The sticker

A thin patch goes on your forearm before a workout. Inside it, a network of microfluidic channels, each narrower than a human hair, collects your sweat as it emerges from your pores, routes it through chambers filled with chemical reagents, and changes color depending on how much sodium you’re losing. You snap a photo with your phone. An app tells you what to drink.

This is the Gx Sweat Patch, made by Epicore Biosystems, a Northwestern University spinoff backed by $32 million in venture capital and a global partnership with PepsiCo’s Gatorade brand. It launched in 2021. The US Air Force, Navy, and Army have contracts for it. Chevron, United Airlines, and Bechtel use a version called Connected Hydration to monitor dehydration in workers on oil rigs and tarmacs. Epicore’s Discovery Patch is FDA Class I-registered for clinical research.

But the interesting part isn’t what the patch already does. It’s what the patent landscape says the same underlying technology is about to do.

219 patents and counting

The US Patent and Trademark Office has granted 219 utility patents for devices that combine sweat collection with biosensing — a category that barely existed before 2010. In 2005, there was one. By 2017, the count hit 11 per year. In 2024 and 2025, it reached 23 per year. More than half of all sweat-sensor patents ever issued were granted in the last five years.

The assignee list tells the story of who saw it first. The University of Cincinnati holds at least 16 patents. Northwestern University has five. Philips has seven. Samsung, Nokia, and Nippon Telegraph and Telephone each have filings. A startup called CoreSyte has nine. And Mercedes-Benz Group AG has one that we’ll get to in a moment.

The bottleneck that broke

Sweat is a surprisingly rich clinical fluid. It carries glucose, cortisol, lactate, uric acid, amino acids, electrolytes, and even drug metabolites. On paper, the biomarker list rivals blood. The problem was practical: you don’t sweat enough.

Not continuously. Not in controlled amounts. Not when sitting in a clinic. A jogger’s forearm produces plenty. A resting patient in an air-conditioned hospital produces almost none. Without a reliable supply of sweat, the sensor is useless.

Jason Heikenfeld’s lab at the University of Cincinnati spent over a decade on this problem. The solution was iontophoresis: a current of 0.2 milliamps drives a chemical stimulant (carbachol, pilocarpine, or methacholine) into the upper layer of skin, locally activating sweat glands to produce a controlled flow of at least 0.1 nanoliters per minute per gland. The technique works for up to six hours from a single dose, even when the wearer is cool, dry, and sitting still. Cincinnati’s most recent patent on the method, granted in April 2025 (US 12,262,990), claims a protocol for delivering multiple doses throughout the day to sustain continuous sweat production.

That was the lock. Once you could produce sweat on demand, everything else was engineering.

Three labs, three pieces of the puzzle

Cincinnati solved how to get the sweat. Northwestern solved how to move it.

John Rogers, a materials scientist at Northwestern’s McCormick School of Engineering, built what became the field’s reference architecture for soft, skin-mounted microfluidics. His 2016 paper in Science Translational Medicine, cited over 1,250 times, described a device that bonded conformally to skin, routed sweat through branching capillary-burst valves into sequential reservoirs, and measured chloride, lactate, glucose, and pH through colorimetric assays visible to a phone camera. No electronics on the patch. No battery. Just engineered fluid mechanics and chemistry.

Rogers founded Epicore Biosystems in 2017 to commercialize the work. In August 2025, his team and Northwestern clinicians published a study in Proceedings of the National Academy of Sciences showing that a wireless version of the patch measured sweat chloride — the gold-standard diagnostic for cystic fibrosis — as accurately as a laboratory sweat test, in a trial of 20 CF patients and 7 healthy controls. For a disease whose diagnosis currently requires a visit to a specialized center, a sticker that works at home changes the map of who gets screened.

Wei Gao’s lab at Caltech attacked the third problem: power. A sweat sensor that needs charging is a sweat sensor with a dead battery. In 2023, Gao published a device in Nature Electronics powered entirely by a flexible perovskite solar cell (20 by 27 millimeters, 4 millimeters thick) that achieved 31% power conversion efficiency under normal indoor lighting. Not sunlight. Office light. “We care about more real-life conditions, including normal office and home lighting,” Gao told Caltech News.

Gao’s group also expanded what a sweat patch can read. A 2022 study in Nature Biomedical Engineering demonstrated continuous detection of all nine essential amino acids, multiple vitamins, and various hormones and drugs from sweat — a range that was previously impossible with wearable electrochemistry. In 2025, the lab published a technique in Nature Materials for printing molecule-selective nanoparticles onto flexible substrates, a step toward manufacturing recognition layers at industrial scale.

The escape from sports

Mercedes-Benz Group AG was granted US Patent 12,304,500 in May 2025. The title is bland: “Method for operating a vehicle.” The claims are not. The patent describes a wearable device that reads cortisol concentration in the driver’s sweat, transmits the measurement to the car’s computer, and adjusts vehicle parameters — steering sensitivity, braking thresholds, warning alerts — if the driver’s stress hormone is elevated. One claim describes identifying the driver by their sweat profile, linking it to a stored database of drivers with license data and experience levels.

A car that knows who is driving by the chemistry of their palms, and decides whether to tighten the brakes based on how stressed those palms are. That is a real claim in a real patent granted to a real automaker.

It’s not alone. Nokia Technologies patented a device in 2025 (US 12,318,073) that uses heat to activate sweat glands and route fluid to biometric sensors. NTT, Japan’s telecom giant, holds a 2024 patent (US 12,089,824) for a microchannel device measuring sodium and potassium ions in sweat through electrodes formed directly on channel walls. City University of Hong Kong patented a battery in 2025 (US 12,279,887) that is activated by sweat itself — a graphene-coated magnesium-air cell that generates power when sweat reaches it, then uses that power to run the biosensor. The device and the fuel are the same fluid.

In the ClinicalTrials.gov registry, a trial beginning July 2025 (NCT07047664) is testing a sweat patch for early kidney disease detection. Another (NCT07414810) is recruiting for a point-of-care sweat chloride device for CF monitoring starting in August 2026.

Who should care

The continuous glucose monitor created a market worth over $10 billion by making one blood analyte painless to track. Dexcom and Abbott’s Libre showed that people will measure something about themselves continuously if you remove the needle.

Sweat carries dozens of measurable analytes. Cortisol (stress). Lactate (exertion). Chloride (cystic fibrosis). Uric acid (gout, kidney function). Amino acids (nutrition). Drug metabolites (chemotherapy compliance). If even one of these reaches the clinical reliability of a CGM — and the chloride and cortisol data are furthest along — the addressable market is comparable, and the competitive moat belongs to whoever solves the last engineering mile: reliable, continuous, affordable sweat sampling on demand.

Cincinnati showed you can force sweat production all day. Northwestern showed you can route it through channels on the skin and manufacture them at consumer-product scale. Caltech showed you can power the whole system from a desk lamp and print the sensors at scale.

The adjacent possible here isn’t a better fitness tracker. It’s a clinical lab that fits under a Band-Aid and never asks for a blood draw. 219 patents say people are building it.

Method note

Patent counts draw from 9.3M US utility grants sourced from USPTO bulk grant XML, January 2005 through September 2025. The sweat-sensor filter requires co-occurrence of “sweat” with sensing-related terms (sensor, analyte, biomarker, diagnostic, monitor, detect, biosensor, wearable) in the full text. This is a conservative filter; the actual count of sweat-adjacent patents is higher. Assignee counts combine variant spellings manually. Literature citation counts are from OpenAlex (357M papers). Clinical trial data is from ClinicalTrials.gov (541K studies). Web-sourced claims are cited inline by publication name.