Expanding to New Healthcare Markets: The Growing Impact of Medical AI

Artificial intelligence in medicine is most often associated with detecting abnormalities in medical images, analyzing test results, and helping doctors make diagnoses. But lately, I’ve been encountering more and more developments that go far beyond these applications and challenge the way we think about what medical technology can achieve. Some are so ambitious that their potential impact could extend well beyond healthcare, affecting entirely different areas of our lives and giving rise to new technologies, business models, and even markets that don’t yet exist.

One of the most interesting areas I’ve been following lately is drug discovery, the search for and development of new drug compounds using computational methods and artificial intelligence. DeepMind’s work on AlphaFold has played a major role in advancing this field. Its database of more than 200 million predicted protein structures has given researchers a wealth of information to explore potential molecular interactions and has also opened new opportunities for startups targeting specific medical conditions.

Suppose someone wants to build a startup to develop a treatment for a relatively rare disease that affects only a small percentage of women in a particular age group. The first question: why should they target such a limited market when they could pursue a much more common condition? But even after narrowing the global population down to this specific group, they could still have thousands of potential patients. And if their treatment requires patients to take a daily medication for ten years or for the rest of their lives, even that relatively small patient population could represent a substantial market.

A commercially viable pharmaceutical business doesn’t necessarily need tens of millions of customers. Of course, there is an enormous gap between estimating a potential market and generating actual revenue, including the costs of research, clinical trials, manufacturing, and regulatory approval. Still, a startup can focus on a single underserved disease and build an entire research program around it.

For example, Recursion Pharmaceuticals is a US-based biotechnology company developing AI-driven treatments for diseases with significant unmet medical needs. One of its investigational drugs, REC-4881, targets familial adenomatous polyposis, a rare inherited condition affecting an estimated 50,000 people across the US and five major European countries. The company has reported encouraging Phase 1b/2 trial results, including a reduction in abnormal tissue growths among patients.

It is a real challenge to find a molecule that can target a protein associated with a particular disease. This requires screening vast numbers of chemical compounds to identify those that can bind to the target protein, alter its activity, or affect related biological processes. And it’s not just the chemical composition that matters. The three-dimensional structure of each molecule plays a crucial role in determining how these interactions occur.

Testing every possible combination in a physical laboratory is practically impossible. Even with substantial financial resources, conducting that many experiments would take far too long. This is why part of the research process is moving into virtual environments, where computational models and simulations can predict molecular interactions and identify the most promising compounds for further testing. This approach reduces the number of experiments needed in the lab, although the predictions still have to be validated through physical testing.

An entire ecosystem of startups with different business models is emerging around these technologies. Some build molecular simulation infrastructure and make it available to pharmaceutical companies and research teams. Others use existing tools to develop their own drug candidates. It’s an interesting example of how a single technological advancement can create business opportunities at multiple levels of the market. A company can build a commercially successful business without ever bringing its own drug to market, provided its technology proves valuable to other developers.

A great example is Isomorphic Labs, which develops AI technologies for drug discovery and molecular design. The company has established research partnerships with Novartis and Eli Lilly and, in 2025, expanded its collaboration with Novartis to include additional drug discovery programs. It illustrates how the technology used to discover and design molecules can itself serve as the foundation for a business working with major pharmaceutical companies.

Some startups are taking the idea of drug discovery even further. I recently met a team working on technologies related to skin cancer. They are interested in identifying genetic predispositions to the disease and exploring ways to treat it at the DNA level. Rather than developing a drug for a broad group of patients, they are investigating whether treatments could be