Skip to content
Advertisement
Novartis CEO says AI could cut drug development time to eight years

Novartis CEO says AI could cut drug development time to eight years

Vasant Narasimhan argues artificial intelligence could shave four years off the typical 12-year path to market and improve the odds of success

Novartis CEO Vasant Narasimhan thinks artificial intelligence could shrink drug development timelines from about 12 years to eight years. He also thinks it could meaningfully raise the chances that a drug candidate actually makes it to patients.

Narasimhan made the case in an interview with Bloomberg’s Francine Lacqua on “The Pulse.” He said AI is already helping the Swiss drugmaker run clinical trials faster, analyze data and spot promising new drug candidates.

His comments carry extra weight because of a second job. Narasimhan also sits on the board of Anthropic, the AI developer.

Where the four years could come from

The biggest reported gains so far are in early discovery. AI-assisted generative chemistry has reportedly cut the stretch from launching a research program to selecting a drug candidate. That phase used to take three to four years. It now takes around 18 to 24 months.

Clinical operations are getting a similar treatment. AI platforms at Novartis can reportedly run thousands of trial enrollment simulations in minutes. The same work previously took weeks.

Advertisement

By late 2026, around 90% of Novartis development content is reported to be AI-assisted. That shift has reportedly trimmed clinical study report timelines by 25%.

Statistical programming for 2026 trials has reportedly become 90% faster as well. That is the unglamorous code that turns raw trial data into results regulators can review.

The odds problem

Novartis is targeting an increase in the probability of success for first-in-human drugs from 8% to 12%. In other words, the company wants to move from roughly one winner in every twelve or thirteen candidates to about one in eight.

Big pharma meets big AI

Narasimhan’s seat on Anthropic’s board is one of the clearest examples of that overlap at the leadership level.

For Anthropic, having a sitting pharma CEO in the boardroom offers a direct line into one of the most data-heavy, highly regulated industries in the world. For Novartis, it provides a front-row view of where frontier AI models are heading.

The research focus here is less about science fiction and more about plumbing. The targets are information latency and operational bottlenecks: the delays that pile up when data sits waiting to be cleaned, analyzed, written up or approved.

What this means for the industry

There are reasons for caution. Many of the efficiency figures are internal, and some are tied to trials running in 2026. Early operational gains do not guarantee that drugs will clear late-stage trials or regulatory review faster.

Policy is another wildcard. Proposed US restrictions on biotech investments in China could complicate the picture, potentially slowing technological progress in medicine development.

The eight-year target is best read as a direction of travel rather than a finished achievement. Narasimhan said AI could get the industry there, and the early data from Novartis suggests the efficiency gains are real at the operational level.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.