The AI startup ecosystem in 2025 is witnessing a transformative wave, and Periodic Labs is leading the charge. Founded by former OpenAI researcher Liam Fedus and ex-Google Brain scientist Ekin Dogus Cubuk, the startup emerged from stealth mode last month, securing an astonishing $300 million seed round. Led by Felicis, the round also drew investments from top-tier venture capitalists and notable angel investors, signaling strong market confidence in AI-driven scientific innovation.
From Idea to Disruption: The Genesis of Periodic Labs
Periodic Labs began with a conversation seven months ago between Fedus, a key contributor to ChatGPT, and Cubuk, renowned for his work in machine learning and material science at Google Brain. Inspired by discussions around how generative AI could reshape scientific discovery, the duo realized the timing was perfect to integrate AI, robotics, and high-precision simulations into a single research platform.
Cubuk explained, “Advances in robotic automation, simulation accuracy, and large language model reasoning capabilities have aligned perfectly. This is the moment to bring AI fully into experimental science.”
The Technology Stack: AI Meets Robotics
Periodic Labs’ innovation relies on the convergence of three technological breakthroughs:
- Automated Robotic Labs – Robotic systems capable of handling powder synthesis and creating new compounds reliably.
- High-Fidelity Simulations – AI-powered simulations that model complex physical systems, enabling rapid hypothesis testing.
- Advanced Language Models – LLMs analyzing experimental data, suggesting corrections, and guiding discovery processes.
This integrated approach allows AI to propose experiments, physical robots to execute them, and LLMs to interpret outcomes—effectively creating a closed-loop AI-driven scientific lab.
Data as the New Gold: Leveraging Failed Experiments
Fedus and Cubuk also recognized the value of “failed” experiments. Each attempt generates data, which is a critical asset for refining AI models. Unlike traditional science, which often prioritizes published successes, AI science rewards exploration and iterative learning, creating a new paradigm for research motivation.
“Integrating real-world experiments into AI processes is the next frontier,” Fedus emphasized.
The VC Frenzy: How Felicis Won the Deal
Fedus’ public announcement of leaving OpenAI set off an unprecedented wave of investor interest. Venture capitalists and angel investors were eager to participate, sending detailed proposals and even “love letters” expressing their enthusiasm.
The first call they took was with Peter Deng, a former OpenAI colleague turned Felicis partner. Deng recalls a defining moment during a walk in San Francisco’s Noe Valley: “Liam told me that to truly do science, we need a fully equipped wet lab for AI. That’s when I knew we had to invest.”
Within weeks, Periodic Labs was officially incorporated, equipped with a team of world-class scientists and AI engineers, including specialists in material science and generative AI platforms.
Mission and Early Focus
Periodic Labs’ initial goal is ambitious: discovering next-generation superconductors. These materials could revolutionize energy efficiency and enable a new era of high-performance technologies. While robots are still being trained for lab automation, the team has already begun experiments and simulations, using AI to accelerate material discovery.
Cubuk highlighted the company’s collaborative culture: “Each team member lectures on their expertise weekly, ensuring everyone understands all components of our AI-driven research ecosystem.”
The Broader Context: AI-Powered Scientific Innovation
Periodic Labs exemplifies the growing trend of AI startups that merge machine learning, robotics, and material science. OpenAI itself is exploring similar initiatives, aiming to build AI platforms that accelerate scientific discovery.
Investors backing Periodic Labs include Andreessen Horowitz, DST, Nvidia’s NVentures, Accel, alongside angel investors like Jeff Bezos, Eric Schmidt, and Jeff Dean. The funding not only validates the vision but also provides a blueprint for how AI startups in 2025 are reshaping traditional research and innovation landscapes.
Read More: CHRO as the New CEO Strategist: Driving Growth Through People Capital
