{"id":229209,"date":"2025-10-21T05:31:08","date_gmt":"2025-10-21T05:31:08","guid":{"rendered":"https:\/\/www.newsbeep.com\/au\/229209\/"},"modified":"2025-10-21T05:31:08","modified_gmt":"2025-10-21T05:31:08","slug":"top-openai-google-brain-researchers-set-off-a-300m-vc-frenzy-for-their-startup-periodic-labs","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/au\/229209\/","title":{"rendered":"Top OpenAI, Google Brain researchers set off a $300M VC frenzy for their startup Periodic Labs\u00a0"},"content":{"rendered":"

Periodic Labs, a new startup by one of OpenAI\u2019s most respected researchers, Liam Fedus, and his former Google Brain colleague, Ekin Dogus Cubuk, came out of stealth last month with an enormous $300 million seed round<\/a>. It was led by Felicis and included a who\u2019s who of angels and other top VCs.\u00a0<\/p>\n

The startup began when Fedus had a conversation with\u00a0Cubuk (whose friends call him \u201cDoge\u201d) about seven months ago. Cubuk was one of Google Brain\u2019s foremost machine learning and material science researchers.\u00a0After endless Silicon Valley takes on how generative AI would radically change scientific discovery<\/a>, they decided that the pieces were finally in place to make this a reality. Or at least to found a startup that attempted it.<\/p>\n

\u201cThere are a few things that happened in the LLM field, in experimental science and in simulations that kind of made this the right time,\u201d Cubuk told TechCrunch. <\/p>\n

For one, he said, robotic arms that could handle powder synthesis \u2014 the process of mixing and creating new materials \u2014 had recently proved themselves reliable.\u00a0For another, machine learning simulations had become efficient and accurate enough to model complex physical systems such as those needed to develop new materials.\u00a0<\/p>\n

And, third, LLMs now had powerful reasoning capabilities\u00a0\u2014 in part through the work of Fedus and his team at OpenAI.\u00a0Fedus was one of the small team that created ChatGPT to begin with and was running OpenAI\u2019s uber-important post-training team<\/a>, which refines models after their initial development.<\/p>\n

Stitched together, the picture was clear: A simulation could theoretically discover new compounds, a robot could mix the materials, and an LLM could analyze the results and suggest course corrections. AI-automated material science was ready to be built.<\/p>\n

In fact, Cubuk was one of the researchers who published a groundbreaking paper in 2023<\/a> documenting a precursor Google research project. The team built a fully automated, robotic-powered lab and created 41 novel compounds from recipes suggested by language models.\u00a0<\/p>\n

Techcrunch event<\/p>\n

\n\t\t\t\t\t\t\t\t\tSan Francisco
\n\t\t\t\t\t\t\t\t\t\t\t\t\t|
\n\t\t\t\t\t\t\t\t\t\t\t\t\tOctober 27-29, 2025\n\t\t\t\t\t\t\t<\/p>\n

Equally importantly, the founders realized that even failed experiments would be valuable for their new startup because data is the lifeblood of AI. AI science offered an entirely new source for real-world training and post-training data. This\u00a0could, the founders believe, turn the existing scientific motivation system on its head, which seeks success, not exploration, rewarded via paper publication and grants.<\/p>\n

\u201cMaking contact with reality, bringing experiments into the [AI] loop \u2014 we feel like this is the next frontier,\u201d Fedus told TechCrunch.\u00a0<\/p>\n

Felicis wins the deal; OpenAI does not invest\u00a0<\/p>\n

After that discussion with Cubuk, Fedus went to the powers that be at OpenAI to share his resignation and his plan. He then gleefully tweeted<\/a> to the world that he was leaving with what appeared to be OpenAI\u2019s blessing and investment.\u00a0\u00a0<\/p>\n

This is what I sent to my colleagues at OpenAI:<\/p>\n

Hi all, I made the difficult decision to leave OpenAI as an employee, but I\u2019m looking to work closely together as a partner going forward. Contributing to the mission of OpenAI and working with world-class teams to create and\u2026<\/p>\n

\u2014 William Fedus (@LiamFedus) March 17, 2025<\/a><\/p>\n

That investment didn\u2019t actually materialize, however. OpenAI is not a backer of Periodic, the founders confirmed to TechCrunch. And while Fedus declined to say why, they actually didn\u2019t need OpenAI\u2019s money.\u00a0\u00a0<\/p>\n

Fedus\u2019 tweet set off a frenzy of VCs courting the company. \u201cThere was almost like a feeling of being reverse pitched. One investor actually wrote a love letter to Periodic Labs,\u201d Fedus laughed, explaining that neither he nor Cubuk \u201cknew what to make of it.\u201d Others sent multi-page documents pitching themselves.<\/p>\n

