vs active biologic clinical trials, FDA Biologics License Applications (BLAs) and Chinese registered trials, analyzed structurally: how technology platforms are shifting, where targets are crowded, how the US–China landscape flipped, a fact hidden behind approval counts, and what this dataset can't see.
Scope: interventional trials started on or after January 1, 2015, where the intervention is a biologic or its name contains a biologic marker (antibody, ADC, CAR-T, recombinant protein, etc.).
Annual FDA biologics approvals jumped from 29 to 43 in 2024, a record high. Splitting approvals by whether the active ingredient is new makes it clear where the increase came from.
The 2024–2025 surge in approvals came almost entirely from biosimilars and follow-on biologics — novel BLAs have held between 15 and 19 a year for a decade.
Follow-on biologics jumped from 10 in 2023 to 24 in 2024 and 22 in 2025. Over the same three years, novel BLAs were 19, 19 and 17, and Type 1 NMEs were 17, 17 and 13. The priority review share fell at the same time, from a historical ~50% to 20.9% in 2024, which fits the view that the increase came from biosimilars on the standard review track. The sponsors with the most approvals in 2024–25 were Accord BioPharma, Samsung Bioepis, Celltrion and Biocon, all of them biosimilar makers.
Growth itself doesn't signal an inflection; the change in growth does. For each track we use three-year rolling windows. Growth compares the latest three years (2023–25) with the middle three (2020–22). Prior growth compares the middle three with the earliest three (2017–19). Acceleration is the difference between the two. The x-axis shows whether a track is still rising, and the y-axis shows whether that rise is strengthening or fading —The bottom-right quadrant is the inflection warning zone: still growing, but momentum has turned down。
The entire ADC complex has entered the inflection warning zone at once:ADC modality, HER2 and TROP2 are three independent tracks, all still growing fast, but acceleration has turned negative for every one。
ADC grew +72% over the past three years (deceleration -9.0pp), HER2 +81% (-12.3pp), TROP2 +115% (-72.3pp). These three use independent counting methods -- one by modality keyword, two by target keyword -- yet point the same direction, so this looks unlikely to be an artifact of a single classification rule. By contrast, bispecific/multispecific (+123%, acceleration +67.1pp) and CD3 (+115%, +134.1pp) are still accelerating, and these two also corroborate each other: CD3 is the shared arm of T-cell engagers, so the two tracks should naturally move in sync. For betting on "the next ADC," these two signals lead ADC itself.
Note: China also falls in the warning zone (+34%, -36.8pp) -- still the fastest-growing region, but its expansion rate has clearly slowed from 2020-22. This does not contradict Section 06's finding that "China overtook the US for the first time in 2025": the overtake is about level, while the deceleration is about rate of change.
Annual new biologics trial starts fell after peaking in 2021 (the COVID vaccine surge), rebounding to 1,698 in 2025. The overall focus remains on early-stage work: Phase I and Phase I/II combined account for 42%.
Modalities are classified by regex matching on intervention names and titles. The four main modalities have taken very different paths: ADCs grew fourfold, cell therapy keeps climbing, mAbs peaked and declined, and vaccines went through one full COVID boom and bust.
Using 32 target-keyword groups (generic-name stems + target abbreviations) to match intervention names and titles, 5,253 trials were matched. PD-1 leads by far with 2,115 trials, but its annual curve already peaked back in 2019.
In 2015, the number of biologics trials involving China was 19% of the US figure. China has risen steadily since then, while the US has declined year by year after peaking in 2020; the two were nearly even in 2024 (534 : 541), and in 2025 China overtook the US for the first time (640 : 526).
China data in the previous six sections comes from the site-location field of the international registry, so it covers only trials that chose to register there. This section adds China's domestic registry, which includes many investigator-initiated trials that never appear in the international registry.Note: this is clinical study registration data and does not include NMPA approval decisions, and cannot be used to infer China's approval counts or approval trends.
The previous seven sections all measure Clinical stage activity: a molecule only shows up in this dataset once it is registered for a trial. So the force most likely to rewrite every curve above within five years, AI-driven drug design, doesn't register at all in the inflection signals. This section charts that blind spot.
The twelve AI drug-discovery companies have 43 trials combined across the full international registry; the number within this page's biologics scope is only 2 trials.Isomorphic Labs and Xaira: 0 each。
These gaps aren't a scraping error. They come from two structural causes. First, the two leading companies are still preclinical: Isomorphic expects its first candidate to enter the clinic in late 2026, and Xaira was only founded in 2024, so neither has any registered trials. Second, the companies that do have trials (Insilico Medicine with 11, Recursion with 9, Exscientia with 7) focus on small molecules. This page covers only biologics, so they are excluded by definition.
Generate Biomedicines is the only company that is both "AI-driven" and focused on biologics. It has 6 trials in the full registry, and 2 of them fall into this page's sample (GB-0895 for asthma, and GB-4362 with enfortumab vedotin and pembrolizumab for advanced urothelial carcinoma, both Phase I). In other words, on this trials, AI-designed biologics account for just two dots.
The evidence from the previous eight sections, condensed into two numbers:Industry band (whether fundamentals are improving or deteriorating) and Market band (how much is already priced in). Together the two form quadrants. The real warning sign isn't weakness on both, but Market clearly stronger than fundamentals cell. The price range is daily bars from 2026-04 to 2026-10.