Absci: Can Generative AI Finally Make Antibody Discovery Faster, Smarter – and More Investable?

Absci is an AI antibody discovery and engineering platform that combines generative models with high-throughput wet-lab experimentation to design, evaluate, and optimise therapeutic biologics. It addresses a central bottleneck in antibody drug discovery: identifying differentiated, developable candidates for difficult targets can require large experimental campaigns, lengthy iteration cycles, and sequential trade-offs between affinity, specificity, safety, manufacturability, and half-life. Absci’s approach is designed to make these early discovery choices more systematic by generating candidate antibody structures and sequences against selected epitopes, including targets with limited prior structural information.

The platform draws on proprietary experimental datasets alongside public biological data to train models for de novo antibody design, candidate scoring, and multiparameter lead optimisation. AI-generated designs are then tested in Absci’s wet-lab environment, with functional validation data feeding back into subsequent model-development cycles. This lab-in-the-loop model can help research teams prioritise a more focused set of candidates for experimental characterisation, rather than relying solely on broad conventional screening approaches.

For biologics R&D workflows, the potential value lies in linking computational design to experimental feedback while considering several therapeutic attributes together. Absci states that its platform can test binding affinity and target specificity for millions of AI-generated designs per week and operates data-to-validation cycles in as little as six weeks; actual programme timelines and candidate quality will depend on target biology, assay design, development requirements, and experimental results.

Last checked on September 30, 2026: Absci remains active as a generative-AI biologics company and most recently announced preliminary plans for a Phase 2 endometriosis trial of its AI-designed anti-PRLR antibody, ABS-201.

What is Absci?

Absci is an AI antibody discovery and engineering platform for biopharma research teams developing therapeutic biologics. It uses generative AI, proprietary high-throughput wet-lab data, and iterative experimental validation to support de novo monoclonal-antibody design, target discovery, and multi-parameter lead optimisation across properties such as affinity, specificity, developability, half-life, and manufacturability. Its differentiating technical approach is a lab-in-the-loop workflow in which wet-lab validation data feed back into its models; Absci’s Origin-1 platform reports validation on selected “zero-prior” epitopes using orthogonal assays and cryo-EM structural analysis.

Intended Use and Limitations

• Absci is a Nasdaq-listed clinical-stage biopharmaceutical company developing therapeutic biologics using generative AI and high-throughput wet-lab experimentation.

• Its partnership with Almirall expanded to a second dermatology target in 2025 after the parties reported delivery of AI-designed functional antibody leads for their first target.

• Absci’s live partnership materials identify AstraZeneca as a collaborator using its de novo AI antibody-creation platform for a potential oncology antibody therapy.

• Eli Lilly made a $40 million strategic investment in Absci as part of a June 2026 $100 million underwritten offering linked to ABS-201’s clinical-development strategy in endometriosis.

• Absci reported $201.1 million in cash, cash equivalents, and marketable securities as of June 30, 2026, following the financing.

• The company has advanced ABS-201, an AI-designed anti-prolactin-receptor antibody, through Phase 1 and announced preliminary plans for a global Phase 2 endometriosis study expected to begin in mid-2027.

• Absci’s platform combines generative AI with high-throughput wet-lab data and iterative experimental validation, rather than relying solely on computational antibody design.

• Absci reports that its Origin-1 de novo antibody-design platform has been evaluated using orthogonal functional assays and cryo-electron microscopy structural analysis for selected “zero-prior” epitopes.

• No ISO certification, HIPAA alignment, GDPR alignment, FDA clearance, CE marking, or comparable product-level compliance certification was specified in the reviewed live-site and recent public materials.