Beyond the Weekly Shot: The Race to Put GLP-1s in Implants and Patches

Almost every GLP-1 drug on the market is a weekly injection because that is what the chemistry allowed first. The next fight is over the delivery method itself: how to give the same drug without a weekly shot. Two credible attempts moved forward within two weeks of each other in mid-2026.

On July 7, 2026, Vivani Medical said Novo Nordisk had agreed to evaluate NPM-139, a matchstick-sized semaglutide implant designed to dose once or twice a year from a single placement under the skin. On June 24, Anodyne Nanotech raised a $12.6 million Series A to move ANN-101, a once-weekly GLP-1 skin patch with dissolving microneedles, into its first human trial. Both promise to strip the needle, the cold storage, or the weekly ritual out of GLP-1 therapy. Neither has finished a Phase 1 trial. And the last company to sell a GLP-1 implant idea to the FDA, Intarcia, was rejected three times and gave up in 2024 after a clean cardiovascular-safety trial could not offset manufacturing and drug-release problems. The drug is the easy part. The device is where these programs live or die.

GLP-1 drugs are peptides, and peptides are hard to deliver any way except by needle. Semaglutide weighs about 4,113 daltons. Skin passively absorbs molecules under roughly 500 daltons, so semaglutide is eight times too large to cross intact skin on its own. Swallowed, a peptide meets stomach acid and digestive enzymes that treat it as food; oral semaglutide (Rybelsus) gets around this with an absorption enhancer and still lands near 1% bioavailability, which is why the pill needs far more drug per dose than the injection.

A subcutaneous injection sidesteps all of that. It deposits the peptide under the skin, where it enters circulation intact and the fatty-acid chemistry that gives semaglutide and tirzepatide their week-long half-lives can do its job. That is a real achievement, and it is also the ceiling: the shot works, but it is a shot, on a schedule, that has to stay cold. Every alternative delivery route is an attempt to keep the drug working while removing one of those three frictions.

The case for changing how GLP-1s are delivered rests on how badly people stay on them. A Truveta analysis of 125,474 patients found that 64.8% of those without diabetes had stopped their GLP-1 within a year. A Prime Therapeutics study put three-year persistence in obesity without diabetes at 8.1%, meaning roughly one in twelve was still on the drug. Some of that is side effects, some is cost and insurance churn, and some is the plain friction of a weekly self-injection that has to be refrigerated.

Delivery is the lever these companies are pulling on the friction part. A patch removes the needle and the fridge. An implant removes the weekly decision entirely: a clinician places it once, and the patient does nothing for six months or a year. The bet is the same one behind monthly injections, only pushed further. A patient who never has to remember a dose cannot forget one.

Vivani Medical, based in Alameda, California, builds a subdermal implant it calls NanoPortal: a small titanium device, placed under the skin in a quick office procedure, that releases drug at a steady rate with little of the peak-and-trough seen after an injection. Its obesity candidate, NPM-139, holds semaglutide and is designed for once- or twice-yearly dosing. In a rat study reported in August 2025, a single implant maintained about 20% weight loss for more than six months.

The July 7, 2026 announcement with Novo Nordisk is smaller than it sounds and worth reading precisely. Novo agreed only to evaluate NPM-139 internally, and Vivani's release states plainly that the agreement carries no exclusivity over NPM-139 or the NanoPortal platform and discloses no payment. It is an evaluation, and the release says as much. The more concrete milestone came June 25, when an Australian ethics committee cleared SLIM-1, a roughly 10-per-arm, four-week Phase 1 study comparing a low-dose NPM-139 implant against weekly Wegovy at 0.25 mg. Vivani reached the semaglutide implant by way of an earlier exenatide version, NPM-119, whose first-in-human LIBERATE-1 trial reported in August 2025 that the implant released drug without a harmful early burst and caused no gastrointestinal side effects in implant recipients over about nine weeks.

The GLP-1 implant is not a new idea, and its history is a warning. Intarcia Therapeutics spent more than a decade and well over a billion dollars on ITCA-650, an osmotic mini-pump the size of a matchstick that pumped exenatide continuously for up to a year from under the skin. The science mostly worked. Its FREEDOM-CVO trial, in more than 4,000 patients, met the FDA's cardiovascular safety bar in 2016.

