The Firmware Clock: Every Connected Device Has an Expiry Date You Didn't Read

Buy a deadbolt and you can reasonably expect it to outlive the door. Buy a smart lock and you have bought two products bolted together: a mechanical device that will happily work for fifteen years, and a software product with a lifespan nobody printed on the box. The second one decides when the first one stops being useful. Almost nobody checks it before paying.

This is the firmware clock, and it runs on every product with a radio and an app. Smart locks, thermostats, doorbells, cameras, hubs, TVs, wearables, appliances with a companion app. The clock starts the day the product ships and stops the day the maker decides the model is no longer worth engineering time. That date is rarely announced in advance, almost never printed on the packaging, and it does not track the physical condition of the thing you own at all.

What actually breaks when the clock runs out

Support ending is not a single event. It arrives in stages, and each one is easy to miss until it isn’t.

First, security updates stop. Nothing visible happens. The device keeps working, and you keep trusting it, while newly discovered flaws in its software go unpatched forever. Second, the app decays. Mobile platforms change their requirements every year, and an app nobody is maintaining eventually fails to meet them, gets removed from the store, or simply stops launching after a phone OS update. If setup, pairing, or day-to-day control runs through that app, you now own a device you cannot configure. Third, and most abruptly, the cloud service goes dark. Servers cost money to run, and when a product line is discontinued or a company is acquired, those servers are a line item with no revenue behind it. When they are switched off, every device that depended on them for authentication, remote access, or basic logic becomes inert.

The pattern is consistent: the hardware outlives the software, sometimes by years, and you are left with a working machine and nothing to talk to.

Cloud dependence is the highest-risk design

The single best predictor of how badly a device ages is how much of its intelligence lives on someone else’s server.

A camera that records to a local card or a network drive is a camera. A camera that streams to a vendor’s cloud, authenticates against a vendor’s account system, and shows you footage only through a vendor’s app is a subscription to a service that happens to include a lens. A thermostat that holds its schedule on-device keeps heating your house when the company folds. A thermostat that fetches its schedule from a server each morning does not. Same box, same sensors, completely different failure behavior.

Ask a blunt question about anything you’re considering: if this company’s servers went offline permanently tomorrow, what would this device still do? If the answer is “most of what I bought it for,” the risk is manageable. If the answer is “nothing,” you are not buying hardware. You are renting access to hardware you paid for in full, and the lease has no stated end date.

The security dimension is the part people underrate

An abandoned connected device is not neutral. It is an unpatched computer with network access sitting inside your home, and it will stay there for years because nothing about it looks broken.

This matters more in some categories than others. Cameras and doorbells carry a live feed, locks control physical access, hubs sit at the center of everything else, and routers are the front door to your whole network. A device in any of those categories that no longer receives security updates is an active liability, not a slightly outdated gadget. Retire it, or at minimum isolate it on a guest network with no access to anything else you own.

Open standards are the real hedge

Local control is what survives corporate decisions, and open standards are how you get it.

Matter, Zigbee, Z-Wave, HomeKit, and ONVIF for cameras all share one useful property: they let a device be driven by something other than the manufacturer’s cloud. A Z-Wave sensor paired to a hub you control does not care whether its maker still exists. An ONVIF camera can be recorded by third-party software indefinitely. A Matter device can be adopted by a different controller when the original app dies. None of this makes a product better on day one. It makes it recoverable on the day the vendor stops caring, which is the only day this argument is about.

The corollary is uncomfortable for anyone who likes new things: a startup with a gorgeous app is a bigger gamble than a boring incumbent with an ugly one. Proprietary protocols concentrate all your risk in one company’s survival, and small companies pivot, get acquired, and shut down product lines far more often than large ones. That is not a reason to avoid young companies. It is a reason to prefer, from one, the model that speaks a standard you could re-home elsewhere.

The pre-purchase check

Five minutes, before you pay.

Search the maker’s site for the words “end of life,” “product lifecycle,” or “security updates.” A published minimum support window is a strong buy signal, and its absence is information too. Then find the firmware release notes for the previous generation of this product and check the date on the most recent entry. A company still patching a four-year-old model behaves differently from one that ships and forgets.

Next, confirm which standard the device speaks, and whether it can be controlled without the vendor’s app or account. Then run the offline test: unplug your internet, or read owner reports from people who have, and find out what the device still does. Finally, weigh the company itself. How long has this product line existed, and what happened to the last one it replaced?

So what should you do

Treat software support as a spec, not an afterthought. For anything with a radio, favor devices that keep their core function locally, that speak a standard you could move to another controller, and that come from a maker with a published support window and a visible history of honoring it. Pay a premium for that where the device is load-bearing: locks, cameras, hubs, network gear. Accept more risk where it isn’t, like a light bulb or a plug you can cheaply replace. The question was never whether the hardware will last. It’s whether anyone will still be maintaining the software that makes the hardware worth owning, and that’s a question you can answer before you buy instead of four years after.

Real Buyer Experiences is reader-supported and independent. We synthesize independent testing and owner reports into one honest verdict - we never trade a recommendation for payment. How we review →