I believe during the video he uses pricing offered by Tesla(EDIT: SpaceX rather) as the optimistic pricing for Starship! So taking the "cheap" numbers!
Tesla doesn't have anything to do with Starship launch prices. Let's look at costs to SpaceX for launches, since those are the most relevant numbers:
Reused Falcon 9 launch: $15M[1] in the best case, but let's double it just to ballpark the average, so $30M. Falcon 9 can loft around 50 satellites per launch. 30M/50 = $600k/satellite.
Starship: Aspirational goal of $2M cost to launch. Let's just round that up to $10M for whatever might be more expensive than expected. Starship has been predicted to be able to launch around 400 Starlink satellites at a time, the last figure I saw was that Starship launches would carry "100 plus" Starlinks[2]. $10M/100 = $100k/satellite. This is assuming Starship is launching v2.0 satellites though, rather than the current v1.5, which are way heavier (1 metric ton vs ~290kg)[3].
OK, yeah so the video mentioned an optimistic price of $250k/satelite. $100k is definitely better! It sounds like $250k would be matching the v1.5 satelite cost.
I guess the overall thing would be to break things down and figure it out. I'm skeptical mainly from a "possible market share" perspective ($99/month works well for some, way less for others, and honestly the competition is not wired broadband but cell networks). But I am here to be proven wrong!
The idea is to use starship which should be really cheap.