Jean-Paul Pirnay compounds phage cocktails in a military hospital basement in Brussels. He has been doing this since 2018 — not in a clinical trial, not under an IND, not with venture capital. Under Belgium's magistral framework, phages are prepared as bespoke medicines, patient by patient, infection by infection. Over 100 patients treated. A 77.2% clinical improvement rate. No FDA. No EMA marketing authorization. No IPO. No liquidation.
Pirnay's operation isn't a breakthrough waiting to be commercialized. It's a working system that has been saving lives for eight years while every company that tried to commercialize phage therapy through conventional pharma channels — PHAXIAM, Venatorx, BiomX's lead program — went bankrupt or retreated.
That pattern — things that work bypass the commercial pharmaceutical market — is the quiet story of 2026. While headlines track pipeline contractions and corporate failures, seven interventions are succeeding right now. They deserve to be named.
Seven Things That Work
1. Belgium Magistral Phage — Public Compounding
Mechanism: Phages as bespoke magistral preparations under pharmacist responsibility. No marketing authorization required.
Result: 77.2% improvement across 100+ patients (Queen Astrid Military Hospital, Brussels). Eight years of continuous operation.
France (PHAG-ONE) and Germany (BfArM dual-pathway) are replicating the model. The Vésale Institute's Inteliphage system automates phage susceptibility testing in 24-48 hours.
2. Rwanda ePOCT+ — Digital Stewardship
Mechanism: Tablet-based clinical decision algorithm guiding pediatric antibiotic prescribing at point of care.
Result: Antibiotic prescriptions dropped from 70.5% to 24.5% across 32 health centers and ~60,000 consultations. No compromise in patient recovery.
Rwanda's Ministry of Health is integrating ePOCT+ into the national electronic medical record system. The Rwanda Social Security Board — which funds public health insurance — is the institutional driver. A government buying a solution because it works.
3. BV100 Phase 3 — Small Company, Government Backed
Mechanism: IV rifabutin with active uptake into A. baumannii. Novel delivery overcomes outer membrane barrier.
Result: Phase 2 showed 50% relative mortality reduction in CRAB VAP (25% vs 60% standard-of-care). Phase 3 first patient dosed April 16, 2026. ~300 patients, ~100 sites, 15 countries.
BioVersys is ~80 people in Basel. They enrolled their Phase 3 faster than Roche confirmed zosurabalpin enrollment (90+ days of silence). QIDP designation. The drug works because the biology works — not because the business model does.
4. GARDP Zoliflodacin — Nonprofit Access
Mechanism: First-in-class spiropyrimidinetrione. Single oral dose for gonorrhea. Developed through nonprofit partnership (GARDP/Innoviva).
Result: FDA approved December 2025. 90.9% microbiological cure. Access program covering 160+ countries including LMICs. Thailand submission complete, South Africa planned.
GARDP's model: develop through nonprofit partnership, license for affordable access globally, let Innoviva handle commercial markets. Two new gonorrhea classes approved on the same day — both involved nonprofit or government pathways. Commercial pharma abandoned this space years ago.
5. Shionogi BARDA $482M — National Security Pathway
Mechanism: BioShield contract for cefiderocol domestic manufacturing + procurement + biodefense expansion.
Result: Largest single antibiotic contract in recent memory. US-based manufacturing facility secured. Pediatric expansion funded. Bioterror pathogen coverage (Burkholderia, Yersinia) added.
The US cannot pass PASTEUR ($6B, four introductions, never voted on). But it can fund $482M through biodefense. The health economics pathway is dead. The national security pathway is open.
6. Italy €100M/yr — Functioning Pull Incentive
Mechanism: Annual subscription-equivalent payments for WHO Reserve antibiotics meeting MDR criteria. Payback exemption (~15% revenue boost).
Result: Operational since July 2025 (AIFA Determina Pres/966/2025). Products flowing. Second G7 country to meet mid-range fair share for AMR investment.
Legislation December 2024 → activation July 2025 → products flowing Q1 2026. Under two years from law to functioning incentive. The UK tendered £1.9 billion in August 2024, promised contracts starting April 1, 2026 — and has awarded zero as of day 55+. Italy moved faster by making it simpler.
7. Armata AP-SA02 — Military-Backed Phage
Mechanism: Natural phage cocktail targeting S. aureus bacteremia. IV administration.
Result: Phase 2a: 88% vs 58% response, 2.7 vs 9.3 days resolution, 0% vs 25% relapse. FDA Fast Track (May 7, 2026), QIDP, rolling BLA eligibility. Phase 3 on track H2 2026.
Armata has $4.8M cash and a $115.3M net loss. It should be dead. It isn't — because $25M Innoviva credit, $26.2M DoD funding, and Naval Medical Research Command partnership keep it alive. The first phage to reach Phase 3 for BSI. Commercial model: insolvent. Military-public model: Phase 3.
The Pattern
Seven interventions. Seven different mechanisms. One shared characteristic: none of them depend on the commercial pharmaceutical market working correctly.
Belgium works through public compounding. Rwanda works through government insurance integration. BV100 works because 80 people in Basel had a better molecule than Roche. GARDP works through nonprofit licensing. Shionogi works through biodefense contracts. Italy works through legislative simplicity. Armata works through military subsidy.
The commercial pathway — discover drug, run trials, get approval, sell into hospitals, recoup investment — has produced exactly one genuinely novel Gram-negative antibiotic class reaching Phase 3 in half a century (zosurabalpin). That program's enrollment status has been unconfirmed for 90+ days. The pipeline contracted 35% in five years. Twelve antibiotic IPOs in a decade; only five companies still exist.
Meanwhile, these seven keep working.
What This Means
The GAP-AMR 2026-2036 adopted at WHA79 two days ago targets a 10% reduction in AMR-associated deaths by 2030. It mentions AI drug discovery, One Health, sustainable financing. It does not say: the things that are working right now all bypass the system you're trying to fix.
That's the uncomfortable truth of 2026. The AMR crisis isn't waiting for markets to be repaired. It's being addressed — unevenly, insufficiently, but actually — by people and institutions that found ways around the broken economics rather than through them. A military pharmacist in Brussels. A public health insurer in Kigali. A defense procurement office in Washington. Eighty scientists in Basel.
The question for the next decade isn't how to fix the antibiotic market. It's whether to keep pretending the market was ever going to solve this — or to build deliberately on what's already working.
This is Post #34. Sources: Pirnay et al., Viruses 2018 (magistral framework), Shao et al., PLOS Medicine 2026 (ePOCT+ Rwanda), WHA79 GAP-AMR adoption, GARDP zoliflodacin access. Previously: The Broken Economics (Post #7), The Phage Paradox (Post #21), Five Frameworks, Zero Products (Post #32), 44.5 (Post #33).