Novo Nordisk and Eli Lilly still own obesity by sales, but Q2 2026 made the pipeline behind them look a lot less like a two-horse race. Roche's enicepatide (formerly CT-388) posted −22.7% weight loss in a Phase 2 dose-finding trial — within a whisker of Lilly's approved Zepbound (−22.5%) — and it did so without the tolerability spike that usually shadows the most potent dual agonists. Amgen, AstraZeneca and Pfizer all moved their own candidates forward. The cross-trial map below places the newcomers against the field. One caveat up front, and it matters for everything that follows: these are separate trials at different doses, durations and phases — not head-to-head comparisons.
The one that matters: Roche's enicepatide
Of the four challengers, enicepatide is the only one that changes the competitive picture. It is second only to Lilly's triple agonist retatrutide (−28.3% in Phase 3) on raw efficacy, and its adverse-event discontinuation (~7%) sits in the same band as semaglutide and tirzepatide rather than spiking. Roche has already started the confirmatory Phase 3 programme (ENITH-1 and ENITH-2). Roche also controls a second shot on goal: petrelintide, a once-weekly amylin analog licensed from Zealand (which keeps milestone-and-royalty economics, not the full profit) that lands at a gentler −10.7% but with tolerability close to its own placebo arm — and the two are being combined in a Phase 2 fixed-dose study. For now, that is combination optionality the incumbents do not have in the clinic.
The contrarian bet: Amgen's MariTide
MariTide (maridebart cafraglutide) is mid-pack on efficacy (−19.9% in Phase 2) but interesting for a different reason. It is a GIP receptor antagonist paired with a GLP-1 agonist — the mechanistic opposite of tirzepatide, which agonises GIP. If the Phase 3 MARITIME programme holds up, the real differentiator is convenience: an antibody-peptide conjugate dosed monthly, as few as four to six injections a year against 52 for the weekly incumbents. Its tolerability figure is not dose-comparable and should not be read off the chart directly.
The rest: worth watching, not owning the narrative
AstraZeneca's elecoglipron is an oral GLP-1 at −11.8%, and Pfizer's berobenatide is a monthly injectable at −12.1% (its adverse-event rate, like MariTide's, isn't dose-comparable to the rest of the field) — both earlier, both mid-pack, and neither with a clear efficacy edge. Pfizer is also worth a raised eyebrow here: its previous oral GLP-1, danuglipron, was discontinued in 2025. Convenience — and, for an oral small molecule, a real manufacturing and cost-of-goods edge over the peptides — could still matter commercially, but on the data so far these are watch-list names, not thesis-movers.
The honest counter-argument
Here is what the leaderboard does not say. Phase 2 top-dose numbers reliably erode as trials scale to broader Phase 3 populations, so none of these four is a confirmed −20%-weight-loss drug yet. More importantly, the incumbents' moat was never top-line efficacy alone. It is manufacturing and supply at a scale no challenger can match this decade — though, as Novo's and Lilly's own multi-year shortages showed, scale has been a constraint even for the leaders — plus payer and pharmacy-benefit-manager (PBM) formulary lock-in, and a multi-year head start on completed cardiovascular-outcomes data (SELECT for semaglutide, SURMOUNT-MMO for tirzepatide). Challengers are racing there too — Amgen already has MariTide cardiovascular and heart-failure trials running — but none will own that outcomes data before roughly 2029–2030. Roche, the nearest threat, is still years from a Phase 3 readout. Q2 2026 raised the efficacy ceiling of the category. It did not move the moat.
What to watch
The next real signal is confirmatory Phase 3 data: Roche's ENITH-1/2 for enicepatide, ZUPREME-3/4/5 for petrelintide, and Amgen's MARITIME suite. Until those read out, the challengers are expanding what is possible in obesity — not yet redistributing who profits from it.
A note on the data: this is a cross-trial comparison, not head-to-head. Values are top-line results at differing timepoints, phases and estimands; several curves were digitised from published figures and are approximate; a handful of assets (mazdutide, ecnoglutide) rest on China-only trials and are shown for context. See the obesity drug hub for the underlying per-drug detail.