Ex-OpenAI and Google DeepMind researchers raise $300 million to automate science

Periodic Labs, a new startup founded by AI and materials science veterans, emerged from stealth Tuesday with a massive $300 million seed round. The funding comes from a who’s who of tech investors, including Andreessen Horowitz, DST, Nvidia, Accel, Elad Gil, Jeff Dean, Eric Schmidt, and Jeff Bezos.

Founders With Deep AI Pedigrees

The company was founded by Ekin Dogus Cubuk and Liam Fedus, both of whom bring heavyweight experience from the front lines of AI research:

• Cubuk led the materials and chemistry team at Google Brain and DeepMind, where he helped develop GNoME, an AI system that discovered over 2 million new crystals in 2023.
• Fedus, a former VP of Research at OpenAI, helped create ChatGPT and led the team behind the first trillion-parameter neural network.

Periodic’s small team also includes researchers who worked on major projects like OpenAI’s Operator agent and Microsoft’s MatterGen, an LLM for materials discovery.

Automating Scientific Discovery

Periodic Labs’ mission is ambitious: to automate scientific discovery by building “AI scientists” that can run autonomous laboratories. In these labs, robots would conduct experiments, collect data, iterate, and improve—a continuous loop of machine-led research.

The startup’s first target is the development of new superconductors, which could outperform existing materials and potentially require less energy. Beyond superconductors, the team aims to invent other breakthrough materials.

Another major goal is data generation. Unlike large language models trained on the internet—an increasingly “exhausted” data source—Periodic wants to create fresh, high-quality datasets from physical-world experiments, fueling the next wave of scientific AI.

A Growing Field

While Periodic Labs has assembled one of the most formidable teams in this emerging space, it’s not alone. Startups like Tetsuwan Scientific, nonprofits such as Future House, and academic initiatives like the University of Toronto’s Acceleration Consortium are also exploring ways to use AI to automate chemistry and materials discovery.

But with $300 million in seed funding and backing from some of the tech industry’s most influential leaders, Periodic Labs is immediately positioned as one of the field’s most well-funded and ambitious players.