Belkin's July 2025 announcement that it will end Wemo cloud services and app support on 31 January 2026 quietly turned a lot of working smart plugs and switches into a warning label for anyone who buys technology for a classroom. The lesson is short: if a device stops doing its job the moment a vendor's servers go dark, you were renting it, not owning it. The good news is that you can test for that risk before you spend a cent, and the test is easy.

What happened

In July 2025, Belkin said that support for the Wemo app and the cloud services behind it will stop on 31 January 2026. Wemo is a range of smart-home gadgets such as plugs, light switches and cameras that owners control through a phone app. A lot of that control travels through Belkin's servers rather than staying inside your home. When the app and those servers are retired, the remote control, schedules and automations that rely on them stop working. The hardware still has power and still physically switches, but it loses the software brain that made it worth buying. The decision prompted a long community discussion; one detailed thread lives on Hacker News.

Why this is a procurement problem, not just a gadget problem

Buying technology for a school, a lab or a team is a multi-year bet. You are not just paying for the box; you are paying for the years of use you expect to get out of it. When the smart part of a device lives on the vendor's servers, you do not control that timeline. The vendor can change its business, drop a product line, or retire an app, and your working hardware becomes a paperweight on a date you did not choose.

For South African budgets the sting is sharper. Replacing kit that still switches on is a hard line item to justify in rand, especially when the original purchase was signed off as a long-term investment. There is a local-conditions angle too: anything that needs a constant connection to a distant server is already fragile here. Load shedding and patchy connectivity mean cloud-only devices can go dark for reasons that have nothing to do with the vendor at all. A tool that only works when the internet is perfect is a tool that will fail you during a lesson.

The works-offline test

Before you commit to any connected device, run one experiment with a demo unit or a colleague's sample. Set it up normally, then cut it off from the outside world by switching the router off or putting the controlling phone or tablet into airplane mode, and see what still works.

• Does the core function still happen? A plug should still switch, a robot should still run its last program, a sensor should still read.
• Can you still change its behaviour locally, or does every edit need the vendor's app and account?
• Does it recover cleanly when the connection comes back, or does it need a full re-setup?
• If the vendor vanished tomorrow, is the device a brick or just a slightly less convenient tool?

If the honest answer is that it becomes a brick, you have found a cloud dependency, and you should price that risk into the decision.

A buyer's checklist for cloud-dependent tech

You will not always avoid the cloud, and you should not have to. Plenty of good tools use it for genuine extras. The goal is to know what you are buying. Ask the vendor these questions, in writing, before a large order:

1. Which features stop working if your servers or app are switched off?
2. What is your track record and stated policy for supporting older products?
3. Can the device be controlled or programmed on a local network with no internet at all?
4. Do we need a per-user account to do the basics, or only for optional features?
5. If the account system closed, would our existing units keep running?
6. Are the files and exports we create ours to keep, or locked inside the app?

Answers that lean toward works on its own are worth paying a little more for. Answers that lean toward trust us to stay in business deserve a discount for the risk you are absorbing.

What this means for coding and robotics kits

Robotics and coding kits are especially exposed, because a locked-out kit is a locked-out lesson. If a class set of ten kits suddenly needs an account server that happens to be offline, a 40-minute lesson can evaporate. When we buy and build for classrooms, our rule is that the child's program should run on the board itself, not on a distant server. The sheenbot∞ board keeps the code on the device, and its block editor is designed to work without a login so a lesson does not hang on a screen that will not load. It is the same principle behind the test above: the smart part should live where the teacher can reach it.

This shapes how we run the academy too, where we favour tools that keep working when the network does not. If you want to see the difference in practice rather than take our word for it, a free trial class is the quickest way, and our winter-holiday workshops put the hardware straight into a child's hands.

Takeaway

The Wemo shutdown is not really a story about one company or one line of plugs. It is a reminder that smart often means dependent, and dependence is a cost you pay later, on someone else's schedule. Before the next purchase order, run the works-offline test and ask the six questions above. Buy tools you own rather than tools you rent by permission, and your future budget, along with your next lesson, will thank you.