Short answer: they can be reliable, but I wouldn’t treat “smart” as automatically trustworthy. I get why you keep seeing them—cities are under pressure to cut overflow and unnecessary pickups. The better ones use ultrasonic or time-of-flight fill-level sensors, send data over NB-IoT or LoRa, and flag a bin only after a couple of consistent readings. In a well-run pilot, that can work well.
Where they get unreliable is at the edges. Grease, wet waste, plastic bags, and compacted rubbish can confuse sensors. Batteries die, solar panels get shaded, lids get damaged, and connectivity drops in basements or dense streets. A bin can also be “full” by sensor height but still have usable capacity, or look empty while waste is stacked at the back. Without calibration and cleaning, false alerts creep in.
The data layer matters too. If the dashboard just shows red dots and no one owns the alerts, reliability is irrelevant. You need a clear rule for what “full” means, a maintenance schedule, battery and network monitoring, and a way for drivers to confirm or reject reports. Redundant sensors or a second reading can help, but it adds cost.
So: reliable enough for many cities if deployed carefully, not reliable as a plug-and-play gadget. If you’re evaluating them, ask for real pilot data over several months, false-alert rates, battery life, cleaning and calibration intervals, and what happens when connectivity fails. A small controlled rollout will tell you more than a glossy spec sheet.