Is GLP-1 the New Penicillin? What the Market Shift Means for CDMOs and CROs

Every few decades, a drug arrives that doesn't just only treat a condition but ultimately changes the underlying logic of medicine. Insulin in the 1920s. Penicillin in the 1940s. Statins in the 1980s. Each one revealed something unexpected about human biology, unlocked treatments for conditions that seemed entirely unrelated, and restructured the industry around it.

GLP-1 receptor agonists are making a credible case to join that list.

The comparison to penicillin is bold, and it's worth being rigorous about it. But across the science, the economics, and the supply chain disruption already underway, the parallels are harder to dismiss than the sceptics would like.

‍

A drug class that keeps surprising

GLP-1 agonists were developed to manage blood sugar in type 2 diabetes. That alone would have made them a significant advance. What nobody fully anticipated was the breadth of what followed.

The cardiovascular data arrived first and drew attention across the clinical community. A meta-analysis covering more than 83,000 patients confirmed significant reductions in major adverse cardiovascular events, all-cause mortality, cardiovascular mortality, and kidney disease progression. These weren't modest improvements. They were findings that changed prescribing behaviour and prompted cardiologists and nephrologists, not just endocrinologists, to sit up and take notice.

Then came the addiction data. A study analysing over 500,000 patients with opioid use disorder found that GLP-1 prescriptions were associated with significantly lower rates of opioid overdose and alcohol intoxication. A separate analysis of patients with alcohol use disorder found similar results. There are now more than 15 clinical trials globally exploring GLP-1 use across substance use disorders. The mechanism appears to relate to how GLP-1 receptors interact with brain reward circuits, the same pathways involved in addictive behaviour. Nobody designed these drugs to do this. Researchers are now investigating potential applications in Alzheimer's disease, liver disease, sleep apnoea, and neurodegeneration.

This is precisely what made penicillin historically significant. Not just that it worked, but that it kept revealing new territory. A compound developed for one purpose turned out to be intercepting disease processes across conditions that looked, on the surface, completely unrelated. GLP-1 is following the same pattern.

‍

The acceleration that penicillin never had

Fleming discovered penicillin by accident in 1928. It took until World War II, with government-backed industrial coordination across multiple manufacturers, to produce it at meaningful scale. The journey from discovery to widespread deployment took the better part of two decades. For its era, that was considered remarkable speed.

GLP-1 is operating in a different world entirely, and AI is a significant reason why.

Traditional peptide development once required decades of iterative laboratory work and billions in investment. Machine learning models helped predict semaglutide's pharmacokinetic profile during development, guiding the molecular modifications that extended its half-life and ultimately made weekly dosing viable. Tirzepatide's dual-receptor design, targeting both GLP-1 and GIP, emerged in part from AI-assisted analysis of receptor interactions. Models predicted the synergistic effects before clinical trials confirmed them. The science moved faster because computational tools could explore a solution space that human researchers alone couldn't navigate at the same pace.

The next generation is moving faster still. Researchers have recently used deep learning to design 10,000 novel GLP-1 receptor agonists computationally, screening them for stability, efficacy, and diversity, before identifying candidates with half-lives approximately three times longer than semaglutide. The pipeline now contains over 120 active drug development programmes. Forty-six percent of obesity drugs in development are oral formulations. The therapeutic frontier is being pushed outward at a rate that would have been unrecognisable to the scientists who scaled penicillin in wartime fermentation tanks.

Penicillin had to wait for the right industrial moment. GLP-1 has AI, and that gap in speed has profound implications for everyone who manufactures, develops, and supplies the infrastructure behind these medicines.

‍

The duopoly and its downstream effects

Here the penicillin comparison takes an instructive turn. After World War II, penicillin production was initially led by a handful of US manufacturers, with Pfizer dominant among them. But the manufacturing process, while technically demanding, was broadly replicable. Competition arrived relatively quickly, and the price of penicillin fell from $3,995 a pound in 1945 to $282 a pound by 1950. The market opened up because the barriers to entry, though real, weren't insurmountable.

GLP-1 is following a different script. At the end of Q2 2025, Eli Lilly and Company reported to hold approximately 57% of the global GLP-1 market, with Novo Nordisk holding most of the remainder. Together, they currently control a market projected to grow from $53 billion today to over $157 billion by 2030. Over 60 companies are developing competing drugs and more than 135 candidates are in clinical trials, but meaningful competition at commercial scale remains some way off. For now, two companies are setting the pace for the entire ecosystem.

The trickle-down effects of that concentration are already visible, and for CDMO and CRO leaders, they're impossible to ignore.

Novo Holdings' $16.5 billion acquisition of Catalent in early 2024 was explicitly designed to secure fill-finish manufacturing capacity for Wegovy. The move sent immediate shockwaves through the outsourcing ecosystem. Smaller clients found themselves displaced from manufacturing slots they'd relied on for years. The acquisition reshaped supplier relationships, capacity availability, and strategic priorities across the sector almost overnight. Eli Lilly, meanwhile, committed $27 billion in 2025 to build four new domestic manufacturing sites, and has publicly acknowledged difficulty finding the specialist talent those capital expenditure projects require.

The concentration of manufacturing investment inside two companies' supply chains has created real constraints for the broader market. Peptide synthesis capabilities are being scaled at a pace the industry hasn't seen in a generation, but much of that capacity is locked into serving the two dominant players. For CDMOs and CROs operating outside those relationships, the strategic question is where the next wave of opportunity sits.

The answer is likely in the challengers. With dozens of new GLP-1 programmes advancing through clinical development, many at smaller companies without in-house manufacturing capabilities, the demand for flexible, specialist outsourced partners will grow considerably. The duopoly won't hold indefinitely. The penicillin market opened up. So will this one. But the window between now and meaningful competition is where positioning decisions made today will determine who captures the growth that follows.

‍

What this means for our industry

The penicillin analogy has genuine limits. These drugs are expensive, they require continuous use, and equitable global access remains an unsolved problem that no amount of clinical data resolves on its own. Fleming's discovery ultimately became a global public health tool within a generation. GLP-1's trajectory will depend heavily on pricing, policy, and manufacturing scale in a way that penicillin, developed partly under government coordination, did not.

But the core argument holds. We're looking at a drug class whose therapeutic scope keeps expanding in ways nobody fully predicted, accelerated by AI-driven discovery tools that have fundamentally compressed the timeline from hypothesis to clinical candidate, and currently concentrated in the hands of two companies who are actively reshaping supply chains, talent markets, and outsourcing relationships in real time.

For those of us who build, resource, and support the infrastructure that brings medicines to patients, this isn't a background trend to monitor. It's the central story of the next decade. The organisations that understand that now, and position accordingly, are the ones that will matter when the market opens up.

Vector Talent Insights is a fortnightly newsletter for leaders across the CDMO, CRO, Biotech and pharma services space. We write about the trends, talent pressures, and strategic shifts shaping the sector. Subscribe to stay in the know.

‍