In the high-stakes realm of contemporary medicine, discovering a cure ought to be mathematically unfeasible. Researchers calculate there are 10^60 potential chemical compounds—10 raised to the 60th power, a figure larger than the number of stars in the observable universe—yet only a tiny fraction will ever develop into usable drugs. Today, a coming-together of Big Pharma and venture capital funding is wagering billions that AI can at last unravel this boundless complexity.
A front-runner in this movement is , a drug-design firm spun off from parent company Alphabet. As an indication of the market’s hunger for AI-powered biotech, the company closed a huge in March 2025. Joshua Kushner’s Thrive Capital led the funding round, with Google Ventures joining in—this confirms a change in investment thinking: approaching biology less as a game of luck and more as an engineering challenge.
“Nowadays, no one would imagine designing an airplane manually, and you certainly wouldn’t want to fly in one that was,” says Thrive Capital partner Vince Hankes. “Yet every drug we have is created that way.” Looking ahead, he forecasted that all drugs should be designed “using strong software, smart systems, and simulations—exactly how we build airplanes now.”
The catalyst: AlphaFold
The recent surge in funding for this industry comes from the success of AlphaFold 2, an AI system that cracked the “protein folding problem” by forecasting 3D protein shapes from DNA sequences. This milestone—for which Isomorphic founder Demis Hassabis received a —provided the first major evidence that AI could shorten biological research steps that once took years down to just minutes.
The conventional drug discovery process is tough: It usually takes more than 10 years and costs over $2 billion to launch a new drug, and a shocking 90% of candidates fail during clinical trials. For years, chemists have used brute force methods—boiling mixtures and conducting countless lab tests—to find effective compounds; Isomorphic’s chief scientific officer Miles Congreve likens this to playing “Whac-a-Mole.”
To improve these odds, big pharma companies are teaming up with tech firms to tackle “undruggable” diseases. Isomorphic has formed partnerships with and ; the second of these expanded their collaboration in 2025. These joint efforts aim to unlock the secrets of protein mutations common in pancreatic, lung, and colorectal cancers—targets that have long been hard to treat.
The reality check: ‘Wet-lab’ work
Even with all the money and computing power flowing in, moving from algorithmic predictions to actual clinical cures is still risky. As of January 2026, Isomorphic Labs hasn’t yet advanced any drug to the key clinical trial stage, though rivals like Insilico have drugs being tested in China.
Adding AI to physical science isn’t smooth. “It’s a reality check when you get to actual scientific work and hands-on lab experiments,” noted . Even with top-tier software, biology’s unpredictability stays. , who heads biomedical research at Novartis, said that while AI might one day cut five years from the average discovery time, human safety trials can’t be skipped using algorithms.
Looking toward a virtual future
The end goal of this Big Pharma and VC investment isn’t just to speed up the existing system. Hassabis shared that he wants to create a “virtual cell” that can predict how treatments will work before they’re used on patients. The aim is to build a scalable process that produces “dozens of drugs annually,” pushing the industry toward personalized medicine—where someday, patients might get their specific disease’s phenotype tested at a pharmacy and receive tailor-made treatments.
Right now, though, the industry is at a critical point—with billions in funding and advanced code—to show that it can turn biology’s messy randomness into an equation that can be solved reliably every time.
For this article, journalists used generative AI to assist with research. An editor checked the information’s accuracy before it was published.