The first post in this series ended with the molecule. Semaglutide represented the current apex of a decades-long engineering effort to take a two-minute gut hormone and turn it into a once-weekly drug — and the structural logic behind that transformation was reasonably well understood. What it did not address is what happened next: two pharmaceutical companies, one receptor pathway, and a clinical data trail that has produced the most effective weight loss drugs ever tested in humans. The competition now is not about whether GLP-1 agonism works. It is about who can extend it furthest, and how.
Two Companies, One Pathway
Novo Nordisk and Eli Lilly arrived at GLP-1 from different directions and with different bets.
Novo Nordisk’s trajectory was more direct. As a company whose core business had been insulin since the 1920s, the incretin pathway was a natural extension of existing metabolic expertise. Liraglutide (2010) and semaglutide (2017) were sequential iterations on the same structural strategy — extend half-life, optimize albumin binding, improve receptor affinity. By the time Wegovy was approved in 2021, Novo Nordisk had more than a decade of GLP-1 clinical experience.
Eli Lilly’s parallel track led somewhere structurally different. Rather than optimizing a single-receptor agonist, Lilly pursued dual agonism — a single molecule that activates both the GLP-1 receptor and the glucose-dependent insulinotropic polypeptide (GIP) receptor simultaneously. The hypothesis was that GIP receptor activation, which had been studied for years without a clear therapeutic application, might be additive or synergistic with GLP-1 signaling in ways that pure GLP-1 agonism could not achieve. Tirzepatide (Mounjaro for diabetes, Zepbound for obesity, approved 2022) was the result: a 39-amino acid peptide with dual GIP/GLP-1 activity, structurally derived from the native GIP sequence with GLP-1 receptor binding engineered in.
These are not two companies selling the same drug at different prices. They are testing two different hypotheses about which receptor combination produces the best metabolic outcome — and the clinical data has, so far, favored the dual agonist.
The Efficacy Progression: What the Data Actually Shows
One of the more striking features of this drug class is how clean the efficacy progression has been. Each generation has meaningfully outperformed the last, and the improvements track directly to the molecular changes driving them.
Exenatide (Byetta, 2005): The proof-of-concept agent. In clinical trials, exenatide produced approximately 3–5% mean body weight loss. Its limitations were pharmacokinetic — twice-daily injection, moderate receptor affinity, half-life of 2.4 hours — rather than mechanistic. The biology worked; the molecule needed refinement.
Liraglutide 3mg (Saxenda, 2014): The SCALE Obesity trial reported approximately 8% mean weight loss at 56 weeks versus placebo. Liraglutide’s albumin-binding strategy extended half-life to ~13 hours and allowed once-daily dosing, improving compliance. The efficacy gain over exenatide was real but moderate — the structural improvements were incremental.
Semaglutide 2.4mg (Wegovy, 2021): The STEP-1 trial reported ~15% mean body weight loss at 68 weeks. This was not a marginal improvement over liraglutide — it roughly doubled the effect size. The structural changes responsible (longer fatty acid linker, DPP-4 resistant position 8 substitution, optimized albumin binding affinity) collectively produced a molecule that spent far more time at therapeutic concentration. Higher effective exposure translated directly to greater efficacy.
Tirzepatide 15mg (Zepbound, 2022): The SURMOUNT-1 trial reported ~20–22% mean body weight loss at 72 weeks — the highest of any approved agent. The additional GIP receptor agonism appears to amplify the GLP-1 effect through complementary mechanisms: GIP signaling enhances insulin secretion from a different pathway, and there is evidence that combined GIP/GLP-1 activation produces greater suppression of food intake than either receptor alone. The precise mechanism of the synergy is still being characterized, but the clinical signal is unambiguous.
The trajectory across these four data points — 4%, 8%, 15%, 21% — is not noise. Each jump corresponds to a specific structural or mechanistic advance.
The Oral Delivery Problem
Injectable GLP-1 analogs work. The engineering question that has occupied both companies since semaglutide’s approval is whether the same efficacy is achievable in a pill — and the answer is: partially, expensively, and not yet with small molecules.
The reason injections exist is not historical inertia. GLP-1 analogs are peptides. Oral administration exposes them to the proteolytic environment of the GI tract, which degrades them before they reach systemic circulation. The half-life engineering described in the previous post solved the problem of circulation stability; it did nothing for intestinal stability or transcellular absorption across gut epithelium.
