In 2023, Johnson & Johnson exited infectious disease research entirely. Novartis had already left. Sanofi, AstraZeneca, Allergan — gone. Of the world’s twenty largest pharmaceutical companies, exactly three still maintain antibiotic R&D programs: GSK, Otsuka, and Shionogi. The rest decided that antibiotics cannot generate sufficient return on investment. They are correct. The market is structurally broken.
But bacteria do not wait for markets to be fixed. Three people die from drug-resistant infections every minute — 39 million projected deaths between 2025 and 2050. The pipeline that should be producing new weapons against them has shrunk 35% in five years, from 92 projects to 60.
Into this vacuum steps a single organization with an outsized burden: CARB-X — the Combating Antibiotic-Resistant Bacteria Biopharmaceutical Accelerator. On April 8, 2026, it opens the most consequential antibiotic funding round in years. The chemistry that gets seeded in this window will determine what arrives at patients’ bedsides in the 2030s. What does not get built now may never exist.
The Machine
CARB-X was founded in 2016, when the early-stage antibiotic pipeline was stalling and Big Pharma was accelerating its retreat. A decade later, it has become the single largest non-governmental funder of early-stage antibiotic research in the world.
The coalition funding it has expanded steadily. Wellcome renewed its partnership in 2026 with $60 million. BARDA remains the anchor funder. The Gates Foundation, GAMRIF, Novo Nordisk Foundation, and the governments of Italy ($21 million at the 2024 G7), Canada, Germany, and Japan all contribute. In September 2024, CARB-X was named in the UN Political Declaration on Antimicrobial Resistance — the first time a specific funding mechanism was recognized at that level.
There is a paradox here. CARB-X’s indispensability is also the field’s fragility. One program now shoulders an enormous portion of the weight that should be distributed across dozens of funders and hundreds of companies. If CARB-X contracts, so does the pipeline.
The Five Targets: Chemistry as Strategy
The centerpiece of the 2026 funding round is the Novel Chemistry Challenge: new molecular scaffolds with activity against five validated bacterial targets. This is not abstract R&D. Each target was chosen because nature or medicine has already proven it matters — but the chemical diversity exploiting it is dangerously thin. CARB-X is calling for new scaffolds, new chemotypes, new ways of attacking these five pressure points.
Three are clinically validated. Two are preclinically validated. Together, they form a strategic bet on the molecular architecture of the next generation of antibiotics.
Target 1: The Ribosome — Clinically validated
The most targeted structure in antibiotic history. Tetracyclines, macrolides, aminoglycosides, oxazolidinones — all bind the bacterial ribosome. And resistance has eroded every class. The ribosome has approximately 50 known antibiotic binding sites, and most are already exploited. What CARB-X wants: new scaffolds that find unexploited binding sites or engage the ribosome in ways current drugs cannot.
The proof that new ribosomal chemistry can work is already in hand. Lariocidin, discovered in 2025, binds a completely novel ribosomal site via RNA backbone interactions — resistant to point mutations. Cresomycin uses a pre-organized molecular structure to overcome diverse resistance mechanisms. Both are preclinical, but they demonstrate the ribosome still has unexplored territory.
Target 2: Penicillin-Binding Proteins (PBPs) — Clinically validated
Beta-lactams — penicillins, cephalosporins, carbapenems — remain the most prescribed antibiotic class in the world. Their target: PBPs, the enzymes that cross-link peptidoglycan in the bacterial cell wall. The problem is not the target but the counterweapons. NDM, KPC, OXA carbapenemases destroy beta-lactams before they reach PBPs. Current BL/BLI combinations (Zaynich, Emblaveo) are sophisticated stopgaps, but they are still playing defense against an expanding enzyme arsenal.
CARB-X is looking for post-beta-lactam PBP chemistry: molecules that reach PBPs through routes that beta-lactamases cannot intercept. Nature has independently validated the broader cell wall target through convergent phage evolution — three unrelated bacteriophages evolved proteins (Li et al., Nature, Feb 2026) that disable MurJ, the flippase essential for peptidoglycan export. Three separate evolutionary experiments, one conclusion: cell wall assembly is an enduring vulnerability.