It still failed. The FDA issued a complete response letter in 2017 over manufacturing, a second in 2020 after a resubmission, and in September 2023 an advisory committee voted 19 to 0 that the implant's benefits did not outweigh its risks, citing kidney-injury signals and inconsistent drug release. The agency issued a final refusal in August 2024. The lesson for Vivani and anyone following is specific: an implant lives or dies on the device itself. The hard part is making each unit identically and holding a predictable release rate for months, a manufacturing bar well beyond a weekly pen, and the one the category has lost on before.

The patch approach attacks the 500-dalton skin barrier head-on. Instead of asking a large peptide to diffuse through intact skin, a microneedle patch carries the drug on hundreds of sub-millimeter needles that pierce the outer layer and dissolve in the fluid beneath it, delivering the peptide directly to the living tissue below. No syringe, and because the drug rides in solid form, no refrigeration.

Anodyne Nanotech, a Boston company spun out of Tufts, is the newest entrant. Its HeroPatch platform reported in preclinical large-animal work that it delivered GLP-1 with more than 50% bioavailability, reaching exposure equivalent to a 3.6 mg subcutaneous semaglutide dose, with a claimed payload of up to 30 mg per patch. The $12.6 million raised on June 24 is meant to carry ANN-101 into its first human trial; a second Anodyne program pairs GLP-1 with an apelin peptide aimed at preserving muscle during weight loss. Anodyne is not alone: in March 2026 the company formerly known as Vaxess rebranded to Terrestrial Bio and raised $50 million for a competing weekly GLP-1 microarray patch. The cautionary note here has a name too. Anodyne's new board includes Vikram Lamba, former chief executive of Zosano Pharma, a microneedle pioneer whose migraine patch drew an FDA rejection and pushed the company into bankruptcy in 2022.

The needle-free route already on the market is the pill, and it works by dodging the peptide problem rather than solving it. Eli Lilly's orforglipron, approved as Foundayo in April 2026, is a small molecule, not a peptide, so it survives the gut and can be taken any time of day without food or water rules. Novo's oral Wegovy, a high-dose semaglutide tablet, reached US pharmacies in January 2026 and leans on an absorption enhancer to push enough peptide across the gut wall.

Stacked up, the delivery menu now runs from most to least frequent: a daily pill, a weekly injection, a weekly patch, a monthly injection, and a twice-yearly or annual implant. Only the first two rungs are approved. The weekly patch and the long-acting implant are both entering first-in-human testing in 2026, which in practice means the back half of the decade before either could reach a pharmacy or clinic, if they clear the same trials everything else has to. The menu is real; most of it is still on paper.

Nothing changes this year, or likely next. If you are on Wegovy, Zepbound, or Foundayo, the choice in front of you is still a weekly shot or a daily pill. Implants and patches are years of trials away, and the field's own track record says some of these programs will not make it.

There is also a trade-off worth understanding before the marketing arrives. A weekly injection is easy to stop: skip a dose and the drug clears in days. A twice-yearly implant keeps releasing drug for months. If it causes nausea, or you need to come off it for surgery or pregnancy, the device is still under your skin working, and the answer is a second procedure to remove it. Vivani argues its implant can be taken out, which is exactly the feature that matters, but it is a feature that a pen does not need. The convenience of never thinking about a dose is real, and so is the loss of a simple off switch. The best delivery method for a given patient is the one whose trade-offs fit how they actually live, which is not always the one with the fewest doses.

GLP-1 delivery is entering the phase that follows a successful drug class: once the molecule works, the competition moves to how you get it into the body with the least friction. Two 2026 milestones marked the shift, a semaglutide implant Novo Nordisk is now evaluating and a microneedle GLP-1 patch heading into human trials, alongside a pill class that is already reshaping prescriptions.

The near-term reality is modest. Every implant and patch program is preclinical or just entering Phase 1, every efficacy number so far is from animals or company labs, and the one company that got a GLP-1 implant all the way to an FDA decision was rejected on the device, not the drug. That is the sentence to keep in mind as the announcements accelerate. The weekly shot is not going away soon. What is arriving is a real contest over whether the next decade of GLP-1 therapy comes in a pen, a pill, a patch, or something placed under the skin and forgotten for a year.