Periodic Labs

We had a great conversation with The New York Times about what we are building at Periodic Labs. Thanks to Cade Metz for the story.

We’re thrilled to announce we are leading a $300M founding round for Periodic Labs. Right now, science is bottlenecked by the slow and manual process of experimentation and verification. To move faster, we need to combine frontier-level artificial intelligence with feedback from the real world. Periodic Labs does just that. Periodic builds both state-of-the-art AI scientists and the autonomous labs in which they can form hypotheses, run physical experiments, and iterate at a speed and scale that we believe no human-led lab can match. The world-class team behind Periodic Labs has helped create ChatGPT, DeepMind’s GNoME, OpenAI’s Operator, neural network attention, the Materials Project, and MatterGen, and has already partnered with a semiconductor manufacturer to make sense of their experimental data in order to iterate and innovate faster. Their first goal: designing better semiconductors and paving the way for breakthroughs in computing, transportation, and energy. We’re excited to partner with Periodic as they set out to transform scientific discovery. Congratulations to William (Liam) Fedus, Ekin Dogus Cubuk, and the entire team!

I am excited to announce what William (Liam) Fedus and I have been working on: Periodic Labs, a world class team of experimentalists, theorists, and LLM experts. Scientific discovery is inherently an out-of-domain task. Experimental iteration is required for significant advances, regardless of the form of intelligence that is modeling the world. We are building experimental labs that will unlock the next frontier for LLM reasoning. Deeply grateful to our advisory board, Prof. Carolyn Bertozzi, Prof. Mercouri Kanatzidis, Prof. Steven Kivelson, Prof. Zhi-Xun Shen, and Prof. Chris Wolverton, for their guidance and support.

We are proud to announce Periodic Labs. Our mission is to accelerate science. Our founding team co-created ChatGPT, DeepMind’s GNoME, OpenAI’s Operator (now Agent), the neural attention mechanism, MatterGen; have scaled autonomous physics labs; and have contributed to important materials discoveries of the last decade. We’re fortunate to be backed by investors who share our vision, including Andreessen Horowitz who led our $300M round, as well as Felicis, DST Global, NVIDIA (NVIDIA’s venture capital arm), Accel and individuals including Elad Gil, Eric Schmidt, Jeff Dean, and Jeff Bezos. We’ve come together to scale up and reimagine how science is done. If you want to help build the first generation of AI scientists, we’re hiring: https://lnkd.in/gvZv56Q7.

Today, Ekin Dogus Cubuk and I are excited to introduce Periodic Labs. Our goal is to create an AI scientist. Science works by conjecturing how the world might be, running experiments, and learning from the results. Intelligence is necessary, but not sufficient. New knowledge is created when ideas are found to be consistent with reality. And so, at Periodic, we are building AI scientists and the autonomous laboratories for them to control. Until now, scientific AI advances have come from models trained on the internet. But despite its vastness — it’s still finite (estimates are ~10T text tokens where one English word may be 1-2 tokens). And in recent years the best frontier AI models have fully exhausted it. Researchers seek better use of this data, but as any scientist knows: though re-reading a textbook may give new insights, they eventually need to try their idea to see if it holds. Autonomous labs are central to our strategy. They provide huge amounts of high-quality data (each experiment can produce GBs) that exists nowhere else. They generate valuable negative results seldom published. But most importantly, they give our AI scientists the tools to act. We’re starting in the physical sciences. Progress is limited by our ability to design the physical world. We’re starting here because experiments have high signal-to-noise and are fast, physical simulations effectively model many systems, but more broadly, physics is a verifiable environment. AI has progressed fastest in domains with data and verifiable results — for example, in math and code. Here, nature is the RL environment. One of our goals is to discover superconductors that work at higher temperatures than today's materials. Significant advances could help us create next-generation transportation and build power grids with minimal losses. But this is just one example — if we can automate materials design, we have the potential to accelerate Moore’s Law, space travel, and nuclear fusion. We’re also working to deploy our solutions with industry. As an example, we're helping a semiconductor manufacturer that is facing issues with heat dissipation on their chips. We’re training custom agents for their engineers and researchers to make sense of their experimental data in order to iterate faster. Our founding team co-created ChatGPT, DeepMind’s GNoME, OpenAI’s Operator (now Agent), the neural attention mechanism, MatterGen; have scaled autonomous physics labs; and have contributed to some of the most important materials discoveries of the last decade. We’ve come together to scale up and reimagine how science is done. We’re backed by investors who share our vision, including Andreessen Horowitz who led our $300M round, as well as Felicis, DST Global, NVentures (NVIDIA’s VC arm), Accel and individuals including Jeff Bezos, Elad Gil, Eric Schmidt, and Jeff Dean. Their support will help us grow our team, scale our labs, and develop the first generation of AI scientists.