Target 3: Type II Topoisomerases — Clinically validated
Fluoroquinolones target DNA gyrase and topoisomerase IV. Ciprofloxacin resistance is now so widespread that it effectively functions as a first-line drug only where local surveillance confirms susceptibility. But the December 2025 approval of zoliflodacin — a spiropyrimidinetrione that binds gyrase at a different site from fluoroquinolones — proved that new topoisomerase chemistry can bypass existing resistance.
The catch: zoliflodacin targets gonorrhea specifically. CARB-X wants scaffolds with broader Gram-negative activity. The gyrB D429N cross-resistance mutation already circulating in N. gonorrhoeae is a reminder that even novel topoisomerase chemistry faces evolutionary pressure from day one.
Target 4: LpxH — Preclinically validated
LpxH is a manganese-dependent phosphoesterase in the lipid A biosynthesis pathway. Inhibiting it triggers a dual kill mechanism: the outer membrane loses integrity and toxic lipid A intermediates accumulate inside the cell, functioning as internal detergents. The enzyme is present in roughly 70% of Gram-negative bacteria, covering Enterobacteriaceae, P. aeruginosa, and A. baumannii — all three WHO critical priority pathogens.
Multiple groups have now produced LpxH inhibitors with whole-cell activity and in vivo efficacy. Duke University’s JH-LPH series achieves >1,000-fold bactericidal reduction against K. pneumoniae. Uppsala’s JEDI compounds (PNAS 2024) show in vivo efficacy in bloodstream infection models. Roche holds 5 patents (2023–2024). A comprehensive Duke review (ACS Infectious Diseases 2025) charts the full progression from AstraZeneca’s first inhibitor to today’s potent leads. No clinical candidates yet — that is exactly the gap CARB-X aims to close.
Target 5: LolCDE — Preclinically validated
LolCDE is the transporter that shuttles lipoproteins from the inner to the outer membrane in Gram-negative bacteria. Gram-positives lack this system entirely, which means a drug targeting LolCDE is inherently Gram-negative selective. The proof-of-concept compound, lolamicin (Nature 2024), demonstrated something remarkable: it killed E. coli, K. pneumoniae, and E. cloacae while sparing the gut microbiome — preventing secondary C. difficile infection in mice.
The limitation: lolamicin has no activity against P. aeruginosa or A. baumannii, and resistance emerged relatively quickly in vitro through LolC/LolE amino acid changes. It remains preclinical. But the principle — a pathogen-selective, microbiome-sparing antibiotic — is precisely what the field has been asking for. A 2025 review in Infection highlights lolamicin as the leading example of “smart” antibacterial therapy. CARB-X wants more robust chemistry around this target: broader spectrum within Gram-negatives, higher barriers to resistance.
The LPS Convergence
Step back from the individual targets and a pattern emerges. Three of the most promising new antibiotic programs in the world — zosurabalpin (Phase 3), BAL2420 (Phase 1), and the LpxH inhibitor programs (preclinical) — all target different steps in the same pathway: lipopolysaccharide biosynthesis and transport.
| Compound | Target | Pathway Step | Stage | Spectrum |
|---|---|---|---|---|
| Zosurabalpin (Roche) | MsbA flippase | LPS flipping | Phase 3 | A. baumannii only |
| BAL2420 (Basilea) | LptA bridge protein | LPS periplasmic transport | Phase 1 | Enterobacteriaceae |
| JH-LPH / JEDI series | LpxH enzyme | Lipid A biosynthesis | Preclinical | Broad Gram-negative |
This is not accidental. For decades, the outer membrane of Gram-negative bacteria was considered impenetrable — the reason no new Gram-negative antibiotic class had been approved since 1968. The LPS pathway is the membrane’s supply chain. Attacking it at multiple points creates the possibility of combination strategies that could devastate Gram-negative pathogens the way multi-drug regimens transformed tuberculosis and HIV treatment. Three entry points into one essential system.
