What Periodic Labs does
Periodic Labs is building AI scientists paired with autonomous laboratories to accelerate scientific discovery, starting with the physical sciences. The company automates the scientific method itself, using AI agents to generate hypotheses, run experiments in automated labs, and learn from the resulting data.
Key capabilities
Periodic Labs develops AI agents trained on experimental data produced by its own autonomous labs, which generate large volumes of physical-science measurement data per experiment. It trains custom agents for industrial partners to analyze experimental data and optimize materials design and discovery. The company emphasizes physical experiments because they offer high signal-to-noise, relatively fast results, and a verifiable environment, similar to domains like mathematics and code where correctness is measurable.
Who it's for
Periodic Labs targets industry partners in semiconductor manufacturing and materials science, as well as academic researchers through a grant program. Use cases include discovering superconductors with higher operating temperatures, designing advanced materials, and helping semiconductor makers solve engineering challenges such as chip heat dissipation, with longer-term ambitions across frontier research areas.