The New York company that decided the humble spare key deserved a neural network - and built 7,000 robots to prove it.
Here is a fact that should be more surprising than it is: the American locksmith industry is worth something like $12 billion, and until fairly recently almost none of it ran through an app, a screen, or a price you could see before someone showed up at your door. You lost a key, you called a number on the side of a van, and you hoped. This is roughly the market structure of the 1970s, preserved into the smartphone era like an insect in amber.
KeyMe's founder, Greg Marsh, discovered this the way most people discover the locksmith industry - involuntarily. He needed his home's locks changed, the experience was miserable, and instead of simply complaining about it, he did the founder thing and asked whether the whole process could be rebuilt. The answer he arrived at, in 2012, was a machine. Not a person with a truck, but a kiosk: a self-service box that could scan your key, understand its geometry, and cut a copy on the spot.
The technically interesting part is that keys are harder than they look. A house key is a small piece of brass whose value lives entirely in a series of precise cuts, and there are thousands of variants. Rather than build a mechanical library, KeyMe trained neural networks on those key types - using computer vision to look at a key and infer its structure with enough precision to reproduce it. The company says its hardware and software support roughly 50 times more key types than a traditional key-cutting machine. That is the sort of number that sounds like marketing until you remember that the alternative is a human squinting at a blank and a grinding wheel.
Once you have a robot that can read keys, a few things follow almost automatically. You can put it where people already are - so KeyMe kiosks live inside Kroger, Walmart, IKEA, Menards and Rite Aid, among others. You can let people store a digital scan of a key in the cloud through a mobile app, so a lost key becomes a printable file rather than a locksmith call. And you can move beyond brass: KeyMe's kiosks handle car keys, including the ones with transponders that talk to the car, plus RFID fobs, which are the keys that don't look like keys at all.
The kiosks scaled the way good hardware scales when it hides inside someone else's foot traffic. In 2016 the company had around 200 machines and had fabricated its first million keys. By 2025 that number had crossed 7,000 kiosks, and the company expects to cut more than 10 million keys in a single year. Somewhere in there, copying a key at the grocery store stopped feeling like science fiction and started feeling like a chore you could knock out between the produce aisle and checkout.
But a key-copying kiosk has a natural ceiling, which is that people do not lose keys very often. So in 2020 KeyMe did the sensible expansionist thing and put humans back into the loop - not the old van-and-a-number humans, but a nationwide dispatch network of professional locksmiths for the jobs a machine can't do: lockouts at midnight, re-keying a new apartment, installing and repairing locks for homes, businesses and cars. The pitch is a tidy division of labor. Robots handle the routine and predictable; people handle the emergencies and the judgment. The customer, ideally, never has to know which one they're getting until it shows up.
Then there is the part of the business that is either a clever second act or a sign that every hardware company eventually discovers advertising. Each KeyMe kiosk has a screen. Screens attract eyeballs. Eyeballs, when counted, become a media network. In 2021 KeyMe launched exactly that - a retail media network that by 2025 was generating more than 2 billion advertising impressions. It is a familiar move: the machine you built to do one useful thing turns out to be, incidentally, a place to sell attention. The keys may be the product, but the screen is real estate.
None of this happened without a fight. In 2021, KeyMe went to trial in Marshall, Texas - the patent-litigation capital of America - against The Hillman Group, whose minute key kiosks are the most direct competitor. A jury found that KeyMe did not infringe Hillman's patents. When you build a physical machine that undercuts an incumbent, litigation is less a risk than a milestone; KeyMe cleared it.
The money followed the machines. KeyMe raised across eleven rounds, from a $300,000 angel check in 2012 to a $35 million Series E in January 2020, with Comcast Ventures, Battery Ventures and, tellingly, 7-Eleven among its backers along the way. Roughly $200 million total has gone into the unglamorous proposition that copying a key should be fast, transparent, and available wherever you happen to be standing. It is not the flashiest thesis in venture capital. It may be one of the more legible ones.
Walk up to a machine in a store, scan your key, and get a copy in minutes - brass house keys, transponder car keys, and RFID fobs included.
Digitally scan and store your keys in the cloud. Lose one, and you can print a spare later at any kiosk instead of calling anyone.
Dispatch a vetted professional for lockouts, re-keying, lock installation and repair - residential, commercial and vehicle.
The kiosk screens double as ad inventory, generating 2+ billion impressions a year for brands reaching shoppers in-store.
"The kiosks use computer vision and AI to scan and infer a key's structure with high precision - enabling faster, more accurate duplication."
| Round | Year | Amount | Notable Investors | Scale |
|---|---|---|---|---|
| Angel | 2012 | $300K | Ravin Gandhi | |
| Seed | 2013 | $2.3M | Battery Ventures | |
| Series A | 2014 | $7.8M | - | |
| Series B | 2016 | $20M | Comcast Ventures, 7-Eleven | |
| Series C | 2017 | $15M | - | |
| Series D | 2018 | $25M | - | |
| Growth | 2019 | $50M | - | |
| Series E | 2020 | $35M | - |
Greg Marsh starts KeyMe in New York after a frustrating lock change; raises $300K in angel funding.
Launches self-service key-copying kiosks and a mobile app; raises $2.3M seed from Battery Ventures.
Raises $7.8M to expand the kiosk footprint and its computer-vision technology.
$20M round led by Comcast Ventures with 7-Eleven; hits ~200 kiosks and 1M keys made.
Fuels national expansion of the kiosk and services network.
Raises $35M and launches residential, commercial and vehicle locksmith services.
Wins its patent case against Hillman Group and launches a retail media network.
Operates more than 7,000 kiosks and expects to cut over 10 million keys in the year.
Kiosks across grocery locations.
Kiosks in stores nationwide.
Kiosks in retail showrooms.
Home-improvement store kiosks.
Pharmacy retail kiosks.
Early host & Series B investor.
KeyMe is a technology company that copies keys through self-service robotic kiosks and provides 24/7 locksmith services through a nationwide network of professional locksmiths.
KeyMe kiosks can duplicate standard brass house keys, car keys (including keys with transponders), and RFID key fobs using computer vision.
KeyMe kiosks are in over 7,000 retail locations, including Kroger, Walmart, IKEA, Menards and Rite Aid stores across the United States.
KeyMe was founded in 2012 by Greg Marsh, who created it after a difficult experience getting his home's locks changed; he now serves as president.
Yes. Since 2020 KeyMe also dispatches professional locksmiths for lockouts, re-keying, lock installation and repair for homes, businesses and vehicles.
Watch: KeyMe product demo - copy keys for home, office, vehicle & RFID · Interviews and demos on the KeyMe YouTube channel.