Ozempic, Wegovy, Mounjaro, Zepbound. Everyone knows the drugs. Most investors know the pharma companies that sell them: Novo Nordisk, Eli Lilly. A smaller group knows the CDMOs that manufacture the active ingredients: Bachem, PolyPeptide, CordenPharma.
Almost nobody knows the other three layers of the supply chain. Each layer is an independent bottleneck with multi-year qualification timelines. And right now, all four are binding simultaneously.
The GLP-1 supply chain is not one bottleneck. It is four. Stacked. Each one capable of throttling the entire market.
Layer 1: Peptide CDMO Capacity
The drug substance itself. GLP-1s are peptides — chains of 30+ amino acids synthesised through solid-phase chemistry at metric-ton scale. Before 2022, this was research-grade production. The world woke up and realised nobody had built industrial-scale capacity.
Bachem, PolyPeptide, and CordenPharma form the Western core. Combined capacity expansions exceed $5 billion between 2024 and 2027. CordenPharma alone committed EUR 900 million for 30,000 litres of capacity. All operate near full utilisation with backlogs exceeding available supply.
New entrants need 3-5 years for GMP certification. The barrier is not capital. It is time.
Layer 2: Glass Containment and Elastomer Components
Every injectable GLP-1 needs a physical container: glass vials, cartridges, pre-filled syringes, rubber stoppers, plunger assemblies. These components are not commodities you can switch overnight.
West Pharmaceutical Services (WST) has dominated elastomer components for 90+ years with 30,000+ component variants. Stevanato Group (STVN) makes the glass packaging. Regulatory qualification for a new supplier takes 3 to 5 years — the same timeline as the CDMO layer.
If both Layer 1 and Layer 2 constrain simultaneously, neither additional API production nor additional device assembly can relieve the shortage. The binding constraint is whichever layer has the longer lead time.
Layer 3: Autoinjector and Pen Device Assembly
The physical device that delivers the drug. The global autoinjector market reached $107 billion in 2025 with projections of $300 billion by 2031. Becton Dickinson has secured long-term contracts including two next-generation GLP-1 programmes with major pharma.
This layer serves all injectable biologics across rheumatology, oncology, and autoimmune indications, not only GLP-1s. That diversification makes it the most resilient layer — and the longest-duration investment thesis in the stack.
Layer 4: Oligonucleotide CDMO Capacity
Oligonucleotides — ASOs, siRNAs, gRNAs — share the same solid-phase synthesis chemistry as peptides. The industry calls this TIDES for a reason. The oligo CDMO market sits at $2.55 billion in 2024, growing at 22% CAGR toward $6.73 billion by 2029.
Bachem holds a strategic oligo manufacturing collaboration with Eli Lilly since 2022. This layer is earlier in its lifecycle — lower catalyst density, same structural physics.
The Paired-Position Framework
These four layers have different falsification profiles. The CDMO layer faces direct risk from oral non-peptide GLP-1s like orforglipron (recently approved). The device and containment layers do not, because injectable volume grows in absolute terms regardless of oral market share — obesity penetration is still below 5%.
This creates a natural paired-position framework: higher-risk CDMO exposure hedged by longer-duration device and containment exposure within the same end market.
The generic semaglutide wave that began in March 2026 (patent expired, 50+ manufacturers launched at 80% price reduction) is the stress test. It reveals which layers bind first under volume pressure. Early data suggests all four.