Pharmaceutical Industry Shifts to Fewer, Larger AI-Driven Drug Development Partnerships

The total value of global joint research and development deals among pharmaceutical and biotech companies reached $86.7 billion in 2025, up 49 percent from a year earlier, according to health care market research firm IQVIA. The average deal size surged 47 percent to about $1.16 billion per agreement, marking a record high. The number of deals, however, has declined over the past five years, highlighting a shift toward fewer but larger, more targeted investments.

AI and machine learning have become core technologies in collaborative drug discovery, driving large-scale, technology-intensive partnerships. Rather than expanding the number of partnerships, companies are concentrating capital on AI-driven drug discovery platforms that, for now, at least, appear to offer a surer chance of a product hitting the market.

Big Tech firms are increasingly forming alliances with pharmaceutical companies. The AI drug discovery market is expected to grow from about $2.9 billion this year to $13.8 billion by 2033, according to market research firm Grand View Research. Nvidia has partnered with Eli Lilly, the world's largest pharmaceutical company, to invest $1 billion in building a next-generation research lab. Google-backed AI drug startup Isomorphic Labs has also signed large-scale joint development deals with Eli Lilly worth $1.7 billion and with Novartis worth $1.2 billion.

Chinese firms are also scaling up aggressively. CSPC Pharmaceutical, which has developed its own AI platform, secured a $5.3 billion investment from AstraZeneca, and drug design company XtalPi has launched a large-scale project worth $6 billion with U.S. information technology firm DoveTree. Korean companies such as JW Pharmaceutical, Daewoong Pharmaceutical and SK Biopharmaceuticals are also entering the race by developing their own platforms and pursuing external collaborations.

The appeal of AI lies in tackling the industry's core bottleneck: time and cost. Developing a single new drug typically takes 10 to 15 years and costs between 1 trillion won to 2 trillion won ($673 million to $1.3 billion). AI, however, can rapidly analyze vast datasets and identify promising candidate materials for new drugs from millions of possibilities with high precision. Insilico Medicine, a leading AI drug discovery company, completed the process from molecular structure design to early validation in just two months. This was about 15 times faster than the conventional process, which usually takes two to three years, according to the Korea Biotechnology Industry Organization.

But transforming AI-discovered candidates into approved drugs remains a major hurdle. While AI excels at identifying promising candidates, proving their effectiveness in the human body and overcoming complex regulatory systems in each country are entirely different challenges, according to an adjunct professor of biomedical science at the Catholic University of Korea. Industry experts say this is driving a new model, pairing Big Tech's algorithms with Big Pharma's capital and clinical expertise. To overcome these hurdles, large-scale collaborations combining the technological capabilities of Big Tech and the capital of Big Pharma are becoming increasingly essential.

In 2020, Britain-based Exscientia spurred expectations by identifying an anticancer candidate using AI, but the project was halted in 2023 during clinical trials and later sold to a competitor.

Pharmaceutical companies invest billions developing breakthrough therapies, yet a staggering percentage of patients prescribed complex treatments never start them, and the majority who do fail to stay on long enough for the drug to work. Three forces are driving this crisis: Complex therapies dominate new approvals. Biologics, injectables, gene therapies, rare disease and other complex treatments are starting to lead new FDA approvals, each requiring intensive, ongoing patient education. Physicians lack time to educate. With 7 to 12 minutes per visit, doctors can barely confirm a diagnosis, let alone walk a patient through a complicated testing and dosing regimen or address injection anxiety. Patients turn to unreliable sources. Overwhelmed patients fill the education gap with ChatGPT, social media, and influencers, often finding misinformation that breeds confusion and hesitation.

Pharmaceutical companies spend a ton of money on developing new drugs. Most fail, but those that win regulatory approval and reach the market can generate billions of dollars in sales, all while patents bar competitors from copying them. Patents last for years, but they eventually expire. At that point, a drug's sales plummet once generic copies become available to patients.