A new startup, Periodic Labs, has just emerged from stealth mode with a massive $300 million in seed funding, and it’s already making waves. Backed by some of the most influential names in tech—like Andreessen Horowitz, Nvidia, Jeff Bezos, and Eric Schmidt—this company has big ambitions: to automate scientific discovery using AI.
Founded by Ekin Dogus Cubuk and Liam Fedus, the startup is built on a foundation of deep expertise. Cubuk comes from Google Brain and DeepMind, where he helped develop an AI called GNoME, which discovered over 2 million new crystal structures in 2023. These materials could one day revolutionize how we build technology. Fedus, on the other hand, is a former OpenAI VP who contributed to the development of ChatGPT and led the creation of the first trillion-parameter neural network.
The team at Periodic Labs isn’t your average startup crew—it includes researchers who have worked on leading-edge AI projects like OpenAI’s Operator and Microsoft’s MatterGen, which focus on discovering new materials through large language models.
So what exactly is Periodic Labs trying to do? Their goal is bold: to create AI scientists. That means building autonomous labs where robots run experiments, collect data, and improve over time—essentially, machines that learn through hands-on trial and error, just like human scientists. Their first big target? Inventing next-generation superconductors that are more efficient and energy-friendly than current options.
But it doesn’t stop there. As these robot labs run tests—mixing, heating, and manipulating materials—they’ll generate new physical data. This data, which doesn’t exist anywhere on the internet, will help AI systems become even smarter by giving them something fresh to learn from.
In a blog post, the company points out that current AI models have pretty much “exhausted” the internet for training data. So, instead of relying on existing information, Periodic Labs wants to create new data from the real world—something that could push AI to a whole new level.
While Periodic Labs may have one of the most star-studded teams and generous funding, it’s not alone. Other players, like Tetsuwan Scientific, Future House, and the University of Toronto’s Acceleration Consortium, are also exploring how AI can accelerate breakthroughs in chemistry and material science. But with its dream team and cutting-edge mission, Periodic Labs is definitely one to watch.
About the Author
