China Now Registers More Clinical Studies Than the U.S. — While Western CROs Write Off $2.7 Billion of Backlog insight cover
Industry NewsIQV · ICLR · MEDP•19 min read

China Now Registers More Clinical Studies Than the U.S. — While Western CROs Write Off $2.7 Billion of Backlog

China's trial engine is now the cheaper, faster option, and Washington's own data says so: HHS found in September 2026 that China registered more clinical studies than the United States for the first time in 2024, and China's slice of industry-sponsored trials rose from 9% in 2018 to roughly 20% in 2023. The service layer is already repricing — Chinese contract research organizations are growing revenue 14% to 46% while their Western peers grow 8% to 9% and have just marked down $2.1 billion of backlog at IQVIA and $558 million at Fortrea. The damage is not to trial volume, which keeps expanding; it is to the Western full-service model, whose contracted backlog is priced off Western sponsor budgets and whose China-origin work now carries Washington-mandated inspection cost and Chinese export-control risk at the same time. The offsets: Beijing's October 2026 export rules protect platforms but exempt licensed products, and firms such as Medpace with little China exposure are still compounding at low-teens growth.

Published Oct 30, 2026Updated Oct 30, 2026

China's share of industry-sponsored trials

9% to ~20%

2018 to 2023, attributed to centralized hospital networks and operational efficiency; Nature, 2026

Registered clinical studies versus the U.S.

More than the U.S.

First time, 2024; HHS memo, September 2026

IQVIA backlog write-down

$2.1bn

Inactive-trial review, Q3 2026, reported Oct 27, 2026

Fortrea backlog write-down

$558m

Inactive studies and scope changes, Q3 2026, reported Oct 29, 2026

HHS's September 2026 memo put a number on what FDA Commissioner Marty Makary warned about in February: the United States \"walked into a mess\" behind China on Phase 1 trials run in 2024. China registered more clinical studies than the U.S. for the first time in 2024, and China's share of global Phase 1 trials passed the U.S. in 2021. This is not a data-quality artifact that a better inspection regime can wave away. It is a cost-and-clock artifact, and it is already reordering who gets paid for running the world's trials.

China's share of industry-sponsored trials

9% to ~20%

2018 to 2023, attributed to centralized hospital networks and operational efficiency; Nature, 2026

Registered clinical studies versus the U.S.

More than the U.S.

First time, 2024; HHS memo, September 2026

IQVIA backlog write-down

$2.1bn

Inactive-trial review, Q3 2026, reported Oct 27, 2026

Fortrea backlog write-down

$558m

Inactive studies and scope changes, Q3 2026, reported Oct 29, 2026

WuXi AppTec continuing-operations growth

+46.3%

Nine months to Sept 30, 2026

Tigermed China revenue growth

+24.0%

H1 2026, reported Aug 28, 2026

The cost curve

China Now Registers More Clinical Studies Than the U.S.

The mechanisms are administrative, not exotic. China's clinical trial regulation requires ethics committees to issue an opinion within 30 days of accepting a submission, allows expedited reviews in 72 hours, and runs a 30-day trial review channel that commits sponsors to start the study within 12 weeks of approval. Rewritten Good Clinical Practice rules took effect on September 1, 2026, aligning every drug trial with ICH E6(R3). The result is a faster, cheaper readout, not a different standard of evidence.

  • Ethics committees must issue an opinion within 30 days of accepting a trial submission; emergency reviews can clear in 72 hours.
  • The 30-day trial review channel for innovative drugs and biologics commits sponsors to start the study within 12 weeks of approval.
  • China's revised Good Clinical Practice rules, effective September 1, 2026, replace the 2020 version and align all drug trials with ICH E6(R3).
  • Phase III multinational protocols filed for priority review must carry clinical centers in at least two countries, China included.
  • Researchers attribute China's rise from 9% of industry-sponsored trials in 2018 to about 20% in 2023 to centralized hospital networks, operational efficiency and regulatory reform.
China's trial system measured against the U.S., from HHS and FDA statements and China's own trial regulations.
MeasureChinaUnited States
Share of global Phase 1 trialsPassed the U.S. in 2021 (HHS memo, September 2026)Fell behind China on trials run in 2024 (FDA Commissioner, February 2026)
All registered clinical studiesMore than the U.S. for the first time in 2024No longer the leader, per the same HHS memo
New drug approvalsNearly a third of new global approvals (Global Data and Morgan Stanley data cited February 2026)Projected at 35% of FDA approvals by 2040 on the same data
Ethics reviewOpinion within 30 days of acceptance; 72 hours in emergencies\"Clunky processes that take too long,\" per the FDA Commissioner, February 2026
Trial start30-day review channel commits sponsors to start within 12 weeksIND clearance process flagged as a bottleneck, February 2026
HHS is now paying to close the same gap from the inside. Its ARPA-H funding of a trial-speed program targetscutting timelines in half and costs tenfold, on the backdrop of a still ten-year, $1-2 billion drug development cycle with failure rates above 90%. The agency's own briefing notes that China and Australia offer speed, lower cost and less regulatory burden than the United States, and that trials in Australia can start six to twelve months earlier for as little as half the cost.

