This AI Startup Lets Biology Teach the Machine, Not the Other Way Around

Jan de Vries · 2026-07-17

Most AI drug discovery companies start with a target. They design a molecule against it, then test it in cells. Sightera Biosciences flips that script completely. The Antwerp-based startup, a spin-off from the University of Antwerp and Antwerp University Hospital, just closed a $3.2 million pre-Seed round led by Entourage, Anacura, and QBIC. That money will go toward expanding their team, building out their AI-native drug discovery platform, and pushing their lead oncology program toward preclinical candidate selection. ### What Makes Sightera Different? Here's the thing about most AI drug discovery: the models are trained on public data. Generic data. Data that doesn't capture the messy reality of human disease. Sightera takes a radically different approach. They start with patient-derived biological material. Specifically, they work with samples from patients with advanced, therapy-resistant disease. The kind of "extreme biology" where conventional treatments have failed and the underlying biology has fundamentally rewired itself. From these samples, they build scalable preclinical models - including patient-derived organoids - that preserve that extreme disease biology exactly as it appears in humans. Then they run high-throughput drug screening campaigns to generate rich, dynamic datasets showing how real patient biology responds to therapies. Pieterjan Bouten, General Partner at Entourage, put it well: "Sightera doesn't ask AI to predict biology. It lets biology teach AI." That's not just a clever tagline. It's a fundamentally different approach to drug discovery. ### The Problem with Traditional AI Drug Discovery Most platforms design molecules against a predefined biological target. They optimize for chemical properties - stability, solubility, selectivity. Then they test in complex biological models and hope for the best. That's why so many AI-designed candidates fail in translation to the clinic. The molecules look great on paper but don't work in real human biology. Sightera reverses that paradigm entirely. They start with complex patient-derived biology, learn directly from therapeutic responses, and use those insights to design entirely new small molecules. The biological behavior is engineered from day one. ### The Team Behind the Platform Sightera was founded in January 2025 by a team with deep domain expertise: - Hendrik Vercammen (CEO) - Maxim Le Compte (CSO) - Lars Vanlommel (CBO) - Christophe Deben (CDO) The company spun out from the University of Antwerp and Antwerp University Hospital in January 2026. Their current focus is oncology and fibrosis, two areas where patient biology varies enormously and where traditional drug discovery has struggled. ### What's Next? The $3.2 million round is deliberately structured as a first step. The company is already preparing a larger financing round to take their lead program, SIGHT001, into clinical trials. "Our goal is not simply to accelerate drug discovery, but to improve the probability of clinical success by placing human disease biology at the very beginning of the drug discovery process," said Vercammen. That's a bold claim. But if they can deliver, it could change how the entire industry thinks about AI in drug development. ### Why This Matters Now The timing is interesting. Billions in blockbuster drug revenues are approaching the patent cliff. Pharma companies urgently need faster paths to the next generation of medicines. Traditional approaches take years and fail more often than they succeed. Sightera's approach - learning directly from patient biology - promises to identify promising new molecules in months instead of years. If that holds true in the clinic, it could reshape the economics of drug development. We'll be watching closely to see how SIGHT001 performs in trials. But for now, Sightera represents one of the more interesting approaches to AI drug discovery we've seen come out of Europe.