Variational

Variational AI is a generative AI drug discovery company founded in 2019, dedicated to transforming the way novel small molecule therapeutics are identified and developed. Rather than relying on traditional screening of existing compound libraries, Variational AI generates entirely new, synthesizable drug-like molecules using advanced generative AI — enabling biopharmaceutical partners to explore vast, previously unreachable regions of chemical space faster and more cost-effectively. The company serves R&D teams across the biopharma industry, with an initial focus on oncology and an expanding presence across other high-need therapeutic areas.

At the core of Variational AI's mission is the recognition that drug discovery is increasingly expensive, slow, and constrained. Costs have roughly doubled every nine years since 1949, and the standard Design-Make-Test-Analyze (DMTA) cycle can consume years per program. With an estimated 10⁶⁰ possible drug-like molecules in existence — and current methods having explored far less than a billionth of that space — Variational AI believes the solution lies not in screening more, but in generating better candidates from the outset.

Enki™ — Generative AI Foundation Model for Drug Discovery

• Enki™ is Variational AI's fully-trained generative AI foundation model, purpose-built for early-stage small molecule drug discovery.
• Rather than ranking or filtering known compounds, Enki™ generates novel, diverse, and synthesizable molecules de novo — freeing discovery teams from dependence on experimental or virtual screening libraries.
• Enki™ operates within a smooth, physics-informed latent space that reflects pharmacological behavior, enabling it to generate molecules optimized for success across more than 50 physico-chemical and biological criteria.
• Optimization targets include potency, selectivity, toxicity, pharmacokinetics (PK), ADME properties, synthesizability, and additional physico-chemical parameters — all addressed simultaneously in a multi-objective framework.
• The platform is designed to generate first- and best-in-class compounds across multiple therapeutic areas, with synthesis-aware generation and clear rationales behind each molecular proposal to accelerate decision-making.
• Enki™ also supports hyper-efficient lead optimization when constrained to a defined scaffold, enabling iterative refinement without sacrificing novelty.

Enki™ Workflow — From Target Profile to Synthesis-Ready Leads

• Step 1 – Define the preclinical target product profile (TPP): Users specify on-targets, off-targets, and desired physico-chemical properties as inputs to guide Enki™'s generative process.
• Step 2 – Enki™ generates compounds: The platform produces novel and structurally diverse candidate molecules that meet the defined TPP, exploring uncharted chemical space without relying on known scaffolds.
• Step 3 – Selection and optimization: Teams select compounds for synthesis and testing; Enki™ can further perform lead optimization constrained to a chosen scaffold, reducing iteration cycles and accelerating progression to lead series selection.
• The overall workflow is designed to deliver synthesis-ready leads in weeks, significantly compressing timelines compared to conventional discovery approaches.

Scientific and Technical Capabilities

• Enki™ generates molecules that are both novel and practically synthesizable, providing higher-quality starting points and seeding pipelines with candidates that balance multiple properties simultaneously.
• The platform's physics-informed latent space captures key molecular properties, enabling true exploration of previously unreachable chemical regions with speed and precision.
• Multi-objective optimization across more than 50 criteria — including potency, selectivity, toxicity, PK, ADMET, and synthesizability — is performed concurrently, reducing the need for sequential iterative cycles.
• Enki™ is described as a radical departure from the traditional DMTA cycle, not a marginal improvement, enabling a fundamentally different approach to small molecule drug discovery.

Collaborations, Partnerships, and Notable Use Cases

• Variational AI entered a collaboration with Merck (known as MSD outside the United States and Canada) to apply the Enki™ platform to drug discovery programs, announced in September 2025.
• In collaboration with Rakovina Therapeutics, Variational AI has advanced CNS-penetrant ATR inhibitors using Enki™, with results demonstrating compounds that balance potency and selectivity.
• Variational AI partnered with Life Chemicals for a second time to discover de novo generated selective dual EGFR/FGFR1 inhibitors using generative AI, building on a prior 2022 collaboration targeting the SARS-CoV-2 main protease.
• Variational AI joined the Agora Open Science Trust initiative focused on optimizing PRMT6 inhibitors for spinal and bulbar muscular atrophy (SBMA).
• The company participates in leading industry conferences including Drug Discovery Chemistry and investor-focused events such as the Bloom Burton Healthcare Investor Conference.

Team, Investors, and Governance

• Variational AI was co-founded by Handol Kim (CEO), Jason Rolfe, PhD (CTO), and Ali Saberali, PhD (Head of Platform), and is led by a multidisciplinary team spanning machine learning, medicinal chemistry, computational chemistry, and biopharma R&D.
• AI scientists on the team have been trained at and conducted research with leading institutions including MIT, Caltech, UC Berkeley, Google, Microsoft Research, and D-Wave Quantum.
• The company's Scientific Advisory Board includes experts in cell biology, oncology drug discovery, cheminformatics, and medicinal chemistry.
• Variational AI is backed by top-tier investors including Nimbus Synergies, Merck Global Health Innovation Fund, Flying Fish Ventures, A&E Investments, Nepenthe Capital, and a network of angel investors.
• Board of Directors members include representatives from Flying Fish Partners, Merck GHIF, and Nimbus Synergies, alongside the company's co-founders.

Variational AI's cross-disciplinary approach — combining cutting-edge generative modeling with rigorous medicinal and computational chemistry expertise — positions Enki™ as a practical, science-grounded solution for real-world drug discovery programs. By redefining the unit economics of early discovery, the company aims to help biopharma partners bring novel therapeutics to patients faster and more efficiently across oncology and beyond.