But the first call that they actually took was from Peter Deng, a former OpenAI colleague turned investor for top-tier seed firm Felicis. (Deng left OpenAI for Felicis at the start of 2025.)\u00a0\u00a0<\/p>\n

\u201cLiam is a very big deal within OpenAI, very well loved and an extremely impactful researcher,\u201d Deng told TechCrunch. \u201cWhen I heard he left, I texted him immediately.\u201d\u00a0<\/p>\n

Deng met Fedus for coffee in the Noe Valley neighborhood of San Francisco. Hyped on caffeine and enthusiasm, Fedus invited Deng to finish their conversation on a walk over the area\u2019s famously hilly terrain. Pitch walks may be a Silicon Valley trope, but they also really happen.\u00a0<\/p>\n

The chilly day had turned hot. Deng, wearing a sweater, sweated and scrambled to keep up with the fit and friendly Fedus until the founder said something that \u201cliterally stopped me in my tracks,\u201d Deng told TechCrunch. He told Deng that \u201ceveryone talks about doing science, but in order to do science, you actually have to do science,\u201d Deng recalls.\u00a0<\/p>\n

In other words, they needed to give AI a fully equipped wet lab to try its ideas in a real-world, controlled setting.<\/p>\n

\u201cThe truth about these models is that everything that the models know is within normal distribution. We take a bunch of data, and it can just regurgitate what it knows,\u201d Deng said. \u00a0<\/p>\n

Discovering something new has to involve testing hypotheses.\u00a0\u00a0<\/p>\n

\u201cAnd I committed on\u00a0the spot, in the middle of the hills of Noe Valley, to write the check,\u201d Deng says.\u00a0\u00a0<\/p>\n

Fedus also remembers\u00a0the moment\u00a0Deng asked how he could be involved, and Fedus told him the startup needed cash for laptops and a temporary office. And \u201che\u2019s like, great, I\u2019ll give you money right now. And it was just this huge vote of confidence.\u201d \u00a0<\/p>\n

But Deng didn\u2019t actually whip out his checkbook on the street. He went back to the office elated over the deal only to encounter Felicis\u2019 lawyer, who pointed out that the firm couldn\u2019t promptly sign a contract: The company wasn\u2019t incorporated yet. It didn\u2019t even have a name, much less a bank account to wire funds. \u201cThat\u2019s how early we were.\u201d Deng grinned.\u00a0<\/p>\n

Soon they had all of those things and all the term sheets they could handle. With the $300 million war chest, Cubuk and Fedus hired over two dozen of the most prestigious AI and scientific talent like Alexandre Passos (a creator of o1 and o3); Eric Toberer (a materials scientist who has already made key superconductor discoveries); and Matt Horton, a creator of two of Microsoft\u2019s GenAI materials science tools. And the list goes on<\/a>.\u00a0<\/p>\n

Because the team members are all experts in different areas, from AI to physics, each week one of them gives a grad-level lecture to the others. \u201cWe do feel like a tight coupling is extremely important,\u201d Cubuk said. He wants everyone to understand all parts of what they are building.\u00a0<\/p>\n

Periodic Labs has already set up its lab, too, and is working with experimental data, simulations and testing some predictions. The main initial mission is to find new superconductor materials \u2014 potentially a gold mine discovery. Improved superconductors could power the next era of potent, but lower energy-consuming tech.\u00a0<\/p>\n

But the last part \u2014 the robots \u2014 are not yet up and running. \u201cThey will take a bit to train,\u201d\u00a0Cubuk\u00a0said.\u00a0<\/p>\n

All of this is, of course, a big swing for the fences. AI powered or not, scientific discovery is not typically fast, easy, or predictable. While this team of experts has some indications that they will find what they are looking for \u2014 or make other discoveries along the way (or simply generate valuable data on their failures), there\u2019s no guarantees.<\/p>\n

And we know that model makers themselves are inching their way toward more AI science. Last month, OpenAI VP Kevin Weil said he was launching an OpenAI for Science<\/a> unit at the company to \u201cbuild the next great scientific instrument: an AI-powered platform that accelerates scientific discovery.\u201d<\/p>\n

As for the investor who wrote the love letter, he didn\u2019t win the deal (although Fedus did say that the letter was \u201cvery flattering.\u201d) The other seed investors include\u00a0Andreessen Horowitz, DST, Nvidia\u2019s venture capital arm NVentures, Accel, and angel backers like Jeff Bezos, Elad Gil, Eric Schmidt, and Jeff Dean.\u00a0<\/p>\n

Elad Gil will be speaking about how AI has changed the startup landscape at Disrupt in San Francisco<\/a> on October 29.\u00a0<\/p>\n