Oral semaglutide (Rybelsus, approved 2019 for type 2 diabetes) is the first solution to this problem, and it is instructive about how hard the problem is. Rybelsus is co-formulated with SNAC — sodium N-[8-(2-hydroxybenzoyl)amino]caprylate — an absorption enhancer that works through a specific local mechanism: SNAC buffers the gastric pH in the immediate vicinity of the dissolving tablet, transiently reducing protease activity and facilitating transcellular (rather than paracellular) absorption across the gastric mucosa. The semaglutide is absorbed through the stomach wall, largely bypassing the intestinal proteolytic environment.
The limitation is efficiency. Oral bioavailability of semaglutide via SNAC is approximately 1%. To achieve therapeutic plasma concentrations comparable to the 1mg injectable dose, the oral formulation requires 14mg — roughly 14 times the injected mass of drug. This drives up manufacturing cost, requires strict fasted administration (food dramatically reduces absorption), and sets a ceiling on how high the oral dose can practically go.
Why pursue oral delivery at all under these constraints? Because needle aversion is real, because access and stigma around injectable medications meaningfully affect uptake, and because a pill format expands the treatable population substantially. The SNAC approach is not a full solution — it is a workable one, with known tradeoffs.
Pipeline: Small Molecules and Triple Agonists
The next competitive frontier is being fought on two fronts simultaneously: eliminating the peptide delivery problem entirely, and adding receptor targets beyond GIP and GLP-1.
Orforglipron (Eli Lilly): The most structurally significant pipeline candidate. Orforglipron is a non-peptide small molecule GLP-1 receptor agonist — it binds the GLP-1 receptor at an allosteric site inside the transmembrane domain rather than the extracellular peptide-binding site. Because it is not a peptide, it is orally bioavailable through conventional mechanisms, requires no SNAC co-formulation, and can be taken without fasting restrictions. Phase 2 data showed weight loss of approximately 9–15% depending on dose, competitive with liraglutide and approaching lower-dose semaglutide. Phase 3 results are expected 2025–2026. If the efficacy holds, orforglipron changes the delivery calculus entirely.
Retatrutide (Eli Lilly): A triple agonist — GLP-1, GIP, and glucagon receptors simultaneously. Phase 2 data published in 2023 showed approximately 24% mean body weight loss at 48 weeks, the highest of any agent reported in clinical trials to date. The glucagon receptor addition is mechanistically interesting: glucagon signaling increases hepatic glucose output but also increases energy expenditure and fatty acid oxidation. In the context of a drug that is already strongly suppressing appetite and food intake, adding glucagon receptor agonism appears to increase the metabolic rate component of the energy balance equation rather than simply reducing input further.
CagriSema (Novo Nordisk): Novo Nordisk’s pipeline response combines semaglutide with cagrilintide, a long-acting amylin analog. Amylin is a pancreatic peptide co-secreted with insulin that independently suppresses appetite and slows gastric emptying through receptors distinct from the GLP-1 pathway. Phase 2 data showed weight loss of approximately 15% — comparable to semaglutide alone at similar timepoints, with Phase 3 ongoing.
The pattern across these pipeline candidates is consistent: each new entry is adding receptor targets or eliminating delivery barriers, not simply re-optimizing the existing semaglutide scaffold. The ceiling on achievable weight loss through this class is not yet visible.
Where the Data Leaves Us
The efficacy progression of GLP-1-based therapies is unusually clean for a drug class in active development. Each approved generation has outperformed the last by a margin that exceeds noise, and the pipeline data suggests the trend continues. The mechanistic logic has held: more receptor coverage, better pharmacokinetics, and complementary signaling pathways each add measurable clinical effect.
The unresolved questions are not primarily molecular. Long-term cardiovascular outcome data is still accumulating — early signals from the SELECT trial (semaglutide) are positive, but the dataset for newer agents is thin. Lean mass loss alongside fat loss remains a concern, particularly at higher doses and longer durations. And the cost and access picture is the frontier that molecular engineering cannot touch.
The drugs are, by the standards of their predecessors, remarkably effective. What happens next depends less on the chemistry than on everything surrounding it.