Pharmaceutical giants are employing an array of tactics to minimise the falling revenues that come when patents expire on their top drugs. Among these, experts say artificial intelligence (AI) is playing a growing and potentially revolutionary role.
A wave of expirations, or a ‘patent cliff,’ on pharma’s biggest selling drugs has sped the pace of dealmaking in the past few years. A growing number of deals between pharma and AI companies in recent times are aimed at enhancing internal R&D and easing the pressure to restock drug pipelines.
Go deeper with GlobalData
Discover B2B Marketing That Performs
Combine business intelligence and editorial excellence to reach engaged professionals across 36 leading media platforms.
According to a GlobalData Strategic Intelligence report, the patent cliff between 2024 and 2030 will see the global share of drugs protected under patents fall from 6% to 4%; between 2025 and 2030, an estimated $236bn will be lost in US revenues alone as a result. GlobalData is the parent company of Pharmaceutical Technology.
A similar trend was seen more than a decade ago, when the share of patent-protected drugs dropped from 14% in 2007 to 10% in 2010, and drugs like Boehringer Ingelheim’s Flomax (tamsulosin), Pfizer’s Lipitor (atorvastatin), and MSD’s Cozaar (losartan) lost exclusivity. Despite this precedent, pre-emptive action is difficult thanks to the unpredictable 10–15-year lifecycle of drug development, says GlobalData Healthcare analyst George El-Helou.
Acquisition deals have traditionally plugged the lost revenue of patent expiries, but the transformative impact of AI on the sector has now begun to attract particular interest from big pharma. In October 2025, Eli Lilly announced its landmark AI supercomputer project in collaboration with NVIDIA. In January 2026, AstraZeneca agreed it would acquire Modella AI as GSK acquired AI capabilities from London, UK-based Noetik and San Mateo, California-based Helix.
“[AI is] increasingly seen as a way to strengthen internal R&D productivity,” says El-Helou, part of big pharma’s push to move away from relying on a few blockbusters and towards a steadier flow of differentiated assets.
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalDataAI capabilities come to the forefront
Pharma companies look for several capabilities in AI, says Andrée Bates, PhD, CEO of AI consultancy Eularis. Principally, AI-driven tools are sought to drastically speed up drug development and reduce costs, and Eularis has worked with several high-profile companies to cut drug development limitations to this end, she says.
Bates points to AI biotech In Silico as a prime example of how AI can overturn development in this way. In Silico has reported an average time of 13 months from project initiation to candidate selection in a 2025 Nature Medicine paper published on the company’s AI-developed drug rentosertib. In a Phase IIa idiopathic pulmonary fibrosis (IPF) study (NCT05938920), the drug was found to be safe and indicated positive signals in increasing lung function.
In Silico has partnered with several pharma companies for its AI capabilities, including deals with Qilu Pharmaceutical and Servier announced this year, and a deal with Eli Lilly in November 2025, which is potentially worth $100m. “We are condensing the risk profile. Pharma partners are looking for certainty,” says Alex Zhavoronkov, CEO of In Silico. “This allows us to replenish portfolios before patent expiries begin to impact the bottom line, potentially ending the era of ‘desperation M&A’,” he says.
According to Zhavoronkov, this is emblematic of a changing mindset within pharma. Rather than relying on late-stage M&A which he describes as “a band-aid for revenue,” companies are looking for more targeted deals to acquire reliable, quality assets without taking on the significant overheads that come with whole-company acquisitions.
Beyond this, he says the nature of AI investment is also changing. “Companies like Eli Lilly and Novartis have prepared well because they moved beyond ‘innovation theatre’ and superficial AI pilots,” Zhavoronkov says.
“It’s almost like we’re acting as the external R&D engines to complement [pharma’s] internal R&D engines,” states Brian Alexander, CEO of AI-enabled drug developer Valo Health. More than general interest, Alexander says pharma is engaging in specific, targeted conversations about how AI can enhance their pipelines.
Valo’s own approach centres on AI parsing through data to reveal complex, and often heterogeneous, causes of disease. Valo attracted interest from MSD in November 2025 when the company signed a $3bn deal to find therapeutic targets for Parkinson’s disease.
Pharma’s multi-pronged response to patent expiries
AI may be one of the many ways through which companies are evolving pipeline strategies. The absence of large company acquisitions was widely noted during the recent J.P. Morgan Healthcare Conference 2026 (JPM 2026) in January, traditionally a platform to announce major deals. In El-Helou’s view, 2026 may not see fewer acquisitions, but rather smaller-scale asset-specific acquisitions.
Pfizer, Novartis, and especially Bristol Myers Squibb (BMS) are among the most vulnerable to patent expirations, says El-Helou. BMS is due to lose exclusivity for its most prominent blockbusters, blood thinner Eliquis (apixaban) and cancer immunotherapy Opdivo (nivolumab) during this time.
At the same time, Eli Lilly’s glucagon-like peptide-1 receptor agonist (GLP-1RA) tirzepatide, marketed as Mounjaro and Zepbound, is set to replace MSD’s Keytruda (pembrolizumab) as the world’s top-selling drug by 2030, during which time Lilly’s revenues are projected to increase 105%.
The reshuffled hierarchy in big pharma follows a broader shift in the therapy areas and drug modalities generating the greatest revenues. “You’re seeing a clear shift towards metabolic disorders and peptide-based drugs dominating the future top-sellers, where in the past oncology was the main thing,” says El-Helou.
But though weight loss drug sales are surging, he says oncology is not being abandoned. Instead, major developers increasingly wish to diversify and reduce previous reliance on a few blockbusters, with growing interest in cardio-metabolism, immunology, and other therapy areas.
Big pharma is also not abandoning its traditional strategies to combat patent cliffs, which extend beyond R&D and M&A. At JPM 2026, Bayer noted that expanded US labels for Nubeqa (darolutamide) in prostate cancer and Kerendia (finerenone) in heart failure helped cover lost revenues from patent expiries for blockbusters, but the company was looking for early-stage and preclinical asset deals as its financial health improves.
Secondary patents are also frequently relied upon to extend the lifecycle of branded drugs, says Stephan Neuhaus, PhD, partner and patent litigator at A&O Shearman in Dusseldorf, Germany. For example, while Keytruda’s core chemical patent expires in 2028, secondary patents on formulations for subcutaneous delivery extend beyond 2040.
This series of secondary protections is often termed a ‘patent thicket.’ Though generic manufacturers may enter the market after a core patent expires, Neuhaus says branded developers can “flatten the curve” of falling revenues by allowing their formulations to compete on aspects beyond price, like convenience of administration or breadth of use.
Beyond patents, pharma companies could turn to trademarks to establish brand recognition, which may last for decades, says James Conley, clinical professor at Northwestern University in Evanston, Illinois
AI will be a transformative addition to pharma’s repertoire of patent cliff solutions, says Neuhaus. But AI has its limits here, as many unpredictable variables can scupper pipeline plans even if managed and driven by AI, he notes.
While Bates says AI could render patent cliffs as obsolete—a vestige of the past when drug development took 10–15 years—El-Helou and Neuhaus are more tempered in their expectations. AI may not abolish patent cliffs for good, Neuhaus says, but it could ease falling revenues into gentler slopes.
