Radical Numerics launched with $50 million in seed funding to build multimodal AI models trained directly on biological data spanning DNA, RNA, and proteins. The round was led by Emergence Capital with participation from Obvious Ventures, Triatomic Capital, Factory, and First Spark Ventures. The startup said it is previewing a next-generation genomic language model called Omnii and is partnering with an undisclosed cancer diagnostics company for pancreatic and multi-cancer detection use cases. Scientific advisors include Microsoft CSO Eric Horvitz and Harvard’s George Church, signaling an effort to bridge frontier AI with translational bioinformatics. For biotech, the practical question is how these models translate into validated targets, assays, and real-world decision support—particularly in diagnostic settings where robustness and interpretability are critical.