Biopharma: Unlocking Biopharma’s Next Step: AI as a Productivity Engine

Research evidence excerpt

Biopharma: Unlocking Biopharma’s Next Step: AI as a Productivity Engine

North America InsightMrather than a structural bottleneck. For companies with advanced pipelines and strong

execution capabilities, this framework enhances confidence in both time‑to‑market and

durability of future revenue growth, supporting higher quality earnings visibility over the

medium term.

Exhibit 31: Annual Number of New Drugs Approved by the FDA (2005 - 2025)

NDAs approved Vaccine Approved BLAs approved

Source: Morgan Stanley Research, www.FDA.gov. Note: New drugs approved is broken down into new chemical drugs, new vaccines, and other

new biologics such as antibodies and gene therapies. The data does not include generics, reformulations, or biosimilars.

While headline FDA approval volumes have increased over time, a deeper look at the

system’s conversion efficiency tells a more nuanced story. As illustrated in Exhibit 32 ,

the ratio of NDAs and BLAs approved relative to INDs initiated each year has remained

broadly stable over the past two decades, fluctuating within a relatively narrow band but

showing no clear upward trend. In other words, although the absolute number of shots on

goal (IND filings) has grown alongside approvals, the industry has not materially improved

its success rate in translating early‑stage programs into approved products.

This dynamic suggests that incremental gains in regulatory throughput and scientific

innovation have largely been absorbed by rising biological complexity rather than

improved end‑to‑end efficiency. The industry continues to advance more programs into

development, but those programs are increasingly targeted at harder‑to‑treat diseases,

narrower patient populations, and more complex modalities, limiting improvements in

late‑stage attrition.