Rakovina shows how GenAI can cut drug discovery time while improving precision in complex cancer therapies.

Drug discovery often stalls on two problems. Candidate design takes years. Promising therapies fail from toxicity or poor delivery. Rakovina is using GenAI to tackle both, drug discovery and therapy. Its models generated dual-action cancer compounds and optimized lipid nanoparticles for delivery. Because of this, it shifts GenAI from molecule screening into therapeutic design. The breakthrough is not just faster discovery. It is designing for efficacy. As well as safety, and manufacturability at once.

One result targets a hard problem in oncology. Treating brain-related cancers remains difficult because many drugs cannot cross into the central nervous system. Rakovina used GenAI to co-optimize potency, selectivity, and brain penetrance together. That challenge usually demands long trial-and-error cycles. Preclinical results showed tumor control matched leading comparators, with better tolerability. As a result, this signals GenAI can help solve tradeoffs traditional methods struggle to balance.

A second GenAI system tackled drug delivery. Combination therapies often raise toxicity and formulation risks. Rakovina used GenAI to design nanoparticles that improve uptake and stability for a bifunctional therapy. This compresses formulation work and lowers development friction. More important, GenAI moved beyond molecule generation into delivery engineering. And so this expands where enterprise drug developers can apply these models.

This is where GenAI becomes an R&D engine. It reduces search across billions of molecular possibilities. It helps teams design candidates with fewer dead ends. For pharma, that means lower discovery costs. As well as faster lead progression, and potentially better drugs. For GenAI enthusiasts, this shows models are moving from copilots into scientific invention systems.

Why it matters
This case shows GenAI solving a major enterprise problem: slow, expensive, low-yield R&D.
• Moves GenAI from screening tools into end-to-end drug design.
• Shows models can optimize multiple scientific constraints at once.
• Proves GenAI can reduce risk in both discovery and delivery.
• Signals a repeatable enterprise pattern for R&D-heavy industries.