Periodic Labs Emerges from Stealth With $300m Funding, Aiming to Build AI Scientists for Materials Discovery

Periodic Labs, a startup founded by some of the most influential minds in artificial intelligence, launched out of stealth on Tuesday with a massive $300 million seed round.

The funding, unusually large for a company at this stage, reflects the ambition of its mission: to automate scientific discovery through AI-driven laboratories.

The round was backed by an elite roster of investors from both the tech and scientific worlds, including Andreessen Horowitz, DST, Nvidia, Accel, Elad Gil, Jeff Dean, Eric Schmidt, and Amazon founder Jeff Bezos.

From AI Breakthroughs to Scientific Frontiers

Periodic Labs was founded by Ekin Dogus Cubuk and Liam Fedus. Cubuk previously led the materials and chemistry team at Google Brain and DeepMind, where he helped develop GNoME, an AI tool that discovered more than 2 million new crystals in 2023—materials researchers say could one day power entirely new generations of technology.

Fedus, meanwhile, is a former VP of Research at OpenAI and one of the creators of ChatGPT. He also led the team behind the first trillion-parameter neural network. The company’s small team is composed of scientists and engineers who have worked on major AI and materials science initiatives, including OpenAI’s agent Operator and Microsoft’s MatterGen, a large language model for materials discovery.

Automating Discovery

Periodic Labs describes its mission as building “AI scientists.” In practice, this means creating labs where robots conduct experiments, collect data, and learn from each cycle, iterating in ways that mimic and potentially accelerate the scientific method.

The startup’s first focus is superconductors—materials capable of conducting electricity with little to no resistance. Today’s superconductors are expensive, difficult to produce, and require extreme conditions like supercooling. If Periodic Labs can invent new materials that perform better under easier conditions, it could reshape industries from power transmission to computing.

Another priority is building a vast database of physical-world data generated by its AI scientists. “Until now, scientific AI advances have come from models trained on the internet,” the company said in a blog post. But with the internet as a data source now “exhausted,” Periodic Labs argues that the next breakthroughs will depend on fresh, real-world experimental data.

A Field Heating Up

While Periodic Labs is one of the most heavily funded entrants in this space, it is not alone. Automating chemistry and materials discovery with AI has been a growing focus of academic research since at least 2023. Startups such as Tetsuwan Scientific, nonprofits like Future House, and institutions like the University of Toronto’s Acceleration Consortium are also exploring how machine learning can accelerate breakthroughs in materials science.

But Periodic’s $300 million seed round sets it apart. It represents one of the largest initial financings for a science-focused AI company and suggests that Silicon Valley’s top investors believe the combination of AI and lab automation could lead to transformative discoveries.

Beyond AI Chatbots

The launch highlights a broader shift in the AI sector. For years, breakthroughs like ChatGPT and generative image models have defined public understanding of AI. But with models trained largely on the internet reaching diminishing returns, investors are now backing ventures that can create their own experimental data.

Companies like Periodic Labs hope to do for physical science what ChatGPT did for language: compress decades of trial-and-error into accelerated cycles of learning. Compared with efforts in pharmaceuticals, where AI is already being used to predict drug candidates, materials discovery has lagged. Periodic’s deep pockets and star-studded founding team suggest that may soon change.

The opportunity is believed to be about scale: creating not just one AI scientist, but many, each running thousands of experiments to invent materials that today exist only in theory.