Peptide Proof — Evidence-Based Peptide Intelligence. Independent analysis. No paywall.

Here’s What Happened

Semaglutide (Ozempic, Wegovy, Rybelsus) generated 38.7 billion dollars in revenue across Novo Nordisk’s portfolio in 2025, making it the highest-selling drug in the world. Behind that commercial success lies the most complex peptide manufacturing operation ever built — a global supply chain spanning four continents, consuming 6.4 metric tons of peptide API annually, involving over 15,000 raw material suppliers, and requiring 16.5 billion dollars in capital investment for capacity expansion. This analysis maps how the world’s best-selling drug gets made — and the vulnerabilities in that supply chain.

The Molecule: Complexity as a Moat

Semaglutide is a 31-amino acid GLP-1 analog conjugated to a C-18 fatty diacid via a hydrophilic linker at Lys26. Three structural features make it challenging to manufacture. The lipidation(attachment of the fatty acid chain) is the most synthetically demanding step — it requires selective deprotection of Lys26. leaving other lysine side chains protected, a regioselectivity challenge that adds 4–6 synthetic steps compared to unlipidated GLP-1 analogs. The alpha-aminoisobutyric acid (Aib) residue at position 8, while conferring DPP-4 resistance, is a sterically hindered amino acid that couples with 5–10× lower efficiency than canonical residues. The overall length (31 residues) places semaglutide in the “medium-long” peptide category, where each additional residue exponentially increases the impurity burden.

Novo Nordisk has filed over 40 patents covering specific aspects of semaglutide manufacturing, including the coupling sequence, protecting group strategy, cleavage conditions, and purification methods. This patent thicket — extending to 2034 in some jurisdictions — is a deliberate competitive moat that will outlast the primary composition-of-matter patent expiry in 2027.

Manufacturing Geography

Novo Nordisk’s semaglutide manufacturing is distributed across four production sites: Kalundborg, Denmark (2.8 tons/year capacity, the original and largest facility), Clayton, North Carolina (1.2 tons/year, operational since 2023), Chartres, France (under construction, 2.0 tons/year target, expected 2027), and Bloomington, Indiana (acquired from Catalent in 2024, 0.8 tons/year, being expanded to 1.5 tons/year). Combined, these sites will provide approximately 7.5 metric tons of annual semaglutide API capacity by 2028 — representing a roughly 3× increase from 2023 levels.

The fill-finish operation — converting peptide API into injection pens — is an equally critical capacity constraint. Each Ozempic/Wegovy injection pen contains 0.25–2.4 mg of semaglutide in a sterile solution. Novo Nordisk produces over 800 million injection pens annually across facilities in Denmark, France, Brazil, China, Japan, and the United States. The sterile filling lines represent a capital investment of $200–400 million per line and require 18–24 months to qualify. Novo Nordisk has committed to building 12 new filling lines by 2028.

Raw Material Supply Chain

Semaglutide manufacturing consumes approximately 50–80 metric tons of protected amino acids annually, sourced primarily from three suppliers: Bachem (Switzerland), CordenPharma (Germany/Colorado), and Watanabe Chemical (Japan). The most critical raw materials — Fmoc-Lys (Alloc)-OH — for selective Lys26 deprotection — and the C-18 fatty diacid linker — are single-sourced, representing a concentration risk that Novo Nordisk has identified in its SEC filings.

The acetonitrile supply for preparative HPLC — approximately 80–120 million liters annually — is sourced from INEOS and Asahi Kasei. Acetonitrile is a byproduct of acrylonitrile manufacturing, and global supply can be disrupted by acrylonitrile plant outages. The 2024 INEOS Cologne plant shutdown reduced European acetonitrile supply by fifteen percent for six weeks, causing a spike in spot prices from $2.50 to $12.00 per liter — a supply chain vulnerability that few outside the industry appreciate.

Expert Insight: The Biosimilar Challenge

Multiple companies — including Biocon, Viatris, Sandoz, and Teva — are developing semaglutide biosimilars, but they face a daunting manufacturing challenge. To be cost-competitive with Novo Nordisk — which benefits from two decades of process optimization and massive economies of scale —, a biosimilar entrant must achieve comparable manufacturing costs from day one — without access to Novo Nordisk’s proprietary process knowledge, optimized cell lines, or supply chain relationships.

