OpenAI and Google Brain researchers have launched a $300 million venture capital frenzy to start up their spin labs

League Labs, a new startup founded by one of OpenAI’s most respected researchers, Liam Fedus, and his former Google Brain colleague, Ekin Dogus Cubuk, came out of stealth last month with a massive $300 million seed round. It was led by Felicis and included a group of angels and other top venture capitalists.

The startup began when Fidus had a conversation with Coppock (whom his friends call “Dougie”) about seven months ago. Cubuk was one of the leading machine learning and materials science researchers at Google Brain. After Silicon Valley endlessly covered how GenAI could radically change scientific discovery, they decided the pieces were finally in place to make this a reality. Or at least a startup that tried to.

“There are some things that have happened in the LLM field, in experimental science and in simulations, that kind of made this the right time,” Coppock told TechCrunch.

For example, he said robotic arms that can handle powder synthesis — the process of mixing and creating new materials — have recently proven reliable. On the other hand, machine learning simulations have become powerful and accurate enough to model complex physical systems such as those needed to develop new materials.

Third, LLM students now have strong reasoning abilities – in part through the work of Fidus and his team at OpenAI. Fedus was one of the small team that initially created ChatGPT and manages the all-important post-training team at OpenAI, which improves the models after their initial development.

Taken together, the picture was clear: a simulator could theoretically detect new compounds, a robot could mix materials, and an LLM student could analyze the results and suggest course corrections. AI automated materials science was ready to be built.

In fact, Cubock was one of the researchers who published a groundbreaking paper in 2023 documenting a previous Google research project. The team built a fully automated laboratory powered by robots and created 41 new compounds from recipes suggested by linguistic models.

Just as importantly, the founders realized that even failed experiments would be valuable to their new startup because data is the lifeblood of AI. The science of artificial intelligence has provided an entirely new source of realistic training and post-training data. The founders believe this could upend the current scientific incentive system, which seeks success, not exploration, and is rewarded through paper publication and grants.

“Connecting with reality, bringing experiences into the world [AI] “We feel like this is the next frontier,” Vidos told TechCrunch.

Felicis wins the deal. OpenAI does not invest

After that discussion with Cubuk, Fedus went to the powers that be at OpenAI to share his resignation and his plan. He then happily tweeted to the world that he was leaving with what appeared to be OpenAI’s blessing and investment.

But this investment did not actually materialize. The founders confirmed to TechCrunch that OpenAI is not a backer of the journal. While Fidus declined to say why, they actually didn’t need OpenAI’s money.

Fedus’ tweet sparked a frenzy among venture capitalists courting the company. “It almost felt like the opposite,” Fidos laughed. One investor wrote a love letter to Periodic Labs, explaining that neither he nor Coppock “knew what to do about it.” Others sent multi-page documents promoting themselves.

But the first call they received was from Peter Deng, a former OpenAI colleague turned investor in leading seed company Felicis. (Ding left OpenAI and joined Felicis in early 2025.)

“Liam is a very big figure at OpenAI, very well-liked and a very influential researcher,” Ding told TechCrunch. “When I heard he left, I texted him immediately.”

Deng met Fidos for coffee in San Francisco’s Noe Valley neighborhood. Fueled by enthusiasm and caffeine, Fidus invited Ding to end their conversation on a hike on the region’s famous mountainous terrain. Pitch walks may be a Silicon Valley trope, but they also actually happen.

The cold day has turned hot. Ding, who was wearing a jacket, was dripping with sweat and struggling to keep up with the physicality and friendliness until the founder said something that “literally stopped me in my tracks,” Ding told TechCrunch. “Everyone talks about doing science, but in order to do science, you actually have to do science,” Deng recalled.

In other words, they needed to give the AI ​​a fully equipped wet lab to try out its ideas in a real-world, controlled environment.

“The truth about these models is that everything the models know falls within a normal distribution,” Ding said. “We take a bunch of data, and they can only regurgitate what they know.”

Discovering something new must involve testing hypotheses.

“I immediately committed myself, in the hills of Noe Valley, to writing the check,” Ding says.

Fidos also recalls the moment when Ding asked how to get involved, and Fidos told him that the startup needed money to buy laptops and a temporary office. “He says, ‘Great, I’ll give you the money now.’ And that was just a big vote of confidence.”

But Deng did not actually take his checkbook out into the street. He returns to the office, elated by the deal, only to meet Felicis’s lawyer, who points out that the company cannot sign the contract right away: the company is not yet incorporated. He didn’t even have a name, let alone a bank account to transfer money. “That’s how early we arrived,” Ding grinned.

Soon they had all this stuff and all the term papers they could work with. Using their $300 million war chest, Cubuk and Fedus have hired more than two dozen top science and AI talents such as Alexandre Passos (creator of o1 and o3); Eric Toberer (a materials scientist who has already made key discoveries in superconductors) and Matt Horton, creator of two of Microsoft’s GenAI materials science tools. The list goes on.

Since the team members are all experts in fields from artificial intelligence to physics, each week one of them gives a graduate-level lecture to the others. “We feel that tight coupling is very important,” Coppock said. He wants everyone to understand all parts of what they are building.

The medical journal company has also set up its own laboratory, working with experimental data and simulations and testing some predictions. The main initial task is to find new superconducting materials, which could be a discovery in a gold mine. Improved superconductors could support the next era of powerful, yet less power-hungry technology.

But the last part – the robots – is not yet operational. “They’re going to take some time to train,” Keubock said.

All of this, of course, amounts to a big swing for the fences. Whether powered by AI or not, scientific discoveries are usually not quick, easy, or predictable. While this team of experts has some indication that they will find what they are looking for — or make other discoveries along the way (or simply generate valuable data about their failures), there are no guarantees.

We know that model makers themselves are moving towards more AI science. Last month, OpenAI vice president Kevin Weil said he was launching the company’s OpenAI science unit to “build the next great scientific tool: an AI-powered platform that accelerates scientific discovery.”

As for the investor who wrote the love letter, he didn’t win the deal (although Fidos said the letter was “absolutely excellent.”) Other cornerstone investors include Andreessen Horowitz, DST, NVIDIA’s venture capital arm NVentures, Accel, and angel backers like Jeff Bezos, Elad Gil, Eric Schmidt, and Jeff Dean.

Elad Gil will talk about how AI is changing the startup scene at Disrupt in San Francisco on October 29.