The service layer

The Cost Reset Shows Up First in Who Gets Paid

Reported revenue growth: Chinese CROs versus Western peers

Latest reported periods differ by fiscal calendar; Charles River had not yet reported the September quarter. Fortrea's figure is computed from disclosed revenue of $691.5m versus $701.3m a year earlier.

Unit: % year over year

WuXi AppTec, 9M 2026

Total revenue RMB45.54bn

38.6%

Tigermed, H1 2026

Revenue RMB3,707.6m

14.1%

Medpace, Q3 2026

Revenue $716.0m

8.5%

IQVIA, Q3 2026

Revenue $4,428m

8%

Fortrea, Q3 2026

Revenue $691.5m versus $701.3m

-1.4%

Charles River, Q2 2026

Revenue $1,004.1m

-2.7%

ICON, Q3 2026

Revenue $2,024.8m

-2.9%

Chinese contract research organizations are not nibbling at the growth. WuXi AppTec lifted nine-month revenue 38.6% to RMB45.54bn, grew continuing operations 46.3%, and carried a RMB76.74bn backlog up 36.4% into the fourth quarter — enough to raise full-year guidance to RMB61.0-61.9bn. Backlog grew faster than revenue at every Chinese CRO that disclosed it, which is what a cost curve reset looks like from inside the winners' books.

  • Tigermed posted H1 2026 revenue of RMB3,707.6m, up 14.1%, with Clinical Trial Solutions up 20.7% and China revenue up 24.0% against overseas revenue up just 3.2%.
  • Tigermed added 28 new US FDA investigational new drug projects in H1 2026 and ended the period with 59 ongoing US trial projects, up 45% year over year.
  • Its six co-established clinical sites hold 620 research beds, passed one unannounced FDA inspection with a 100% success rate, and drove site-service orders up 13%, with Phase I accounting for 78% of those orders.
  • WuXi AppTec spent RMB8.0-8.5bn of2026 capex, including a new Changzhou site pulled forward, and guides to RMB14.5-15.5bn of adjusted free cash flow.

The write-down

The First Casualty Is Contracted Backlog, Not Workload

Trial volume is not shrinking; the price of holding it in a Western contract is. In the September quarter, IQVIA cut $2.1bn of contracted backlog after reviewing inactive trials, and Fortrea removed $558m for inactive studies and changed scope. Both said historical results and guidance were unaffected. The real signal is what a backlog is worth when a sponsor knows it can restart the same study at a fraction of the cost in China: contracted backlog stopped behaving like an asset.

September-quarter results for the four listed contract research organizations that disclosed backlog changes.
CompanyLatest revenueBacklog and write-downBook-to-bill
IQVIA$4,428m, +8.0% (Q3 2026)Contracted backlog $32.6bn after a $2.1bn inactive-trial adjustment; $9.4bn converts in the next 12 months1.2x, on $3.2bn of net new bookings
ICON$2,024.8m, -2.9% (Q3 2026)Backlog $24.5bn, up 4.8% sequentially; $639m of cancellations in the quarter1.56x net book-to-bill
Fortrea$691.5m versus $701.3m (Q3 2026)Backlog cut $558m to $7,316m1.12x
Medpace$716.0m, +8.5% (Q3 2026)Backlog $3,167.2m, +5.6%; no write-down disclosedNet new awards $873.4m, +10.6%
Two honest caveats. ICON disposed of Symphony Health Solutions on May 8, 2026, and higher pass-through activity weighed on guidance, so part of its revenue decline is self-inflicted portfolio surgery rather than Chinese competition. And Charles River, whose discovery and safety-assessment work is the earliest layer of the cost curve, posted Q2 2026 revenue of $1,004.1m, down 2.7%, with a trailing-twelve-month net loss of $238.5m. The honest conclusion is narrower and stronger: the service layer's pricing power is eroding fastest where China substitution is cheapest.
  • Medpace is the control case: revenue up 8.5%, net new awards up 10.6%, net margin 19.2% against 16.8%, and guidance of $2.835-2.855bn for 12.0% to 12.8% growth.
  • Its book is weighted to small, sponsor-funded biotech programs with limited China-funded work — the demand least exposed to a global cost curve.
  • That divergence is the warning: a fixed-fee model priced off Western sponsor budgets and Western trial sites is not a neutral toll booth on the way to the patient.

