The adoption of IAG in drug discovery has gone from being a promise to an emerging reality in 2025. Companies like Absci —with a $20M injection from AMD— and Latent Labs —with $50M led by ex-DeepMind— are marking a transformation .
How it works
Unlike traditional methods, IAG seeks:
-
Designing entirely new molecules , optimizing biochemical properties.
-
Simulate complex interactions between drugs and target proteins.
-
Optimize ADMET profiles (absorption, distribution, metabolism, excretion, and toxicity).
The result: viable proposals in weeks, not years, and with lower failure rates .
Featured companies
- Absci + AMD : They are exploring generative AI for biologics, displacing Nvidia in some partnerships .
- Latent Labs : They design synthetic proteins with multimodal AI, linked to high-caliber investors .
- Isomorphic Labs (Alphabet/DeepMind): Raises $600M for GAM (Generative-Aided Medicine), boosting clinical phase designs .
Preliminary results
- Reduction of molecular design time by 95% .
- Shortened preclinical phase.
- Greater accuracy and lower risks in early stages.
Benefits and challenges
Pros :
- Cost savings (reduction of 60–80%).
- Democratization of compound generation
- Less use of animals thanks to accurate simulations.
Against :
- Ambiguous regulation regarding generative tools.
- Who is responsible if an AI-generated drug fails?
- Need for transparency, since models can generate "shadow molecules" without expecting replicable results.
Are we facing the third medical revolution?
Between Watson Oncology and CRISPR, AI could be the next major disruption, displacing paradigms of "burning down chem libraries" with workflows entirely generated by AI. To integrate these models in the lab, examples like GPTeal (Merck), ImmuneAI, and the massive training of employees—56,000 at J&J—highlight the industry's adaptation to this new era.