As a requirement for getting regulatory approval to launch your satellite, I think you have to submit a debris management plan, which can include raising a geostationary satellite to a graveyard orbit after EOL (GEO is really far from earth, it would take a lot of fuel to de-orbit them), or lower a low-earth orbit to an altitude at which it would decay in a few weeks or years.
Generally, I think these small satellites are quite cheap both to launch and manufacture, do not have a great lifespan, and certainly not have lots of fuel. So they are most likely intended for LEO, and a low one with that, where atmospheric drag is small, but could bring a satellite down in a few (~5-7) years. Hopefully smaller than the rate at which the operator renews them.
Picking LEO has many benefits: lower latency, higher bandwidth, smaller antennas, higher resolution for pictures, with a few drawbacks: more earth shadow, faster ground velocity, smaller coverage for earth sensors.
And there is precedent that even if an American company launches their satellite from somewhere that is not the USA, they still have to follow all the American regulation. See FCC and Swarm Technologies over the SpaceBee 1-4 launch.
[1] https://www.spaceacademy.net.au/watch/debris/orblife.htm
The 5-year lifespan is for them to be intentionally deorbited. If they die, they become debris subject to atmospheric drag and other forces.
I think it's inportant to note that a traditional 3 ton sattelite and a cubesat will have totally different lifetimes, and that articles appears to assume a larger satellite
wiki citation - https://spacenews.com/starlink-failures-highlight-space-sust...
And some clouds. Children love clouds.