Genesis Molecular AI

Market position

Genesis is positioned among AI-native drug discovery companies; third-party listings name Relay Therapeutics, Atomwise, Verge Labs, Schrödinger, insitro and Nurix Therapeutics as similar companies. Validation cited by third parties includes collaborations with Gilead and Incyte, backing from a16z, Menlo Ventures, BlackRock, Radical Ventures, Felicis, Fidelity and NVIDIA's NVentures, and more than $300M reported raised (with specific totals varying by source between $296M and over $300M).

The company positions itself around models purpose-built for molecular data rather than adapted general-purpose AI, unifying learned foundation models with physics-based methods and large-scale physics-derived synthetic data to address the scarcity of experimental training data; it claims to be the first to demonstrate LLM-like scaling laws with this approach. Other stated differentiators include the Pearl structure-prediction model, an integrated wet lab that closes the design-make-test loop, forward-deployed engineers and drug hunters embedded on partner programs, and access to partner experimental datasets (notably from Incyte) for model training.

Technology

GEMS (Genesis Exploration of Molecular Space) is described as an AI operating system for molecular design that integrates deep learning with physics-based methods to generate, optimize and select molecules for synthesis and testing. Research areas include generative diffusion models, equivariant/geometric 3D deep learning, reinforcement learning with physics-based reward functions for multi-objective optimization, multi-modal models integrating 3D structures with bioassay results, and agentic workflow orchestration. Pearl is the company's generative diffusion foundation model for 3D biomolecular structure prediction, trained on physics-based synthetic data and developed with NVIDIA. A third-party listing reports a corporate tech stack including GitHub, MySQL, Microsoft 365, Google Cloud, Google Sheets, Microsoft Azure, Paylocity and Google Search Console.

Go-to-market

Genesis pursues two channels: multi-target strategic collaborations with large biopharma companies (Gilead, Incyte) in which its forward-deployed engineers and scientists work alongside partner discovery teams on partner-selected targets, and its own internal discovery programs. Publication, open-source contributions, academic collaborations and conference sponsorships (including a joint announcement with NVIDIA at GTC) support visibility in the ML research community.

Large pharmaceutical and biotechnology companies seeking AI-enabled discovery of small molecules against chemically complex, data-poor or canonically undruggable targets; ultimately patients with difficult cancers (breast, colorectal) and autoimmune/inflammatory diseases.

Geography

United States-based with multiple sites: AI research in the Bay Area and New York City and a fully integrated wet lab in San Diego. Headquarters is reported inconsistently across sources as Burlingame, California and South San Francisco. One aggregator states employees are distributed across four continents including North America, Europe and Asia.

History

Internal records date the company's founding to 2019 under the name Genesis Therapeutics. It closed a $200M Series B in 2023 co-led by a16z with new investors including Fidelity, BlackRock and NVIDIA (NVentures). In 2024 it announced a collaboration with Gilead Sciences and received an equity investment and strategic collaboration from NVIDIA. In 2025 it signed an initial collaboration with Incyte, unveiled the Pearl foundation model (jointly announced with NVIDIA at GTC 2025), and appointed Sergey Edunov to lead foundation models. By late 2025 the company was operating under the name Genesis Molecular AI. In May 2026 Incyte expanded the collaboration to at least five additional targets with $120M in upfront consideration and an experimental data-sharing component.

Risks & controversies

Public sources disclose no controversies. Data inconsistencies exist across aggregators: total funding is variously reported as $40.0M, $296M and 'over $300M'; the May 2026 round size is reported as both $40M and $200M; headquarters is listed as Burlingame in one source and South San Francisco in another; and Sergey Edunov's title is given as SVP of Foundation Models in press coverage but CTO on the company website. Pipeline programs are at discovery/pre-development-candidate stage, and several targets are undisclosed.