Who captures the value

The Molecule Follows the Data, Not the Sponsor's Passport

The mechanism behind the write-downs is a reallocation, not a collapse. Cheap, fast trials let a Chinese sponsor generate global-grade data on a fraction of the budget, so it can run more shots on goal and license assets outward. Mainland Chinese innovators signed 157 out-licensing deals in 2025 with a potential value of RMB135.7bn and RMB7bn of upfronts; the first half of 2026 added RMB99.7bn of potential value; and NMPA data cited on September 14, 2026 put 2026 out-licensing above RMB120bn with upfronts above RMB10bn. Upfront payments in nine months of 2026 already exceeded the whole of 2025, which is what a repriced cost curve does to deal flow.

Multiples on trailing twelve-month figures at prevailing market prices; growth periods differ by fiscal calendar.
CompanyEV/EBITDA, trailingRevenue growth, latest reportedBalance sheet
WuXi AppTec10.3x+46.3% continuing operations, 9M 2026Net cash; 30.8% return on equity
Tigermed22.5x+14.1%, H1 2026Net cash
IQVIA16.7x+8.0%, Q3 2026Net debt 3.9x EBITDA
ICON14.2x-2.9%, Q3 2026Net debt 2.0x EBITDA
Medpace28.6x+8.5%, Q3 2026Net cash
Fortrea29.7x-1.4%, Q3 2026Net debt 8.6x EBITDA
Charles River94.6x-2.7%, Q2 2026Net debt 12.9x EBITDA; trailing net loss $238.5m
  • Upstream, the volume follows the trial wherever it sits: central laboratory and early-development services at Labcorp and clinical supplies and research services at Thermo Fisher sell against sample and study counts, so they gain the volume even as price follows Chinese budgets.
  • Labcorp carries $14.59bn of trailing revenue at 16.7x trailing EBITDA; Thermo Fisher carries $47.35bn at 26.4x — steadier earnings bases than the fixed-fee clinical layer.
  • Downstream, the sponsors whose budgets get repriced are large: Pfizer spent $11.98bn on research and development in fiscal 2025, and China-rooted BeOne Medicines carried $2.29bn of trailing R&D expense on $6.13bn of trailing revenue.
  • When the cheapest data in the world comes from a Chinese CRO, the service revenue stays in China while the asset revenue moves to the buyer — which is exactly what the out-licensing numbers show.

Policy

Washington Raises the Bar While Beijing Keeps the Licensing Lane Open

Both capitals are now managing the same arbitrage from opposite ends. On September 2, 2026 the FDA said it is expanding oversight of foreign trial sites, increasing the number of bioresearch monitoring inspectors at its China office and elsewhere, adding remote assessments and pre-submission discussions on the provenance of foreign data, and publishing deficiencies more openly. In August, Rep. John Moolenaar and Rep. Ben Cline asked Acting Commissioner Kyle Diamantas to decline China-generated clinical data for US applications unless the FDA had audited the site in person within the previous 12 months, citing three deaths in separate gene therapy trials run in China. Compliance cost is being added to the cheapest place to run a trial.

  • On October 16, 2026, China's commerce and science ministries added biotechnology entries to the export control catalogue, including outright prohibitions on transferring ownership of cell and gene therapy and siRNA platforms.
  • A joint announcement published October 20, 2026 imposes mandatory pre-export registration on in-scope transfers, effective immediately, with the consultation closing November 15, 2026.
  • Ownership-retaining product licenses and outsourced research or manufacturing services that do not transfer platform ownership stay exempt, so the licensing lane that the cheap trial engine feeds remains open.
  • Technologies for discovering and validating wholly novel targets in cell and gene therapy and siRNA face restricted-export treatment with no equivalent single-product exemption.
Read the two policies together and the asymmetry is the whole story. Washington is raising the cost of trusting Chinese data while protecting its own service vendors; Beijing is protecting its platforms from foreign ownership while deliberately letting licensed products travel. The cheap trial engine keeps running, but the revenue it creates accrues to Chinese providers and the assets it produces are sold to Western buyers.

Horizons

What Breaks This Thesis, and When

  • The next two quarters: the FDA's response to the August congressional letter, and the November 15, 2026 close of consultation on the export-control announcement, which sets whether licensing deals move faster or slower.
  • Conversion, not bookings, is the test for IQVIA: only $9.4bn of a $32.6bn backlog is scheduled to convert within a year, against $3.2bn of quarterly net new bookings.
  • For Tigermed, the proof that the arbitrage is monetized offshore is in US project adds and site-service orders, not in reported overseas revenue, which grew just 3.2% in H1 2026.
  • For WuXi AppTec, the risk is political rather than commercial: it was added to the US Defense Department's 1260H list on June 8, 2026, and a district court injunction on August 7, 2026 barred the government from enforcing that designation.
  • For Charles River, the next quarterly print is the earliest read on whether discovery demand follows the same cost curve.
  • Over one to three years, the likely end state is bifurcated: Chinese providers run the majority of global study volume at Chinese price points, while Western sponsors keep only the work their regulators and investors require on Western soil.
  • The bull case for the Western model is that demand is not dead — Medpace proves a focused, biotech-weighted book can still compound at low-teens growth with expanding margins.
  • The bear case is that a sponsor with a credible cheaper restart option negotiates every Western contract from that position, which is how a $2.1bn and a $558m write-down become a pattern rather than an event.
  • The rule change to watch is structural rather than cyclical: if the FDA effectively requires locally audited data, the cheap-trial advantage survives only for assets already running multinational protocols, which favours Chinese CROs with Western execution arms over sponsors themselves.

