Google sister company patents an under-eyelid glucose sensor
Verily is taking a pretty different route with needle-free glucose monitoring. A newly published patent describes a tiny sensor that would sit under the lower eyelid and try to read glucose from the blood vessels there. The US application, 2026/0305985 A1, was published on October 8, although the work goes back to May 2023.
Most companies chasing needle-free glucose are trying to make the wrist work. We’ve already seen WHOOP explore non-invasive glucose monitoring, while Garmin has patents around estimating HbA1c using optical sensors. Verily is taking a completely different route by changing where the measurement happens.
Verily is basically asking a different question: why use the wrist at all?
Why the eyelid makes sense
According to the filing, the blood vessels in the palpebral conjunctiva, the tissue lining the inside of the eyelid, lie less than 100 micrometres below the surface. Their depth is also fairly consistent from person to person.
That is important for an optical sensor. A short, predictable path to the blood leaves less in the way. At the wrist, light must pass through skin and tissue, with blood vessels at varying depths and plenty of other signals adding noise.
The lower eyelid also provides a natural pocket to hold the device in place. Verily describes a soft, curved, flexible sensor that sits between the eyeball and the lower eyelid, with its optics facing the eyelid tissue. One example in the drawings measures around 18 mm long. The filing also mentions smaller versions, with the goal of keeping the sensor largely or entirely hidden once fitted.
Infrared light and a lot of tears
Rather than relying on a chemical sensor, the device uses short-wave infrared light. Tiny LEDs shine light into the blood vessels in the eyelid while a photodetector measures what reflects back. Because glucose absorbs light differently across the infrared spectrum, the choice of wavelengths is central. The patent covers 900 to 1,700 nm in broad terms, with more specific examples in the 1,300 to 1,700 nm range.
The obvious problem is tears. An optical sensor under the eyelid sits in a constant tear film, which can interfere with the reading. Verily addresses this on several fronts. The sensor can use one wavelength range to pick up the glucose signal and a second to account for water, then subtract the water signal from the first reading.
The hardware helps as well. The casing can have a hydrophobic surface that repels tears around the optical components. The filing also describes an oil-based coating between parts of the device and the eyelid to reduce the effect of the tear film. A good share of the patent is devoted to keeping water out of the measurement.
Built to talk to your phone or watch
Power and data are handled wirelessly. Verily describes NFC for communication and potentially for wireless power, so the sensor may not need a conventional battery. Bringing a phone close to the eye would allow it to communicate with the sensor, with Bluetooth and Wi-Fi listed as alternatives.
The patent names smartphones, tablets and smartwatches as devices that could connect to it. The sensor itself would not be a traditional wearable, yet it could fit into a broader wearable health setup.
Glucose is not the only target, either. The filing states that the same approach could detect other substances in the blood, including cholesterol, potassium, iron, magnesium, lactate and insulin. Patent applications are written broadly, though, and naming a substance is a long way from measuring it reliably.
So this is not a sign that a needle-free Verily CGM is just around the corner. What makes the filing worth watching is the location. If optical glucose sensing does eventually work reliably, the sensor that gets there may not sit on your wrist at all.
Source: USPTO
Don’t miss the latest from Gadgets & Wearables
Subscribe to our monthly newsletter and check out our YouTube channel.
You can also follow Gadgets & Wearables on Google News and add us as a preferred source in Google Search.