Can a smartwatch or smart ring measure glucose? How it differs from a CGM

No smartwatch or smart ring has FDA authorization to independently measure or estimate blood glucose without piercing the skin. A wearable can receive and display glucose data from a separately authorized continuous glucose monitor, but the sensor—not the watch or ring—generates the measurement. A wellness score or optical estimate is not interchangeable with CGM data.¹²

The product record must follow the data path. “Glucose on your wrist” can describe an authorized CGM companion display, an unreviewed stand-alone estimate, a manually entered value, or a wellness app. Those are four different systems.

Find the physical sensor

An authorized CGM uses a sensor inserted through skin or implanted under it to measure glucose in interstitial fluid. A transmitter and software process and display the signal. The Eversense record, for example, identifies an implanted sensor, wearable transmitter, mobile app, labeled population, duration, warnings, and limitations.³

A watch or ring claiming to sense glucose independently may cite optical, radiofrequency, sweat, or algorithmic methods. If no separate penetrative sensor exists, FDA’s safety communication applies to the stand-alone noninvasive measurement claim.¹

“FDA registered” does not answer the sensor question

A manufacturer may register a facility or list a device without obtaining product-specific clearance or approval. Search the exact product name, manufacturer, submission number, intended use, and decision letter. A generic registration screenshot cannot prove the watch or ring is authorized to measure glucose.

The approved-cleared-registered guide explains the records. The general-wellness guide explains why a low-risk wellness claim and a disease-management measurement claim have different regulatory consequences.

Beware of “FDA compliant,” “FDA certified,” “clinically validated,” or “medical grade” without a product-specific source. A study can show correlation under one protocol without establishing the product is authorized for treatment decisions.

Display authorization is not sensing authorization

Software and companion displays can be part of an authorized device system. Their job may be receiving, presenting, alerting, sharing, or analyzing sensor data. That does not mean the phone, watch, or ring itself contains an authorized glucose sensor.

Ask a seller to diagram:

1. what touches or enters the body;
2. what creates the raw signal;
3. where the algorithm runs;
4. which device transmits and which displays;
5. what happens if connectivity fails; and
6. which FDA record covers each component and claimed use.

If the answer points to another company’s CGM, the wearable is a display. If the answer points only to LEDs in a ring, require the exact authorization for stand-alone glucose measurement.

Accuracy language needs an intended decision

“Within 10%” can refer to an average, a subset, a range, or a marketing comparison. A usable validation report names the reference method, population, glucose ranges, number of paired measurements, exclusions, calibration, lag, missing data, and error distribution.

Even an authorized CGM has product-specific limitations, interferences, alert settings, and instructions. FDA authorization does not make every reading exact or every product interchangeable. The OTC CGM guide owns the separate question of how authorized nonprescription CGMs can and cannot be used for wellness.

Unverified readings create risk when they influence food, activity, medication, or urgent-care decisions. The central problem is not gadget novelty; it is a number being used for a job the product was not shown to perform.

A marketplace verification path

1. Name the exact claimed measurement. Separate blood glucose, interstitial glucose, a trend, a prediction, and a wellness score.
2. Trace the signal to its physical sensor. Identify whether a separate skin-penetrating or implanted CGM generates the value.
3. Locate product-specific FDA authorization. Match manufacturer, model, submission number, intended use, population, and labeling.
4. Read validation at the decision level. Ask what reference, range, error distribution, calibration, and intended decisions were studied.
5. Preserve a fallback. Document what the labeling says to do when readings conflict with symptoms, fail, or appear implausible.

The decisive question is not “Can the watch show glucose?” It is “Which authorized sensor generated the number, and does the product-specific record support using that number for the decision being considered?”