All those points are true, but it doesn't change the fact that Starlink will be quite profitable for SpaceX.

Currently, each launch of 23 Starlink satellites costs SpaceX around $50 million. To get 1,000 direct to cell satellites in orbit, they'll need to launch 44 times, costing them $2.2 billion. Due to the low orbits, air resistance causes the satellites to reenter within 5-10 years, so to maintain the constellation they'll need to spend $220-440 million per year. These costs will be much lower when they switch from Falcon 9 to Starship.

Now let's say only 1% of the population wants Starlink direct to cell. That's still 80 million people. If SpaceX charges cell companies $10/month per user for the service, that's almost $10 billion per year. And that's not counting the money they make from selling Starlink Internet, which currently has over 4 million subscribers. At $100/month, that's $4.8 billion per year in revenue.

So Starlink is profitable without direct to cell technology, but since they're launching the satellites anyway, they might as well collect more revenue by adding cell capability. DTC only becomes unprofitable if the cost of the extra hardware and mass is less than DTC subscriber revenue.