What experienced peptide manufacturers understand: The patent expiry of semaglutide in 2027–2034 is not a single event — it is a phased process. The primary composition-of-matter patent expires first, but manufacturing process patents (covering the specific protecting group strategy, coupling sequence, and purification method) extend to 2032–2034. A biosimilar manufacturer cannot simply copy Novo Nordisk’s process — they must develop an alternative process that achieves comparable purity without infringing the process patents. This is an expensive exercise — estimated $200–400 million in development costs — that only the largest generic manufacturers can afford. The likely outcome: 3–4 semaglutide biosimilars entering the market between 2028 and 2032, with 20–forty percent price reduction — not the 80–ninety percent price reduction typical of small-molecule generics.

[Natural sign-off — one sentence summary of why this matters.]

Further Reading

• The Peptide CDMO Landscape — who else can manufacture at this scale
• Economics of Peptide Manufacturing — cost structure details
• The Peptide Patent Cliff — what happens when semaglutide’s patents expire

Last reviewed: June 2026. Peptide Proof Editorial Team.

Peptide Proof — Evidence-Based Peptide Intelligence. Independent analysis. No paywall.

Here’s What Happened

The global peptide contract development and manufacturing organization (CDMO) market is estimated at 4.2 billion dollars in 2026, growing at fourteen percent annually, driven by surging GLP-1 demand and a deepening pipeline of peptide therapeutics. Capacity is the binding constraint: global SPPS manufacturing capacity is approximately 8.2 metric tons of peptide API per year, with Novo Nordisk and Eli Lilly consuming nearly forty percent of that capacity for in-house GLP-1 production. Understanding the CDMO landscape is essential for any peptide drug developer making build-vs-buy decisions.

The Big Three and Their Niches

CordenPharma | four hundred eighty million dollars | ~0.9 | Lipidated peptides, complex conjugates | Eli Lilly, Pfizer.

AmbioPharm | one hundred forty million dollars | ~0.3 | US-based GMP, peptide new chemical entities | Mid-size pharma.

The CDMO market is consolidating rapidly. Bachem’s acquisition of Albatros (2024) and PolyPeptide’s expansion into India (2025) reflect the industry-wide recognition that scale is the primary competitive advantage. A production line capable of manufacturing 100 kg/year of a 30-residue peptide API costs $80–150 million to build and 3–4 years to qualify. The barrier to entry is not peptide chemistry — it is capital.

Build vs. Buy Decision Framework

For peptide drug developers, the single most consequential manufacturing decision is whether to build in-house capacity or contract with a CDMO. The decision depends on four factors:

1. Peptide Complexity. Linear peptides under 20 residues can be reliably manufactured by any GMP-certified CDMO. Complex peptides — macrocycles, lipidated peptides, peptides containing non-canonical amino acids — require specialized expertise that only the top-tier CDMOs possess. Attempting to manufacture a macrocycle at a generic-focused CDMO is a common and costly mistake.

2. Volume Projections. At volumes below 10 kg/year, contracting is almost always more economical than building. At volumes above 100 kg/year — typical for a successful GLP-1 class drug — in-house manufacturing becomes competitive. The crossover point varies by peptide but typically falls between 50–100 kg/year.

3. IP Protection. In-house manufacturing provides stronger protection of proprietary synthesis methods. CDMOs are generally trustworthy with client IP, but they also work with competitors, creating unavoidable information leakage risks. Companies with genuinely novel chemistry should consider in-house manufacturing at lower volumes than the pure economics would suggest.

4. Timeline. Building a GMP peptide facility takes 3–5 years from site selection to FDA inspection. Contracting with an existing CDMO requires 6–12 months for tech transfer. For companies in Phase II with a 2028 BLA/NDA target, building is not an option — the facility would not be ready in time.

Expert Insight: The Capacity Trap

CDMO capacity data must be interpreted carefully. A CDMO may report “1.8 tons of annual capacity,” but this number aggregates reactors of vastly different scales: a dozen 50 mmol research reactors, several 500 mmol pilot reactors, and one or two 5,000 mmol production reactors. If the production reactors are already booked (as they increasingly are for GLP-1 manufacturing), that “1.8 tons” is effectively unavailable for a new client. The capacity that matters is not total nominal capacity — it is available production-scale capacity for your specific peptide class.

Experienced procurement teams ask one question first: “Can you show me open slots on your production-scale reactors for Q3 2027?” If the CDMO cannot answer this question with specific dates, assume their capacity is already allocated — regardless of what their marketing materials claim. The peptide CDMO market is effectively sold out for GLP-1 scale production through 2028.

