Periodic Labs company profile: valuation, funding & revenue

This Periodic Labs company profile covers the company's valuation, revenue, funding rounds, investors and history. Periodic Labs works in artificial intelligence, with a focus on AI for science and materials discovery. It was founded in 2025 and is headquartered in Menlo Park, California, United States. Every figure links to a public source, and anything the company hasn't disclosed is marked as such.

Revenue

Funding history

Periodic Labs timeline

Key dates from the sources on this page, oldest first. Metric snapshots are listed after the funding events.

• 2025 — Periodic Labs is founded in Menlo Park by Liam Fedus and Ekin Dogus Cubuk.
• Sep 2025 — Seed: About US$300M, valuing the company at Not disclosed (about US$1B reported in one account). Investors: Andreessen Horowitz, DST Global, NVentures (Nvidia), Accel, Felicis, Jeff Bezos, Eric Schmidt, Elad Gil, Jeff Dean.

Investors

Overview

Periodic Labs is building what it calls AI scientists: systems that combine large AI models with automated physical laboratories so they can propose experiments, run them with robots, learn from the results and propose better experiments, in a continuous loop. Its first target is materials science, with an early focus on finding new superconductors that work at higher temperatures, along with other materials for semiconductors, energy and electronics. The company launched in September 2025 with about US$300 million in seed funding, one of the largest seed rounds ever, from a group that included Andreessen Horowitz, Nvidia's venture arm, DST Global, Accel, Felicis and prominent individual investors such as Jeff Bezos and Eric Schmidt.

Liam Fedus and Ekin Dogus Cubuk co-founded Periodic Labs. Fedus was vice president of research at OpenAI and part of the team behind ChatGPT. Cubuk led a materials and chemistry research team at Google Brain and DeepMind, where he worked on GNoME, an AI system that predicted the structures of hundreds of thousands of potentially stable new materials. The founding team recruited researchers from OpenAI, Google DeepMind, Meta and leading universities, combining expertise in large language models with physics, chemistry and laboratory automation.

What it's building

• AI models for science: models trained to reason about physics and chemistry and to plan experiments.
• Autonomous laboratories: robotic labs that can synthesise and test materials with minimal human intervention, producing high-quality experimental data at scale.
• Experimental data as training data: results from its labs feed back into its models, creating data that doesn't exist on the internet.
• Industrial partnerships: the company has said it intends to work with companies in semiconductors, manufacturing and other industries that need better materials.

Funding

Periodic Labs announced about US$300 million in seed funding in September 2025. Most reports describe the round as led by Andreessen Horowitz, while at least one credits Felicis as a lead; investors named across reports include DST Global, NVentures, Accel, Felicis, Jeff Bezos, Eric Schmidt, Elad Gil and Jeff Dean. One account reported a valuation of about US$1 billion, but we found only a single source for that figure, so we treat the valuation as not disclosed and exclude Periodic Labs from our valuation-ranked tables.

Business model

Periodic Labs is at the research stage and hasn't disclosed revenue. Possible paths include partnerships with industrial companies that pay for materials discovery, licensing discovered materials or processes, and selling access to its AI and lab capabilities. Because a new material can take years to move from discovery to commercial production, the company's long-term value depends on producing discoveries that industry can actually use.

Competition

Periodic Labs competes with research groups at Google DeepMind, Microsoft and Meta that apply AI to materials and chemistry; with AI-for-science startups such as Lila Sciences, CuspAI and Orbital Materials; and, in drug discovery, with companies such as Isomorphic Labs. National laboratories and universities also run autonomous lab programmes. Its distinctive features are its founders' track record in frontier AI, its large early funding and its emphasis on generating its own experimental data.

Periodic Labs at a glance

Periodic Labs is a Silicon Valley AI-for-science company founded in 2025 by former OpenAI research VP Liam Fedus and former Google DeepMind materials researcher Ekin Dogus Cubuk. It combines AI models with robotic labs to discover new materials, starting with superconductors. It launched with about US$300 million in seed funding from Andreessen Horowitz, Nvidia, DST, Accel, Felicis, Jeff Bezos, Eric Schmidt and others.

Timeline

• Early 2025: Liam Fedus leaves OpenAI; the founders begin building the company.
• September 2025: launches publicly with about US$300 million in seed funding.
• 2025–2026: builds out its laboratory and research team.

Why AI needs real experiments

Large language models learn from text and code on the internet, but scientific discovery requires testing ideas against reality. Many promising materials predicted by computers turn out to be unstable or impossible to make, and published data are often inconsistent or incomplete. Periodic Labs' founders have argued that AI progress in science will come from closing the loop: letting AI design experiments, running them precisely and quickly with robots, and learning from both successes and failures. Autonomous labs can run many more experiments than human teams and record every detail, producing clean data for training better models.

Why superconductors

Superconductors carry electricity without resistance, but known materials only do so at very low temperatures or high pressures, which makes them costly to use. A material that superconducts at much higher temperatures could transform power grids, motors, medical imaging and computing. The search is difficult, with many false claims over the years, which makes it a demanding test of whether AI-driven discovery works. Even without a breakthrough superconductor, the methods could find useful materials for chips, batteries and other applications.

Who works with Periodic Labs

As an early-stage research company, Periodic Labs' main audience is industrial partners in semiconductors, energy, aerospace and advanced manufacturing that need better materials, along with the research community.

Risks

• Scientific uncertainty: discoveries may take far longer than investors hope, or not arrive at all.
• Capital intensity: laboratories, robots and compute are expensive.
• Commercialisation: turning a new material into a product requires manufacturing and partners.
• Competition for talent: top AI and science researchers are in high demand.

Questions for partners

• Who owns discoveries made through a partnership?
• How are AI predictions validated before results are shared?
• What timeline is realistic for a usable material?

How it compares with other AI labs

Most well-funded AI labs build general-purpose models and compete on chatbots, coding and agents. Periodic Labs instead applies frontier AI methods to a narrow scientific domain and pairs them with physical experiments. That makes its progress harder to measure from outside, because results come as discoveries and papers rather than product launches, and it means its costs include laboratory equipment and materials as well as computing. Investors backing it are betting that scientific discovery will be one of the most valuable uses of AI over the long term, and that the team best placed to combine AI and experiments will capture a large share of that value.

Latest developments (October 2026)

Periodic Labs was reported in 2026 to be raising at a much higher valuation: Bloomberg reported early talks at about US$7 billion in March, and later reports described negotiations for at least US$500 million at about US$7.5 billion, according to Techleap. We found no confirmed close. The company raised a US$300 million seed round in 2025 to build AI scientists and autonomous labs. See our new AI labs guide.

What to watch

Watch for published results, industrial partnerships, a disclosed valuation and new funding. Our guide to the most valuable AI startups covers related AI labs.

Takeaways

Industry may gain a new route to discovering materials faster. Founders can see how frontier AI researchers are moving into science. Investors should recognise that this is a long-term, high-uncertainty bet, and that its valuation hasn't been confirmed.

Sources

Every figure on this page comes from the public sources below. Where sources disagreed, we used the company's own announcement or a primary news report. Private-company data changes; check the date beside each figure.

1. Maginative