Former OpenAI and DeepMind researchers raise $300M to automate science with Periodic Labs
- October 20, 2025
- 0
Backed by tech giants like Nvidia and Jeff Bezos, the startup aims to create “AI scientists” to revolutionize material discovery.
Backed by tech giants like Nvidia and Jeff Bezos, the startup aims to create “AI scientists” to revolutionize material discovery.
The race to push artificial intelligence beyond text and image generation has gained a new contender with massive ambitions.
Periodic Labs, co-founded by Ekin Dogus Cubuk (former Google Brain and DeepMind) and Liam Fedus (former VP of Research at OpenAI), emerged from stealth mode announcing a $300 million seed round. The funding is supported by industry heavyweights including Andreessen Horowitz, Nvidia, Jeff Bezos, Eric Schmidt, and Jeff Dean.

Periodic Labs’ mission is to automate scientific discovery by building “AI scientists.” The company envisions autonomous labs where robots conduct physical experiments, collect data, iterate, and improve under the supervision of advanced algorithms.
Its first major goal is to develop new superconductors that could outperform current ones while consuming less energy. Such breakthroughs may pave the way for more efficient electronics and energy systems.
Cubuk previously led Google Brain’s materials and chemistry research, including the GNoME AI project that discovered over 2 million new crystals in 2023. Fedus, meanwhile, was instrumental in the creation of ChatGPT and spearheaded the first trillion-parameter neural network.
The team includes other experts who have worked on projects such as OpenAI’s Operator agent and Microsoft’s MatterGen, a large language model designed for material science discovery.
Beyond material invention, Periodic Labs plans to build a vast database with all the experimental data generated by its autonomous labs. This information could fuel its own AI models and become a valuable resource for the broader scientific community.
While other startups and academic groups are exploring similar paths, the combination of financial backing and top-tier expertise positions Periodic Labs as one of the most ambitious efforts to bring artificial intelligence into full automation of experimental science.