The first casualty of China's cheap, fast trials is not the trial itself and not the sponsor. It is the Western contract that turns a decade of development into a fixed fee. China's cost curve did not shrink the world's appetite for clinical data; it moved where that data is produced and who is paid for producing it, and the September-quarter write-downs at IQVIA and Fortrea are the first hard evidence thatthe fixed-fee clinical model is losing its pricing power at the same time.

Where the cost curve reset shows up in the share price

2WuXi AppTec2359.HK--
--Vol --
-
Mixed
  • A RMB76.74bn backlog up 36.4% and 46.3% continuing-operations growth pull full-year revenue guidance to RMB61.0-61.9bn.
  • Adding to the US Defense Department's 1260H list on June 8, 2026, and the August 7, 2026 injunction, keeps federal work off the table despite the demand.
  • Over one to three years the cheap-trial engine compounds the order book; over quarters, Washington's posture decides how much of it bills in dollars.
3Tigermed300347.SZ--
--Vol --
-
Bullish
  • Twenty-eight new US investigational new drug projects in H1 2026 and 59 ongoing US trials, up 45%, show the arbitrage being monetized offshore.
  • Six co-established sites with 620 research beds and a 100% pass rate on an unannounced FDA inspection lock in patient access Western sites cannot match.
  • Over quarters the risk is that FDA data-acceptance policy prices China-only studies out of US filings before the US revenue converts.
IIQVIAIQV--
--Vol --
-
Bearish
  • A $2.1bn inactive-trial write-down cut contracted backlog to $32.6bn, of which only $9.4bn converts to revenue within a year.
  • Q3 revenue rose 8.0% at a 1.2x book-to-bill, so the raised $17.35-17.5bn revenue guide depends on price rather than volume.
  • Over one to three years, a credible cheaper restart option keeps Western backlog an option on the future rather than an asset.
IICONICLR--
--Vol --
-
Mixed
  • A record $24.5bn backlog and a 1.56x net book-to-bill sit against revenue down 2.9% and $639m of cancellations, exposing bookings that do not convert.
  • The May 8, 2026 disposal of Symphony Health Solutions explains part of the revenue decline, so the China read-through is weaker here than at peers.
  • Watch direction: cancellations running above the $639m quarterly pace would confirm that backlog quality is deteriorating, not just revenue timing.
MMedpaceMEDP--
--Vol --
-
Bullish
  • Revenue of $716.0m, up 8.5%, net new awards up 10.6% and backlog up 5.6% show the model working when China-funded work is limited.
  • A 19.2% net margin against 16.8% a year earlier, with EBITDA up 16.0%, funds a $2.835-2.855bn revenue guide at 12.0-12.8% growth.
  • Over quarters, a biotech-weighted book is the cleanest listed way to stay on the safe side of the China repricing.
FFortreaFTRE--
--Vol --
-
Bearish
  • Revenue of $691.5m against $701.3m a year earlier and a $558m backlog cut to $7,316m leave the pure-play clinical CRO shrinking.
  • GAAP net income of $0.6m and diluted earnings of $0.01 per share leave almost no cushion in the fixed-fee model.
  • Over quarters, 2026 guidance of $2,665-2,690m revenue and $210-218m adjusted EBITDA assumes clients restart rather than cancel.
CCharles River LaboratoriesCRL--
--Vol --
-
Bearish
  • Q2 2026 revenue of $1,004.1m, down 2.7%, and a trailing net loss of $238.5m show the preclinical layer repricing ahead of clinical.
  • Net debt at 12.9x EBITDA and 94.6x trailing EV/EBITDA leave no room for another cost-curve shock.
  • Over quarters, the next print is the test of whether discovery demand follows the same China cost curve.
LLabcorpLH--
--Vol --
-
Mixed
  • Central laboratory and early-development services sell against trial volume, so China's roughly20% share of industry-sponsored trials lifts sample volume even as price follows Chinese budgets.
  • Trailing revenue of $14.59bn at 16.7x trailing EBITDA gives a steadier earnings base than the fixed-fee clinical layer.
  • Over quarters volume wins; over one to three years the geographic mix of that volume is what moves the margin.