Finding New Drugs Is Laborious. This AI Startup Is Skipping The Lab Work
Reported by Wired1 source ↓
Potential Sciences, a startup backed by the UK government, is betting that training AI models on computer-generated biochemical data can speed up the plodding process of developing new drugs.
Key points
- Potential Sciences plans to build simulated chemical datasets to train drug-discovery AI
- The startup says its seed funding is in the tens of millions of dollars; it did not disclose the exact sum
- SovAI, the UK government’s AI investment fund, backed the startup
- The approach aims to identify drug candidates, but safety and efficacy testing would still be required
Details
Wired reports that Potential Sciences will use physics-based simulations of atoms, molecules and proteins to generate data for AI models and prediction systems. Founder and CEO Hatem Helal says the goal is to predict how drug candidates interact with human biology and rule out candidates unlikely to work before lab testing.
The startup says its seed funding is in the tens of millions of dollars from SovAI, Twin Path Ventures and other investors; it did not disclose the precise amount or its valuation. The UK government established SovAI earlier in 2026 and has roughly $800 million available to invest, according to Wired.
Whether the models can predict drug-protein binding accurately enough remains unproven. Experts quoted by Wired see potential in simulation-based data but note that identifying candidates would not remove the time and cost of testing drugs for safety and efficacy.
Written from coverage as of
Images
- Wired