Unreasonable Labs exits stealth with $13.5M to reinvent scientific discovery
AI that doesn't just know things — it invents them
Scientific research has an information problem. The data is everywhere, but the breakthroughs stay buried.
What's happening: Unreasonable Labs launched from stealth with a $13.5 million seed round, unveiling an AI discovery engine built to move R&D teams from knowledge retrieval to genuine hypothesis generation. The platform pairs large language models with neurosymbolic mathematical frameworks to connect concepts across disciplines and produce novel, validated scientific insights, simulations, and experimental designs.
Numbers don't lie: The Palo Alto and Cambridge-based company is already running pilot programs with industrial partners in energy transition, materials science, and pharmaceuticals, with the new capital earmarked for scaling its technical team across machine learning, simulation engineering, and computational science.
Who's in: Playground Global led the round, with AIX Ventures, E14 Fund, and MS&AD Ventures participating. The firm's $1.2B AUM portfolio includes PsiQuantum and Agility Robotics. The founding team brings serious pedigree: CEO Yuan Cao was a Senior Staff Research Scientist at Google DeepMind with core contributions to the Gemini model family, while CTO Markus Buehler is an MIT engineering professor with decades of AI-driven materials research. Advisors include Nobel Physics laureate Kostya Novoselov, MIT biotech legend Robert Langer, and Hugging Face co-founder Thomas Wolf.
Looking ahead: Unreasonable is betting that the next frontier isn't faster search — it's machines that reason across disciplines the way scientists do at their best. If the platform delivers on its promise of compressing years of R&D into weeks, the implications for drug discovery, materials science, and energy research are substantial. The race to build AI's scientific layer is on, and Unreasonable just made its opening move.