6 Comments
User's avatar
Waqas's avatar

I led an analysis of AI-enabled therapies that have entered interventional clinical trials -- while most assets (over 120 of them!) are still early (ph 1/2), there are numerous exampled of repurposed drugs and also many with novel targets.

https://ascopubs.org/doi/10.1200/JCO.2026.44.16_suppl.11072

Andrii Buvailo, PhD's avatar

Thanks, it seems i credibly useful. I am going to cite your work in the upcoming 2026 AI pharma industry landscape report.

MadeAi's avatar

AI-driven drug rescue could become a major biotech advantage by finding new uses and patient targets for previously abandoned compounds, reducing both cost and development time.

Ilina's avatar

Great breakdown! Thought of this RichStorm article: https://www.richstorm.co/post/how-ai-increases-the-odds-a-failed-drug-becomes-a-winner.

If ~97.5% of drug candidates die before a patent is even filed, there's a whole earlier graveyard of failures that's invisible before you even get to the late-stage stuff these companies are mining. Curious if the same playbook works at both stages.

Andrii Buvailo, PhD's avatar

Thanks for your comment and for the interesting article. Well, I think there is a subtle balance in picking the right combination of factors when mining "failed" molecules. IMO, as long as safety is sorted, the rest is investable. I mean, safety is the crucial and most complex factor in drug development. It represents the key risk, so as long as this is already sorted, the repurposing/rescue business models become viable. So, technically, only after phase 1 can you get this quality of drug candidates... so the window is quite narrow actually.