Is ChatGPT Down? Why Your Smart Home Isn't Responding
Smart home devices not responding; ChatGPT outage suspected.
Last updated
ChatGPT may be down, but that is almost certainly not why the smart home stopped responding.
Status label: Confirmed as of Aug. 25, 2026. ChatGPT outages happen, and OpenAI’s own status page listed ChatGPT at 99.68% uptime across 15 components for the May–Aug. 2026 window, but that is a ChatGPT availability measure, not a smart-home control measure.[1] Your next move is not to factory-reset a hub. Stand next to one affected device and run the local test below.

The reason is plain: ChatGPT is not a native smart-home control path. It is not the thing your wall switch, Zigbee bulb, Matter device, HomeKit hub, SmartThings hub, Alexa routine, or Google Home speaker normally depends on to turn a lamp on. Older reporting on ChatGPT smart-home experiments described hobbyist and developer routes through Home Assistant APIs and Josh.ai-style experiments, not a mainstream built-in control layer for ordinary homes.[2] That distinction matters. If you built your own OpenAI-powered Home Assistant bridge, troubleshoot that bridge. If you did not, ChatGPT trending on an outage site is almost certainly background noise.
Run the 60-second local test before changing anything
Pick one device that is failing right now. A light is best because the result is visible. Do not start with the reset pinhole, the breaker, or the “remove device” button in the app. Those are recovery tools, not first tests.

- Press the physical switch, button, dimmer, lock keypad, or device control. If the device works physically, the device still has power and is not dead.
- Open the manufacturer or platform app and try the same action. Use the app that normally controls the device, not ChatGPT.
- Trigger or wait for a known local automation if you have one. A motion sensor turning on a light through a local hub, for example, is more useful here than a cloud voice command.
- Check whether phones and laptops on the same Wi-Fi can browse ordinary websites.
- Look for scope: one device, one room, one platform, or the whole house.
A local automation is especially useful because it separates the home from the cloud. If a hub-local rule still fires while app or voice control is dead, the hub and local network are probably alive; the broken piece is likely the cloud account, platform API, phone app path, or voice service. A SmartThings community thread from Sept. 5, 2024 showed that pattern: users reported app trouble while local automations continued, and the official status page did not clearly show an outage at the time.[3] That does not make a forum thread proof by itself. It does show why the thing still working inside your house is often more persuasive than a status page that may lag.
| What you see | Most likely layer | What to do next |
|---|---|---|
| Physical switch works, app and voice do not | Cloud service, account, platform app, or voice assistant | Check the platform status page and recent user reports; wait before resetting devices |
| Physical switch works and local automation still runs | Hub and local network are probably alive | Treat it as a cloud/app/voice path problem, not a device-pairing problem |
| Nothing on Wi-Fi can browse | Internet path, router, modem, ISP, or power | Troubleshoot network and power before touching smart-home pairings |
| Only one room or far corner fails | Wi-Fi coverage, mesh backhaul, or local signal reach | Check access points, mesh nodes, and recent furniture or device moves |
| One hub and its child devices fail locally | Hub, bridge, radio, or power to that hub | Use a hub-focused diagnostic path |
This split matches the practical network diagnosis used for smart homes: if ordinary browsing still works and only one smart platform fails, the problem is more likely that platform or account path; if nothing browses, look at the internet, router, or power path; if only one room is affected, look at coverage and reach.[4]
What the test tells you
Physical control works, but app or voice control is dead
Do not reset the bulb, lock, plug, or hub just because the app is spinning. The device has already proven it can still do its basic job. Now you are looking at the remote-control path: the platform cloud, account authentication, the phone app, the voice assistant, or the connection between them.
This is where official status pages and outage-report sites are useful, but only in their proper place. An official page can confirm a known incident, but it may lag real user experience. A spike in user reports can corroborate that others are seeing the same thing, but it is not a diagnosis of your house. If the switch works and only the platform app fails, waiting is often the least destructive fix.
The cautionary case is Google Home and Nest. During a reported Google Home/Nest outage, users found devices stuck and unresponsive, and reboots or factory resets did not fix the issue because the failure was on Google’s servers.[5] That is the evening nobody needs: the platform comes back, and now someone still has to re-pair the devices that were reset for no reason.
Local automations still run
If a motion sensor still turns on a light, a contact sensor still triggers a siren, or a hub-local routine still runs, your home is giving you a useful answer: the local pieces are probably not the immediate failure. The cloud path can be broken while the house keeps doing local work.
That is why local automations are not just a hobbyist preference. They are a recovery clue. A recipe like a Home Assistant tornado warning automation is not relevant because everyone needs that exact automation; it is relevant because it shows the difference between work that happens inside the home and work that depends on a cloud round trip.
Nothing browses, or the whole home is offline
If phones, laptops, smart speakers, and hubs all lose internet at the same time, stop checking ChatGPT. The failed layer is likely power, modem, router, ISP service, or a mesh node that did not recover cleanly. Smart-home apps often make this look like dozens of device failures because every tile turns gray at once.
If the outage followed a utility blink or a router reboot, move to a power-recovery path instead of deleting devices. Start with smart devices offline after a power outage or the more specific reconnect devices after a power outage playbook once you know the cloud is not the main issue.
One hub or bridge fails locally
If devices tied to one hub fail together and local automations tied to that hub do not run, that is no longer a ChatGPT question. Check the hub’s power, Ethernet or Wi-Fi connection, radio state, firmware state, and whether child devices are still joined. Use the smart home hub troubleshooting branch before doing anything that removes devices from the network.
Only one room fails
A one-room failure usually deserves a boring explanation first: Wi-Fi reach, a dead mesh node, a powered-off access point, a moved hub, or a device sitting at the edge of radio range. This is the branch where walking the house beats refreshing outage maps. If the hallway switch works but the far-bedroom plug does not answer from the app, look at the path between that room and the nearest radio before blaming a global service.

