Weekly brief · Issue 01

The week your contract manufacturer became the critical path

13 August 2026 · covering 3–13 August · 6 minute read

A court told the Pentagon to back off China's biggest CRO. Four days earlier, a clean clinical dossier was rejected because of somebody else's factory. Both belong in the same conversation, and most development plans still don't have one.

The one that matters

On 7 August, a US federal court blocked the Department of Defense from listing WuXi AppTec as a "Chinese military company". If you have work sitting at WuXi, that is a relief. If you read the ruling as an all-clear, you have misread it.

The injunction is preliminary. The court left the door explicitly open for the Pentagon to re-list with better evidence, and the case continues. Nothing about the underlying political direction has changed — only the timetable. A supplier that can be re-listed by executive action in a quarter is not a stable supplier, however good the science is.

Three days later came the other half of the lesson. ITM received a Complete Response Letter for its radiopharmaceutical in gastroenteropancreatic neuroendocrine tumours. Not for efficacy. Not for safety. For deficiencies at a third-party manufacturing site. The clinical package was fine. The programme lost a year anyway.

Two events, one week, same message: the part of your programme you did not build is the part most likely to stop it.

I have watched this from the sponsor side often enough to know how it happens. CMC and supplier qualification get treated as a parallel workstream that reports into the programme rather than as a critical path that can halt it. The clinical team owns the timeline, the technical operations team owns the plant, and nobody owns the sentence "if this site fails its inspection, we lose four quarters."

What to do about it is not complicated, it is just unglamorous:

Put supplier qualification on the same Gantt chart as your pivotal readout, with its own risk budget and its own named owner. If it does not appear next to database lock, it is not being managed.

Write dual sourcing into new contracts now, not at the point where you need it. Technology transfer clauses, defined exit rights and an agreed timeline for a second qualified site cost you very little at signature and are close to unobtainable in a crisis.

Ask your CDMO for its last three inspection outcomes before you ask for its capacity. The second question is the one everyone asks. The first is the one that predicts your approval date.

None of this is new advice. What is new is that it now applies to programmes that thought they were insulated — small sponsors with a single asset, and anyone who chose a partner primarily on price.

Also this week

Regulatory

The FDA published its GDUFA IV commitments — and quietly repriced offshore manufacturing

The commitment letter raises the foreign facility fee differential from US$15,000 to US$25,000 from FY2028, waives fees for new US manufacturing, and introduces four new meeting types (Type 30, 60, 90 and 120) with defined response clocks.

The meeting types are the part worth reading twice. In an agency where meeting requests are being denied outright, a defined clock is worth more than a shorter one.
Regulatory · Neurology

First orexin agonist approved in narcolepsy type 1

Takeda's oveporexton became the first drug approved to address the full symptom range of narcolepsy type 1 — excessive daytime sleepiness and cataplexy together — rather than one at a time.

A first-in-class mechanism carried a bundled endpoint through review. If your programme has a mechanistic story that connects two symptom domains, this approval is now your precedent.
Clinical development

A surrogate endpoint took ninety-two per cent of a company's value with it

Sionna's cystic fibrosis candidate missed on sweat chloride in phase 2a. The programme stopped, the share price fell 92 %.

Surrogate endpoints are cheap to measure and expensive to be wrong about. If a single biomarker is the gate on your whole programme, you have a binary company, not a portfolio — and your investors should be told that in those words.
Safety

Seven deaths reported in Prader-Willi syndrome — raised by clinicians, not by a regulator

Clinicians and patient organisations reported seven deaths and more than a hundred serious adverse events associated with Neurocrine's Vykat XR. The company is standing by the product. No agency action so far.

The signal came from the treating community before it came from pharmacovigilance. In rare disease, the patient organisation often knows before your database does — which is an argument for talking to them regularly, not only at launch.
CRO & service market

Preclinical demand is turning. Consulting services are not.

Charles River reported its best net book-to-bill in nearly four years and raised guidance. In the same week Certara reported services down 3 %, cut about 5 % of staff, and grew only in software.

Read the two together: sponsors are restarting laboratory work but still refusing to buy expensive human advice by the hour. That is exactly the gap fractional and interim arrangements exist to fill.
Pharma business

Jazz paid up to US$1.3 billion for a fifty-five-patient epilepsy programme

Jazz is acquiring Actio Biosciences for US$820 million up front and up to US$500 million in milestones. The lead asset targets KCNT1-related epilepsy, with a trial of roughly fifty-five participants that may be registrational.

Genetically defined neurology at small n is being priced like a platform. If you are sitting on a well-characterised rare neurological population, the constraint on your valuation is the quality of your natural history data, not the size of your trial.

Three more worth a click

Long-term follow-up of tofersen in SOD1-ALS shows stabilisation or improvement in some patients — useful ammunition for anyone arguing a biomarker-based accelerated approval after a missed clinical primary (BioSpace).
Tufts CSDD modelled an AI monitoring agent at up to US$21 million net value per programme — the first economic case for agentic AI in monitoring that is worth arguing with (Applied Clinical Trials).
Fresenius Kabi recalled tocilizumab 400 mg vials over glass particles — check your protocols if you hold it as rescue medication (FDA).

Planning a programme this quarter?

If any of the above touches a decision you are about to make, a short conversation costs you nothing and occasionally saves a year.

Get in touch