Variational AI builds Enki™, a foundation model tailored to small-molecule drug discovery that uses far less compute and data than foundation models in other domains. The company says Enki was trained on a proprietary dataset covering almost 600 targets and generates diverse leads with strong potency and selectivity. Partners synthesize and test around 20 novel molecules from multiple scaffolds per project over a few weeks, and Variational reports a sub-micromolar hit rate of over 50% with a 90% rate of synthetic success. The founding machine-learning team comes from Google, Microsoft, MIT, Caltech, and D-Wave Quantum and works with established drug-discovery leaders. Variational AI is based in Vancouver, BC, and is actively hiring. The company completed an oversubscribed $5.5 million seed extension to support market expansion of its compute-efficient foundation model. Variational AI develops state-of-the-art generative AI to accelerate the discovery and optimization of novel small molecules for drug development. The company uses data-efficient machine learning to rapidly generate novel, diverse compounds optimized for multiple properties to reduce drug attrition. It works with leading biopharmaceutical partners and is developing an internal pipeline across disease areas. Variational AI was founded in September 2019 and combines AI/machine-learning and business specialists. The company plans to use new funding to accelerate discovery programs in COVID-19, cancer, and other disease areas, and to expand its team and advisory board. Operationally, the seed funding improves its ability to hire leaders in medicinal chemistry, cheminformatics, and machine learning and to recruit a Scientific Advisory Board.

Total raised

$9M

Funding rounds

2

Latest round

Seed

Latest activity

Feb 2025