Geostationary is a 44,000 mile round trip. So a request/response for a packet is 90,000 miles. Already at 1/2 of a second latency without any actual networking because of the speed of light. And that is if you are on the equator. Another tenth or two is probably lost at middle latitudes, and worse for higher latitudes. So in practical situations, your latency is a second or more.
And they can only put one satellite in one position, higher power to send the signal, no cheap way to replace/upgrade the satellite, single point of failure.
They cannot compete with a flotilla of low orbit satellites providing multiple powers of 10 better total bandwidth. Especially since Starlink has demonstrated it can communicate with current mobile phones.
If you don't care as much about latency (and many applications don't, e.g. asset tracking, news broadcasting or forwarding etc.), GEO is still very viable.
Are you sure about that? High-throughput LEO inevitably requires steering (whether electronic or mechanical), while high-throughput GEO is possible using a one-time adjusted dish, at least to station-kept satellites (which all HTS these days are, as far as I know).
For mobile applications, steering is obviously required for both, so there GEO has less of an advantage.
The Starlink dish is quite expensive, and I'm not sure if it's really sold at cost or whether there's some subsidy baked in (since it's only usable with their service anyway).
More complicated, yes, but produced at much, much higher volume. That is critical. A (year by year even more) niche market for applications that are ok with high latency, low-ish bandwidth and are stationary will not support proper volume production. Meanwhile the biggest LEO operators will building potentially millions of units.
If you only care about low bandwidth, with LEO you potentially don't even need a satellite terminal e.g look at starlink Direct to Cell, the physics about steering matter less in the case of low bandwidth since you can run a separate service on different frequencies with better radio characteristics.
They already know they can't compete on merit alone, so they'll go for lobbying and regulatory intervention so they can "compete".