$12M+ · AI + Drug Discovery

What they're building

Drug discovery today is a handoff problem. A research team surfaces a candidate. The insight dies at the interface with the next team. The clinical group starts over with partial information. Decisions worth billions get made on incomplete pictures because the systems involved were never designed to communicate. Perceptic is the connective tissue: a shared AI intelligence layer that follows the drug, not the department. The platform has three components. Scout handles asset scouting, rapidly triaging external licensing candidates and competitive programs against a company's strategic objectives, compressing weeks of due diligence into hours. PercepticOS is the internal intelligence layer: it connects a pharma company's proprietary tools and data, lets scientists test hypotheses, and builds a persistent institutional knowledge base that does not vanish when a project ends. Atlas is the clinical data foundation, consolidating internal and external trial data for deeper analysis, and has achieved a 50-fold increase in clinical data extraction speed at production deployments. Every output traces back to its source. No hallucinations. Every claim auditable. The architecture is model-agnostic: customers bring their own data, hardware, and AI tools. Perceptic provides the operating system underneath.

Why this matters

The AI in drug discovery market stood at $1.72 billion in 2024 and is projected to surpass $8.5 billion by 2030. But the market has been building instruments, not infrastructure. Isomorphic Labs, Recursion, Insilico Medicine: all of them operate inside a single stage of the pipeline. Molecule design. Protein structure. Patient recruitment. Each is a better instrument for a specific step. None of them fixes the handoff. That is the problem Perceptic is solving, and it is the one that every pharma executive feels viscerally. A drug approved today took over a decade to develop and cost upwards of $2 billion. A significant portion of that cost is friction, not science. Insights disappearing between teams. Institutional knowledge evaporating between projects. Promising paths unexplored because the researchers did not know someone else had already ruled out a variant. As Flock told Fortune: "Too many critical drug development decisions still happen without a complete view of the evidence." No AI-discovered drug has yet completed human clinical trials and been approved. The scepticism is real and warranted. Perceptic's bet is that the bottleneck is not molecular intelligence. It is the operating system that connects molecular intelligence to the decision-making chain. If they are right, they are building the category's foundation rather than competing inside it.