Periodic Labs is an AI research company building autonomous AI scientists and self-driving laboratories to accelerate scientific discovery. Their closed-loop systems generate hypotheses, conduct physical experiments, and refine understanding autonomously in materials science, chemistry, and related physical sciences. Initial focus areas include superconductors, semiconductor thermal issues, and materials R&D.
Periodic Labs is an AI research company building autonomous AI scientists and self-driving laboratories to accelerate scientific discovery. Their closed-loop systems generate hypotheses, conduct physical experiments, and refine understanding autonomously in materials science, chemistry, and related physical sciences. Initial focus areas include superconductors, semiconductor thermal issues, and materials R&D.
Founded in 2025 by Liam Fedus (former VP of Research at OpenAI, helped create ChatGPT, and led the first trillion-parameter neural network team) and Ekin Dogus Cubuk (led materials and chemistry team at Google Brain and DeepMind, led the GNoME project that discovered 2M+ new crystals). The company emerged from stealth with a record $300M seed round to build AI scientists capable of running and analyzing their own experiments.
Bloomberg reported that Periodic Labs is in deal talks at approximately $7 billion valuation, a rapid step-up from its seed-stage valuation just months after launching.
Multiple outlets covered Periodic Labs as part of a wave of ex-OpenAI and Google Brain founders launching high-profile AI startups, with Periodic Labs standing out for its record-breaking seed round size.
Periodic Labs emerged from stealth with a $300 million seed round, one of the largest seed rounds ever raised, led by Andreessen Horowitz with participation from Felicis, DST Global, NVentures (NVIDIA), Accel, and angels including Jeff Bezos, Eric Schmidt, and Jeff Dean.
Periodic Labs officially launched its platform for autonomous AI scientists—systems that integrate advanced AI models with self-driving laboratory hardware to automate the full scientific method including hypothesis generation, experimentation, and data analysis.
Periodic Labs revealed initial focus on high-impact physical science problems including room-temperature superconductors, semiconductor thermal management, and novel materials R&D—areas with transformative potential for energy and computing.
Periodic Labs revealed its near-term operational milestone: deploying autonomous powder synthesis labs where materials are turned into powder, heated, and mixed to discover new compounds. Initial focus is on superconductors for next-gen chips and power grid infrastructure.
Periodic Labs publicly launched its AI scientist platform, combining large language models, physics simulations, and robotic labs to autonomously design and test new materials. The company's first target is discovering next-generation room-temperature superconductors.
TechCrunch covered the launch of Periodic Labs, profiling how top AI researchers from OpenAI and Google Brain are building autonomous laboratories with AI scientists that can generate and test hypotheses independently.
a16z published an announcement detailing their thesis for investing in Periodic Labs, highlighting the company's potential to accelerate materials science discovery using AI and autonomous labs.
Liam Fedus posted on LinkedIn announcing Periodic Labs, describing the company's mission to build AI scientists and autonomous laboratories for materials discovery alongside co-founder Ekin Dogus Cubuk.
Co-Founder & CEO
Previously OpenAI
Former VP of Research at OpenAI, where he helped create ChatGPT and led the team that built the first trillion-parameter neural network. Co-founded Periodic Labs to build AI scientists that can autonomously generate and test scientific hypotheses in physical laboratories.
Co-Founder
Previously Google DeepMind
Led the materials and chemistry team at Google Brain and DeepMind, where he created GNoME, an AI tool that discovered over 2 million new crystals in 2023. Co-founded Periodic Labs to apply AI to autonomous physical science experimentation.
Co-Founder & CTO
Previously Research Lead, Materials and Chemistry at Google DeepMind
Former head of Google DeepMind's materials and chemistry team. Co-created GNoME, a model that generated millions of candidate crystal structures. Co-founded Periodic Labs to automate scientific discovery using AI and robotics.
$300M raised total
