AI isn’t close to curing cancer. This startup says it knows what it will take.

A biotech startup argues that AI drug discovery has a fundamental problem: existing AI models lack the kind of biological data needed to meaningfully advance medicine, instead relying on animal testing or studies of isolated cells rather than living human tissue. The startup has built robotic laboratory systems that can grow multiple types of human tissue, automatically test drug candidates on them, and generate data about how drugs affect actual human biology, which the company says is essential for training AI models that truly understand human biological processes. This matters because many AI leaders have made bold claims about AI curing cancer, but actual progress has been limited, with most drugs that succeed in animal testing still failing in human trials. The startup's approach represents a different strategy: using better biological data to train AI models and to predict drug effectiveness before expensive human clinical trials begin.