Big Pharma poured billions into AI drug discovery. It could soon pay off

The pharmaceutical industry has made massive investments in artificial intelligence over the past several years, but it's yet to see a payoff in terms of a market-ready, AI-developed drug. That could soon change.

The first drug discovered and designed using AI, Insilico Medicine’s Rentosertib, is likely to enter a Phase 3 clinical trial in the next 18 months, the final stage before seeking FDA approval. And it will reach that milestone in record time.

In the meantime, the coming year is sure to see a stepped-up pace of AI investments, deals, and partnerships by Big Pharma, venture capital firms, and biotech startups. Even OpenAI CEO Sam Altman is investing more than $100 million of his own money in AI-based drug startups.

As a result, Precedence Research estimates that the global drug industry’s investment in AI will reach $2.51 billion in 2026 and $16.49 billion by 2034. 

Fading into the background is the skepticism often heard at the start of 2025 about AI’s ability to have a significant impact on drug R&D. “We are really seeing AI being used across the board now,” said Samuel Scarpino, director of AI + life sciences at Northeastern University and an advisor to many biotech startups. “I don’t think we can point to a clear winner yet… but it is definitely producing a lot of value,” he said.

Pharma giant Eli Lilly $LLY is one of the industry’s leaders in adopting AI. In October, the company announced a partnership with Nvidia $NVDA to build a powerful supercomputer to run an "AI factory," powered by more than 1,000 super-powerful chips, which will be used to discover and test millions of potential medicines.

Lilly also just signed a deal with Insilico Medicine, potentially worth $100 million, to use the startup’s AI platform to jointly identify and develop new therapies. Diogo Rau, executive vice president and chief information and digital officer of Lilly. says the company is encouraging every employee to incorporate AI into their work. "I don't believe any other company in our industry is doing what we do at this scale,” he said.

Rau does acknowledge that “there is still a ton of hype out there,” but predicted that an AI-discovered drug could reach the market by 2030. “AI is doing very, very well at discovering new molecules” that can be turned into drugs, he says.

The end of the decade may seem like a lengthy timeline, except that it typically takes 10 to 15 years to develop a new drug, at a cost of $2 billion and up. Just getting to the point where a drug can be tested in humans can take a minimum of 4 to 5 years, and nine out of 10 drug candidates never make it that far.

That’s why there is so much interest in Insilico’s Rentosertib — it went into human trials in less than two years, a blistering pace for the industry.

Insilico, based in Boston and Hong Kong, was founded in 2014 by Alex Zhavoronkov to take advantage of an AI platform he co-developed to replace animal testing. Rentosertib is its first drug candidate, a treatment for idiopathic pulmonary fibrosis, a chronic and deadly lung disease with no known cure that kills about 40,000 people each year in the United States.

In June, the company announced that Rentosertib improved lung function with few side effects in a phase II study of 71 patients, clearing the way to start a larger phase III trial. Though the drug could still fail, “I don’t want to play down this accomplishment,” says Northeastern’s Scarpino. “The typical time frame for drug development has been dramatically shortened. This could be a game changer.” 

Venture capital investors are betting on such an outcome. According to a November Pitchbook report, VCs invested a total of $2.7 billion in AI drug development firms through the first three quarters of 2025, even while overall investment in biotech startups slowed. The funding difference illustrates the “frenzied AI investment environment and investor conviction in the ability of AI integration to drive greater R&D efficiency,” PitchBook said in its report.

Although drug researchers have been using early forms of AI for decades, excitement over the technology’s ability to speed R&D came with the introduction in 2020 of AlphaFold2, an artificial intelligence system developed by Google $GOOGL DeepMind and Isomorphic Labs that made it possible to predict the three dimensional structure of some 200 million biological molecules. It’s creators were awarded the Nobel Prize in Chemistry in 2024, and the model has been used by more than 2 million people from 190 countries.

Lilly’s Rau cautions that AI is a long way from replacing researchers, although it will shift the roles humans will play. And it has a limited role to play in human trials. “We are still limited on how long it takes the biology to work.” he said. “The clinical trials will still take years and years.”

Nevertheless, a recent survey of people working in biopharma by Endpoints Signal, a division of the Financial Times, found that 48% of respondents describe themselves as heavy AI users. And though only 15% of those surveyed said AI is doing transformative work in research and development today, three-quarters expect AI to transform the drug discovery process by 2030; 12% think the change will come as soon as three years.

One of the biggest proponents of AI in drug R&D is Altman; the OpenAI CEO invested $180 million three years ago to seed Retro Bio, which is partnering with OpenAI to develop drugs that can "reverse age-related diseases." Keeping with Silicon Valley's longevity fixation, Retro’s ambitious goal is to add a decade to the lifespan of healthy humans.

Retro is currently in the process of raising a $1 billion round of funding. The company is also about to start the first human trial for an Alzheimer’s drug it developed, in Australia.

OpenAI is also backing Chai Discovery, founded a year ago, that completed a funding round of $130 million in December, doubling its valuation to $1.3 billion. The company said its AI models are already generating drug-like molecules with high success rates, “Looking back over the last five to 10 years, there’s been so much hype on AI for drug discovery,” co-founder and CEO Josh Meier told Bloomberg. “This is the year things started working.”