$50M Seed Brings Generative Biology to Life with AI That Writes DNA

Artificial intelligence has mastered language and vision—now it is learning to write the language of life. Radical Numerics' $50 million seed round propels generative genomics from research lab to commercial platform, leveraging models trained on genomes from over 100,000 species. For the AI community, this is a landmark demonstration of multimodal, deeptech models moving beyond narrow tasks to full-stack biological reasoning, with open-source roots that raise both collaborative and safety considerations.

Radical Numerics, a startup founded by the creators of generative genomics, has emerged from stealth with a $50 million seed round to commercialize AI that can read, write, and reason in the language of biology. The funding, led by Emergence Capital with participation from Obvious Ventures, Triatomic Capital, Factory, First Spark Ventures, and pre-seed backer Patrick Collison of Stripe and the Arc Institute, signals a new era where artificial intelligence is not just analyzing biological data but actively designing it. The company's founding team—CEO Eric Nguyen, Chief AI Scientist Michael Poli, President Stefano Massaroli, and CTO Armin Thomas—were instrumental in building Evo and Evo 2, the first AI models capable of generating DNA sequences at scale, trained on the genomes of more than 100,000 species.

“

The AI drug discovery market is projected to reach $25 billion by 2035, and Radical Numerics enters a competitive landscape alongside well-funded players like Xaira Therapeutics, which raised over $1 billion.

The leap from academic research to startup was catalyzed by a remarkable milestone. In September 2025, researchers using Evo's open-source weights produced the world's first fully AI-designed functional virus—harmless to humans but profoundly significant for what it represented. It proved that generative AI could craft working genetic code from scratch, a capability that could compress decades of drug discovery and synthetic biology into years. As Nguyen told Fortune, the academic community wasn't picking up the work as fast as he'd hoped, prompting the team to 'show the recipe' by building a company.

The radical premise of Radical Numerics is to move beyond single-purpose AI tools—like those predicting protein structures or analyzing individual genomes—to a multimodal model that understands DNA, RNA, proteins, and the entire molecular interplay of living systems in one unified framework. Three of the four cofounders previously built core technology at Liquid AI, an MIT spinout focused on novel AI model designs, and their Evo models represent a fundamentally new approach to generative biology. With $50 million in seed capital, the company is now positioned to translate this deeptech into real-world applications.

The immediate use cases are transformative. In drug discovery, generative genomics could create novel therapeutic candidates, design gene therapies, or engineer microorganisms for biomanufacturing. The AI drug discovery market is projected to reach $25 billion by 2035, and Radical Numerics enters a competitive landscape alongside well-funded players like Xaira Therapeutics, which raised over $1 billion. However, the startup's core differentiator is its foundation models that can generate functional DNA from scratch, not merely predict molecular interactions.