Type “Bitcoin ATM near me” into Google Maps and you’ll get pins. Some of those pins are machines that haven’t existed for months. USA Crypto Reports checked 200 listings pulled from public directories and found that 41 of them, roughly one in five, no longer existed at the address given. The machine was gone, the store had closed, or the location had never installed one in the first place. This isn’t a glitch. It’s how the data supply chain for Bitcoin ATMs works.
Where the location data actually comes from
Almost every map you’ve ever used to find a Bitcoin ATM, including the big aggregator sites, pulls its listings from the same root source: the operators themselves. Bitcoin ATM location data largely originates from operator self-reporting. A company installs a machine in a gas station or a smoke shop, submits the address to one or more directory services, and that listing then gets copied, scraped, and redistributed across dozens of other sites. There is no independent surveyor driving around confirming that machines are physically present. There is no government registry of installed kiosk locations that a map can check against. The entire system runs on operators telling directories what’s true, and directories trusting them.
That works fine when a new machine goes in. It works badly when a machine comes out.
Nobody is paid to say a machine is gone
Adding a listing gets an operator customers. Removing a listing gets an operator nothing. Operators have little incentive to promptly delist a removed machine, so when a kiosk is pulled from a convenience store, the update to the directory feed often just doesn’t happen, or happens weeks or months later. Nobody at the company is tasked with combing through old listings to prune the dead ones. The commercial incentive points entirely in one direction: keep the count of “locations” looking as large as possible, because that number is part of how these companies market themselves to investors and to the public.
The result is a kind of one-way ratchet. Machines get added constantly. They almost never get formally subtracted. Over time the published directory drifts further and further from the physical reality on the ground, and the gap only closes when someone, usually a frustrated customer standing in front of an empty spot where a kiosk used to be, reports it.
The Bitcoin Depot collapse shows how fast stale data can pile up
The clearest recent example of this problem at scale is Bitcoin Depot’s Chapter 11 bankruptcy filing in May 2026, after which more than 9,000 of its machines were deactivated. That’s not a handful of kiosks quietly disappearing from a few strip malls. That’s a mass deactivation event across the country, all at once. Any aggregator site that hadn’t specifically gone in and pulled Bitcoin Depot’s listings would have continued showing thousands of dead pins, each one pointing a would-be customer toward a machine that no longer accepted cash or dispensed Bitcoin. Bitcoin Depot machines, more than 9,000 of them, were deactivated after its May 2026 Chapter 11 filing, which stale aggregators would still show unless someone actively went in and scrubbed the feed. This is the structural problem laid bare: one company’s bankruptcy can instantly stale-date thousands of listings across the internet, and there’s no automatic mechanism that catches it.
If you’re comparing kiosks, it’s worth understanding the basic vocabulary operators use, which our site glossary covers in plain language, since fee structures and machine types get described with jargon that varies by company.
Why a stale listing costs you more than a wasted trip
A dead pin on a map is annoying, but the deeper risk is what happens if you do find a working machine without checking anything else about it first. The Federal Reserve Bank of Kansas City reported in 2023 that the median fee for buying Bitcoin from a US Bitcoin ATM, self-reported by operators, is 16 percent of transaction value, with a median sell fee of 15 percent. On top of that, exchange rate costs add an estimated 5 to 7 percent, and the same report noted that total fees of 20 percent may not be uncommon. Compare that to the estimated cost of actually running one of these machines, which the Kansas City Fed put at only 3 to 6 percent of revenue, and the markup becomes obvious.
It gets worse. That 16 percent median is self-reported, and operators charging above-average rates commonly disable the reporting feature entirely, which means the published median is biased low. The worst-priced machines in the country are systematically underrepresented in the very data used to describe “typical” fees. Combine that with the fact that Bitcoin transactions are irreversible once confirmed, so there is no chargeback equivalent, and you have a transaction type where the price can be high, the operator can be difficult to identify, and there is no bank on the other end to reverse a mistake. US Bitcoin ATM operators are classified as money services businesses subject to federal anti-money-laundering obligations, and licensing is administered state by state, which means the regulatory floor exists, but it varies depending on where you’re standing, and it does nothing to cap the fee an operator can charge.
None of this is a reason to avoid Bitcoin ATMs outright. It’s a reason to treat the map as a starting point, not a guarantee, and to check the fee screen on the machine itself before you insert cash.
What a “last verified” date is actually protecting you from
Every listing in our verified Bitcoin ATM directory carries a last-verified date for exactly this reason. A date doesn’t mean a machine will definitely be there when you arrive; nothing can promise that in an industry running on self-reported data and one-way updates. What it means is that someone confirmed the listing at a specific point in time, rather than letting a five-year-old scrape sit unquestioned on a page designed to look current. Given that one in five listings we checked had already gone dark, a visible date is the honest way to communicate how much confidence you should place in any single pin.
If you show up somewhere and the machine is gone, or it’s been moved to a different aisle of the same store, or it’s simply switched off, that information is worth more to the next person than it is to you at that moment. You can report a machine that has closed or moved in under a minute, and it directly shortens the gap between reality and what the directory shows. In an industry with no independent inspector and no financial incentive for operators to clean up their own listings, reader reports are close to the only correction mechanism that exists. Before your next trip to a Bitcoin ATM, check the listed verification date, and if you find it wrong, report it rather than just moving on to a different pin.
