How Insilico Medicine Uses AI to Speed Up Drug Discovery
Finding a new drug is not easy. Scientists might spend years testing thousands of molecules, only to find that most of them do not work. The entire process can take nearly a decade and cost over $2 billion. The approach taken by Insilico Medicine stands out in the field. With the help of AI, the company developed a drug candidate for pulmonary fibrosis that reached Phase 2 clinical trials, and it did so much faster than the usual drug-discovery process.
A New Drug Candidate for Pulmonary Fibrosis
The drug candidate was designed to help people with pulmonary fibrosis. This is a disease where the tissue in the lungs develops scars. Over time, the lungs become rigid, and breathing becomes harder. Companies such as Insilico Medicine are focusing research in this area.
It was published in Nature Medicine in 2025. In this paper, it states that the candidate entered phase 2 trials within 18 months, far quicker than the typical ten years associated with drug design. This was an important practical challenge for Insilico. It proved that AI could assist in the entire process, from potential target identification through to designing a viable drug molecule, with Insilico Medicine leading the way.
How the Technology Worked
One of the key instruments used in the project was Chemistry42. It allows creating and evaluating potential molecular structures for future drug molecules in a computer model before they are synthesised in a laboratory. This was a critical part of Insilico Medicine’s technological advancements.
Chemistry42 evaluated about 78,000 molecules. Then, 60 of those molecules were picked up by researchers since they seemed to be the most promising ones. As a result, Insilico Medicine could narrow down the scope of candidates and spend less time, money and effort on laboratory work.
However, finding one promising molecule does not mean that the company managed to revolutionise the field of pharmaceutics since one drug does not make a system good enough for use in many research projects and treating many diseases. For this purpose, Insilico Medicine kept working on its drug-discovery platform called Pharma.AI.
Pharma.AI includes several instruments. Chemistry42 creates virtual molecules. PandaOmics collects and analyses vast amounts of information, including clinical trials’ data and research results. Insilico Medicine integrated these tools to help researchers to find the target for the disease and possible solutions and evaluate them.
Insilico has also improved its platform with the PandaClaw instrument. It provides more agent-like qualities for the system. Simply speaking, now the system can perform more complex actions in a row instead of executing one. Scientists still play an important role, but the system can carry out more connected tasks instead of completing only one small instruction at a time. With Insilico Medicine’s upgrades, the pace of research increased.
Taking the System to a Larger Scale
Even with powerful technology, one company can develop only a limited number of medicines by itself. Therefore, Insilico Medicine chose to license its platform to other pharmaceutical researchers.
According to the report in Forbes, at the time of writing, 13 out of the top 20 pharmaceutical companies were already using this technology to conduct their own drug discovery studies. Some of the illnesses being studied included cancer, cardiovascular disorders, and neurodegenerative diseases. A number of leaders in the sector, including Insilico Medicine included, have expanded the technology’s reach to these illnesses too.
This makes the technology more than just an experimental tool for Insilico. Rather, it is a system that other researchers would utilise in a number of different cases, often enabled by Insilico Medicine’s ongoing collaboration and SaaS model.
A Business Lesson Beyond Medicine
Insilico first found a difficult problem that AI was well suited to handle: studying huge amounts of scientific information and exploring many possible molecules. It then built a process that could be used again and again. Notably, Insilico Medicine then turned it into a product for other businesses.
Sharing the platform with pharmaceutical companies, including possible competitors, may appear surprising. However, this strategy allows Insilico to become more than a company developing its own medicines. Over time, Insilico Medicine may become a leading technology provider for the pharmaceutical sector.
The approach is similar to the way companies such as Google and Amazon built digital systems that many other businesses now depend on. Owning important infrastructure can sometimes create more value than competing only as one company within an industry. Insilico Medicine has followed this strategic model in the medical AI space.
The larger lesson is simple. Businesses should look for problems that AI can solve better than older methods. They should then build reliable systems that can solve those problems repeatedly and at scale, which is exactly the direction Insilico Medicine chose to pursue. Such opportunities exist in many industries, not only medicine.
Insilico Medicine’s progress shows that a successful AI project is not just an impressive experiment. Its real value appears when the technology becomes a practical system that people can use to save time, reduce costs and create better products. In drug discovery, that could ultimately mean developing new treatments for patients more quickly.
