Why Did Amazon Kill Palm Payments in 2026? Biometric Checkout Isn't Dead — It's Just Built for a Different Job.
Amazon discontinued its Amazon One palm-payment system in January 2026, removing every palm reader from physical locations by 3 June 2026 and citing low customer adoption — all enrolled customer data is being deleted. But biometric checkout in unattended retail is not dead. It is splitting in two. The consumer “pay with your palm” experiment is retreating, while the spoof-resistant, fit-for-purpose side — palm-vein recognition and biometric authentication for age gates, employee access and industrial dispensing — keeps growing at roughly 17% a year (Research and Markets). What died was a specific consumer payment convenience. What survived is biometrics used where the machine must prove who you are, not just charge you. For vending and kiosk buyers, the lesson is the same as every other spec decision: biometrics is a requirement to be designed around, not a bolt-on feature.
Amazon One was the poster child of biometric payment.
Now it’s gone.
Amazon killed it in January 2026.
The whole thing — the palm readers, the enrollment data, the “just wave your hand” pitch — wound down by 3 June.
That is not the end of biometrics in vending.
It is the end of the wrong version of biometrics.
What actually happened
Read the timeline.
Amazon launched Amazon One in 2020 at two Amazon Go stores in Seattle.
It let enrolled customers pay by scanning their palm.
It spread to Whole Foods. Then to third-party venues like Panera Bread.
In January 2026, Amazon pulled the plug.
“Limited customer adoption” was the official reason.
Every palm reader came out of physical locations by 3 June 2026.
The enrolled data was deleted.
One number makes the contrast sharp.
Amazon’s Just Walk Out — the cashierless checkout that needs no palm at all — kept growing through the same period.
360-plus locations across five countries.
150-plus third-party stores launched in 2025 alone.
36.7 million items tracked across 17.7 million shopping sessions.
Palm payment died.
No-payment checkout lived.
That is not a coincidence.
The split
Biometrics is not one thing.
It split into two jobs.
| Technology | What it does | Where it landed in 2026 |
|---|---|---|
| Palm scan (Amazon One) | Pay with your hand | Discontinued — low adoption, privacy friction |
| Palm vein | Verify identity, spoof-resistant | Growing — age gates, access, industrial dispensing |
| Just Walk Out | Identify items, not people | Expanding — 360+ locations |
| QR / card / RFID | Credential-based auth | Dominant — still the default |
One of these tried to replace a credit card.
Two of them replace a manual check.
The one that replaced the card died.
The ones that replaced a manual check grew.
Why the consumer version died
Paying with your palm solved a problem nobody had.
A card tap is already near-instant.
Enrolling your palm meant handing over biometric data to save half a second at the till.
The privacy cost was concrete. The convenience was marginal.
So adoption stalled.
And a biometric feature with no adoption is just an expensive, awkward camera.
Amazon is ruthless about killing things that don’t clear the bar.
That is the entire story.
Why the authentication version lives
Now flip the lens.
A machine that must know who you are is a different problem.
An age gate that can’t be borrowed from a friend.
An industrial dispenser that logs which employee took which tool, at what time.
A medicine cabinet that won’t open for the wrong person.
These are not payment conveniences.
They are compliance obligations.
The palm-vein biometric market is projected to grow from US$1.23 billion in 2025 to US$1.44 billion in 2026 — roughly 17% a year (Research and Markets).
Retail payment is about 63% of biometric-palm applications, and palm vein holds about half the technology segment (Fact.MR).
Palm vein survives where palm scan didn’t for three reasons.
-
Spoof resistance. A vein pattern is inside the body. It is far harder to fake than a surface palm print.
-
Hygiene. No contact surface to touch in a public place.
-
Fit for purpose. It answers a question the machine has to ask — “who are you?” — rather than adding friction to a question a card already answered.
That is why the market keeps buying it while the consumer gimmick died.
What to spec in 2026
The lesson is the same as every other vending decision.
Biometrics is a requirement, not a feature.
Five questions to ask before you add a biometric reader to a machine.
-
What must the machine prove? Age, identity, entitlement to dispense — name the job first.
-
Is a credential enough? If a QR code, RFID badge or card answers the question, biometrics is over-engineering.
-
What is the spoof risk? A surface palm print is cheap to fake. A vein pattern is not. Match the sensor to the threat.
-
What does the regulator require? Age-gated vending and employee safety are governed by law, not preference. The spec follows the law.
-
Who maintains it? A biometric reader is a sensor in the field. It fails, drifts, and needs support. Budget for it like any other component.
The payment rail underneath — QR, card, RFID, prepaid account — is a separate decision. The full vending payment method matrix is here. Biometrics sits on top of that rail, not instead of it.
Amazon spent years and untold capital proving one thing for free.
A biometric that only adds convenience loses.
A biometric that removes a manual check wins.
The KioskForce position
We build machines around the authentication the job actually needs.
Not around whatever biometric is fashionable this quarter.
An age gate. An employee badge reader. A QR scanner. A vein reader where spoof resistance is the whole point.
The reader is a spec line.
The decision — what must the machine prove, and to whom — comes first.
That is the difference between a machine that sits in a store and a machine that sits inside a compliance process.
Design the requirement first.
The hardware follows.
If you are specifying a machine that has to know who it is dispensing to, start with the requirement. Not with the sensor.
Sources: Mass Market Retailers — “Amazon to stop Amazon One palm payments” (28 January 2026): launch 2020 at two Amazon Go stores in Seattle, expansion to Whole Foods and Panera Bread, discontinuation with removal by 3 June 2026 citing limited adoption, customer data deleted, Just Walk Out at 360+ locations across five countries with 150+ third-party stores in 2025 and 36.7 million items across 17.7 million sessions. CNBC — Amazon One launch announcement (September 2020). Research and Markets — Palm Vein Biometric Market Report 2026: US$1.23 billion (2025) to US$1.44 billion (2026), ~17% CAGR. Fact.MR — Pay by Palm Market: retail payment ~63% of biometric-palm applications, palm vein ~half of the technology segment. KioskForce operating facts — Nanjing design office and partner factories in Cangzhou — are stated on this site.
Want something like this built?
We design and manufacture custom vending machines, kiosks and the cloud software behind them. Tell us what you have in mind.
Project brief · 60 seconds
What are we building together?
Three quick steps. Your brief goes straight to the engineers who design, build and support KioskForce systems.
