Key Highlights
- Merck taps Variational AI’s Enki™ platform—with a deal value up to USD 349M—to accelerate small-molecule programs across two undisclosed targets
- Fine-tuned generative models trained on Merck’s proprietary datasets set new expectations for speed, precision, and early-stage success rates
- Collaboration underscores pharma’s rapid shift toward foundation-model AI to reshape discovery economics and expand druggability
A Strategic Alliance Targeting Hard-to-Crack Therapeutic Space
Variational AI announced a major collaboration with Merck to deploy the Enki™ generative AI platform for the design and optimization of small-molecule candidates across two Merck-designated targets. The partnership positions Enki™ as a high-impact engine for early-stage drug creation, leveraging both companies’ strengths: Merck’s deep proprietary datasets and Variational AI’s frontier machine-learning capabilities. Merck will retain exclusive rights to advance and commercialize any molecules emerging from the alliance.
Fine-Tuned Foundation Models Built on Merck’s Data Advantage
A core element of the agreement is the training of a fine-tuned Enki™ model using Merck’s proprietary chemistry and biology data. This combined dataset—paired with Variational AI’s curated public corpus and internal molecular intelligence—enables the creation of generative models with higher accuracy, better fitness-to-target profiles, and improved molecular manufacturability. Variational AI’s CEO Handol Kim emphasized that this framework may meaningfully accelerate design cycles while enhancing economic efficiency across discovery programs.
Deal Value Up to USD 349M Signals Rising Confidence in Generative Chemistry
Under the collaboration, Variational AI receives an upfront payment and is eligible for milestone-based payouts totaling up to USD 349 million. This scale of financial commitment reflects Merck’s growing prioritization of AI-native discovery infrastructure, as highlighted by Robert M. Garbaccio, Ph.D., who noted Merck’s intent to boost speed, quality, and efficiency early in the pipeline. The deal aligns with a broader industry trend in which foundation models—not just point-solution algorithms—are becoming central to discovery strategies.
Enki™ as a Platform for Optimized Lead Creation
Enki™ is built on a generative AI foundation model trained on blended proprietary and publicly curated datasets, enabling rapid design of small molecules tuned to partner-defined product profiles. With origins tracing back to researchers from MIT, Caltech, Google Research, Microsoft Research, and D-Wave Quantum, Variational AI aims to push the boundaries of generative chemistry. Its platform focuses on improving probability of success, enhancing molecular novelty, and drastically reducing cycles required to reach optimized leads—supporting a new era of accelerated therapeutic innovation.
Variational AI is a Vancouver-based generative AI drug discovery company founded by scientists from MIT, Caltech, Google Research, Microsoft Research, and D-Wave Quantum. Its proprietary Enki™ platform uses advanced generative AI foundation models trained on internal datasets, curated public molecular libraries, and partner-specific proprietary data to design high-quality, optimized small molecules with improved probability of success. The company aims to redefine early-stage drug discovery by accelerating hit generation, enhancing molecular novelty, and improving downstream manufacturability.
Website: https://www.variational.ai
Merck (known as MSD outside the United States and Canada) is one of the world’s leading biopharmaceutical companies, focused on discovering, developing, and delivering breakthrough medicines and vaccines for some of the most challenging diseases. Through the Merck Research Laboratories, the company invests deeply in next-generation technologies—including AI-powered discovery platforms—to improve the speed, quality, and success of therapeutic development.
Website: https://www.merck.com
By