When ChatGPT and smart-home services really can fail together
There is one narrow exception worth knowing: shared upstream infrastructure. On Oct. 20, 2025, a major AWS incident involving DNS and DynamoDB in US-EAST-1 lasted roughly 15 hours and affected a wide range of services, with TechRadar reporting that Alexa voice control, Ring, and ChatGPT were among services disrupted during the broader outage.[6] In that kind of event, ChatGPT and smart-home platforms can be down at the same time.

That still does not mean ChatGPT caused the bulbs, locks, cameras, or speakers to stop responding. It means multiple services depended on the same upstream cloud infrastructure and were disrupted by the same upstream failure. The fix remains the same: test the local device path, identify the failing layer, and avoid turning a temporary cloud incident into a re-pairing project.
So, is ChatGPT down right now?
Check OpenAI’s status page if you need the ChatGPT answer, but keep it separate from the smart-home diagnosis. OpenAI’s page is the right source for ChatGPT component health; it is not a status page for Google Home, Alexa, SmartThings, Apple Home, Ring, Hue, Aqara, or your router.[1]
ChatGPT has had outages. CNBC reported that a Dec. 2, 2025 ChatGPT outage affected some users, was tied to a routing misconfiguration, was resolved the same day, and drew about 3,000 Downdetector reports.[7] That is a real ChatGPT incident. It is also the wrong explanation for a smart lock, a Zigbee light, or a Google Home routine unless you have a custom integration that explicitly sends your home-control commands through OpenAI.
If your ISP is having trouble, use an outage-recovery path like the AT&T outage smart-home offline fix. If your hub fails the local test, use hub diagnostics. If the switch works and local automations still run, leave the reset buttons alone and watch the relevant platform status page. ChatGPT chatter is not enough evidence to rebuild a smart home.
References
- OpenAI Status, OpenAI
- Can ChatGPT Control Your Smart Home?, MakeUseOf, June 2023
- Smartthings App Outage or just me? 5 Sept 2024, SmartThings Community, Sept. 5, 2024
- What Happens to Your Smart Home When the Internet Goes Down?, Arizona Network
- Google Home outage left Nest devices stuck and unresponsive, Android Central
- Amazon fixes huge AWS outage that broke much of the internet – here's what happened, TechRadar
- OpenAI's ChatGPT suffers outage for some users, CNBC, Dec. 2, 2025
Corroborating context
For protocol background on why this failure happens, see Compatibility & Protocols.
Not currently linked to a known regression. Background on the underlying protocol lives in Compatibility & Protocols.
