Ex-OpenAI and DeepMind talent chase $7B for AI scientists
A new wave of ambition is sweeping through the artificial intelligence sector as a group of former OpenAI and DeepMind researchers seek a valuation of around $7 billion for a startup focused on building autonomous AI scientists. The company, whose name has not yet been publicly disclosed, is positioning itself at the frontier of using AI models to design, run and interpret complex scientific experiments with minimal human oversight.
From large language models to autonomous discovery
The team’s vision goes beyond today’s generative AI and large language models. Instead of merely assisting with drafting papers or writing code, these proposed AI scientists would be tasked with formulating hypotheses, planning experiments, running simulations, and even controlling laboratory equipment through integrated software and robotics.
Backers are said to be attracted by the prospect of dramatically accelerating research in fields such as drug discovery, materials science and climate modelling. By combining advanced AI algorithms with automated labs, the startup aims to compress years of trial-and-error into weeks or days, potentially reshaping how fundamental science and industrial R&D are conducted.
Investors weigh promise against safety concerns
The search for a $7 billion valuation at an early stage underlines how aggressively capital is chasing the next generation of AI infrastructure. Venture firms and strategic investors view autonomous scientific systems as a potential platform play, comparable in importance to the first wave of cloud computing or foundation models.
However, the concept of machine-led discovery is also raising questions about AI safety, governance and reproducibility. Critics warn that delegating more of the scientific process to opaque models could make it harder to audit results or understand how key breakthroughs were reached. Supporters counter that rigorous validation pipelines, transparent data practices and human-in-the-loop review can keep systems accountable while still delivering unprecedented speed.
A new race in AI-native science platforms
If the raise is successful, the startup would instantly join the top tier of privately held AI companies by valuation and intensify competition with existing players building automated labs and scientific copilots. For now, the move signals that the next battleground in advanced artificial intelligence may be less about chatbots and more about who can industrialise scientific discovery itself.