The Forgotten Emergency
Two of CARB-X’s four 2026 funding themes are not about chemistry at all. They focus on neonatal sepsis — prevention and diagnostics. This is deliberate, and the numbers explain why.
Blood culture — the standard diagnostic for sepsis — takes 48 to 72 hours. For a critically ill newborn, that is an eternity. Without pathogen identification, clinicians default to empiric broad-spectrum antibiotics, which drives more resistance, which makes the next infection harder to treat. The cycle is self-reinforcing.
The 2026 AMR Benchmark found zero pediatric antibiotic formulations registered in 17 sub-Saharan African countries. Not reduced access — zero. Only 5 of 35 clinical-stage antibiotic projects globally include any pediatric development. The NeoTest Advanced Market Commitment, modeled on the $1.5 billion pneumococcal AMC that saved an estimated 700,000 lives, is one proposed demand-side solution. CARB-X’s neonatal themes — non-vaccine prevention and diagnostics — target the same gap from the supply side.
This is where the 2026 funding round becomes more than a chemistry exercise. The five molecular targets address the weapons gap. The neonatal themes address the equity gap. Both must be closed simultaneously, or the antibiotics that eventually get built will reach only the patients who already have the most options.
The Translation Gap
Here is the uncomfortable arithmetic. CARB-X has funded 122 projects. Three have reached the market. That is a 2.5% conversion rate.
The valley of death — the span between preclinical proof-of-concept and Phase 1 — is where most antibiotic candidates die, not from scientific failure but from the absence of anyone willing to fund the expensive, unglamorous work of manufacturing scale-up, toxicology studies, and regulatory filings. Even successful CARB-X graduates need advanced development partners. More than ten have secured them. But the corporate graveyard keeps growing.
CARB-X cannot fix the market. It can only keep the pipeline alive until someone does. The UK subscription model, with contracts starting April 1, 2026, is the first permanent demand-side experiment — £5–20 million per drug per year, delinked from sales volume. The EU’s Transferable Exclusivity Voucher, if implemented around 2029, could add more. The PASTEUR Act, introduced four times in the US Congress and never voted on, remains the largest proposed solution and the most politically stalled.
Until these demand-side mechanisms exist at scale, CARB-X’s supply-side work is both essential and insufficient. It is building a bridge to a shore that may not yet exist.
What 11.1 Million Lives Looks Like
CARB-X’s own modeling suggests that regular approval of new, potent Gram-negative antibiotics could avert 11.1 million cumulative deaths over 25 years — the vast majority in low- and middle-income countries where the burden of drug-resistant infection is highest and therapeutic options are fewest.
But there is a 10-to-15-year lag between early discovery and patient access. The compounds seeded in the April 8 funding window will not reach bedsides until the mid-2030s at the earliest. The decisions made this spring — which chemistry to fund, which targets to prioritize, which neonatal approaches to advance — are decisions about who lives and who dies a decade from now.
The pipeline crisis is not about one drug. It is about whether the molecular diversity needed to stay ahead of bacterial evolution exists at all. That diversity must be built now, or it will not arrive in time.
Five targets. Four funding themes. Two application cycles in 2026 — April and Q4. The expressions of interest for the first cycle closed on March 29. The formal application window opens April 8, with public webinars on April 9 and 10.
If even two of the five targets yield clinical candidates within the decade, millions survive who otherwise would not. If none do, the pipeline continues to thin, and the organisms catalogued across these pages grow more dangerous with each passing year.
The chemistry either gets built now, or it never exists.
Previously: The Blind Spot: Why We Can’t Fight What We Can’t See (Post #24). For the pipeline’s novel classes: The New Arsenal. For the market failure: The Broken Economics of Antibiotics. For the global incentive landscape: The Global Experiment.