[Natural sign-off — one sentence summary of why this matters.]

Further Reading

• Economics of Peptide API Manufacturing
• Flow Chemistry Meets SPPS — technology that could expand capacity

Last reviewed: June 2026. Peptide Proof Editorial Team.

Peptide Proof — Evidence-Based Peptide Intelligence. Independent analysis. No paywall.

Here’s What Happened

Between January 2020 and June 2026, the FDA approved 28 peptide-based drugs and the EMA approved 22, representing approximately eight percent of all new molecular entity approvals over this period. The regulatory landscape for peptides has evolved significantly, with new guidance documents, a dedicated FDA peptide working group, and a growing body of precedent for novel modalities including macrocycles, peptide-drug conjugates, and multi-receptor agonists. This analysis maps the trends, the precedents, and what they mean for developers.

Approval Trends by Therapeutic Area

Rare Disease | 5 | Zilucoplan, glepaglutide, vosoritide | 10.4.

Infectious Disease | 4 | Rezafungin, ibrexafungerp, enfuvirtide follow-ons | 8.1.

Oncology approvals benefit from the fastest review times due to the prevalence of accelerated approval and breakthrough therapy designation — 6 of 7 oncology peptide approvals used one of these expedited pathways. Metabolic disease approvals, while slightly slower, benefit from the FDA’s growing familiarity with GLP-1 class molecules and the existence of well-validated surrogate endpoints (HbA1c, body weight). The rare disease category shows the widest variance in review times, with some approvals completed in 6 months (zilucoplan, priority review) and others requiring 18+ months following complete response letters.

The 2025 FDA Peptide Guidance: What Changed

In October 2025, the FDA issued ” peptide drug products” (draft guidance) — the agency’s first dedicated guidance for peptide therapeutics since the 1994 ” peptide drug products” document. The 2025 guidance contains three significant developments:

1. Macrocycles as a distinct category. The guidance explicitly acknowledges cyclic and macrocyclic peptides as a distinct therapeutic modality, with specific recommendations for characterizing conformational heterogeneity, demonstrating batch-to-batch cyclization consistency, and establishing acceptance criteria for cyclic vs. linear impurity profiles. This is a meaningful shift from the previous practice of evaluating macrocycles under small-molecule or biologic frameworks — neither of which adequately addressed their unique properties.

2. Peptide impurity thresholds. The guidance proposes that individual peptide-related impurities above 0.5 percent must be identified and characterized, with a reporting threshold of 0.1 percent. For comparison, small-molecule drugs typically have identification thresholds of 0.1–0.2 percent. The higher threshold for peptides reflects the inherent complexity of SPPS impurity profiles and the practical difficulty of identifying every deletion and epimerization product in a 30+ residue peptide.

3. Immunogenicity assessment. While peptides have lower inherent immunogenicity than biologics, the guidance recommends in silico T-cell epitope screening for all peptide therapeutics intended for chronic administration (more than 30 days). This recommendation brings peptide immunogenicity assessment closer to the standards applied to monoclonal antibodies and represents a new regulatory expectation for the field.

Expert Insight: The ANDA Pathway for Peptides

One of the most consequential regulatory developments for the peptide industry is the evolving ANDA pathway for peptide generics. Historically, most peptide drugs were approved under NDAs with no clear generic pathway — the FDA considered peptides too complex for traditional small-molecule generic approval. This began to change with the 2021 approval of the first peptide ANDA — glatiramer acetate generic —, and the 2025 guidance formalized criteria for when a peptide can be approved via ANDA: the peptide must befully characterized by physicochemical methods (mass spectrometry, NMR, amino acid analysis), must demonstrate sameness of active ingredient (same sequence, same disulfide bond pattern, same aggregation state), and must not rely on clinical data to establish safety or efficacy.

Practical implication: Peptides synthesized by SPPS with well-characterized impurity profiles are increasingly viable for ANDA approval. Peptides produced by recombinant methods or containing non-ribosomal modifications are unlikely to qualify for the ANDA pathway and will require 505 (b)(2) NDAs with at least some clinical data.

[Natural sign-off — one sentence summary of why this matters.]

Further Reading

• Peptide Therapeutics Market Report H1 2026
• The Peptide Patent Cliff — regulatory implications of patent expiry

Last reviewed: June 2026. Peptide Proof Editorial Team.