SES AI abandoned direct competition with major EV battery manufacturers, which compete primarily on production scale and cost. Instead, the company focuses on three areas: energy storage for data centers, batteries for drones, and materials discovery software.
Its Chungju plant in South Korea originally produced military batteries as a joint venture with GM Defense. After GM exited the project, SES AI took full ownership and converted the facility to drone battery production. The plant plans to scale from 330,000 cells annually to 1 million.
The company generated $21 million in revenue last year, up roughly tenfold year-over-year, driven primarily by energy storage sales. Operating losses narrowed to $82.61 million from $109.24 million.
Why This Matters Now
Robotics remains early-stage, with no government subsidies or compliance requirements yet in place. That creates an opening: whoever develops the best battery at the lowest cost will supply the market. Tesla CEO Elon Musk has suggested more humanoid robots could be built than cars or people, signaling massive potential demand.
Hu said similar AI-driven research platforms have already become profitable in pharmaceuticals over the past several years. The battery industry is likely to follow. Some of the largest battery and automotive companies are already using Molecular Universe.
SES AI currently treats Molecular Universe as a cost center requiring long-term investment. Hu said the company is considering spinning out the platform as a separate entity, which could attract dedicated funding and potentially pursue an independent public offering.
For product development teams, the implication is clear: the ability to rapidly discover and validate new materials will increasingly determine which companies can meet robotics demand. Those without access to accelerated research tools face a widening disadvantage.
Your membership also unlocks:
