Leading AI Researchers Depart OpenAI, Google, and Meta to Pioneer AI-Driven Scientific Discovery Startup
In a notable shift within the artificial intelligence sector, more than 20 top researchers have left prominent AI divisions at OpenAI, Google DeepMind, and Meta to join a new Silicon Valley startup, Periodic Labs. This startup aims to leverage AI to accelerate the pace of discoveries in physics, chemistry, and other scientific fields.
Periodic Labs, founded by a co-creator of ChatGPT, represents a new frontier in AI applications, focusing on real-world scientific breakthroughs rather than just language modeling or commercial AI products. The startup has attracted significant talent, including Rishabh Agarwal, formerly of Google DeepMind and Meta’s newly formed superintelligence lab.
Mr. Agarwal’s decision to join Periodic Labs came despite lucrative offers from Meta’s CEO Mark Zuckerberg. Zuckerberg had personally invited Agarwal to join Meta’s ambitious new AI lab with promises of substantial stock options and salary, emphasizing a vision to build “superintelligence,” a form of AI capable of surpassing human intelligence. However, Agarwal opted instead for the riskier venture at Periodic Labs, motivated by the startup’s mission to create AI tools that facilitate scientific advancements.
Meta’s superintelligence initiative has seen significant turnover, with several researchers leaving shortly after joining. Some, like Avi Verma and Ethan Knight, returned to OpenAI after brief stints at Meta. The departures highlight the competitive and rapidly evolving landscape of AI research, where talent mobility reflects differing visions and priorities across companies.
Periodic Labs’ focus on AI for scientific discovery marks a distinct strategic direction compared to the large tech companies concentrating on AI-powered products such as language models and generative AI. By employing advanced AI techniques to analyze and experiment with physical materials and chemical compounds, the startup hopes to revolutionize how new materials and scientific innovations are discovered.
The movement of top-tier researchers from established AI giants to a nascent startup underscores the shifting dynamics in the AI research community. It suggests a growing interest in applying AI beyond traditional domains into solving complex scientific problems, aiming to translate AI’s potential into tangible real-world impacts.
This trend also illustrates the tension between the promise of working at well-funded, prestigious organizations and the allure of startup environments offering freedom to pursue more speculative, high-stakes goals. Many researchers appear willing to forgo significant financial gains for the opportunity to lead transformative projects at emerging ventures.
As Periodic Labs garners more attention and talent, it could become a pivotal player in the AI ecosystem, potentially challenging the dominance of tech giants by pioneering AI’s role in scientific innovation. The company’s progress will be closely watched by industry insiders aiming to gauge the future directions of AI research and its real-world applications.
This phenomenon highlights the expanding scope of AI research and the diversification of aims among the world’s top AI thinkers — from building general-purpose AI models to crafting specialized tools that can accelerate discovery across multiple scientific disciplines.