Reading it cynically/critically/literally, one could interpret it as counting users, not customers (so a family of four sharing a link would count as four users).
They also mention “thousands of customers”, so I guess that reading may be correct.
Also, as to economics, Wikipedia (https://en.wikipedia.org/wiki/Starlink) says there will be up to 30,000 of these satellites. Even if (from https://ecfsapi.fcc.gov/file/1020316268311/Starlink%20Servic...) “hundreds of thousands of individuals spread over diverse locations across all 50 states signed up to register their interest at Starlink.com even without any formal advertising” turns into a million paying customers, that’s about 30 customers per satellite. Can 30 families in rural locations (so, presumably, not in the highest income brackets) really afford to pay for launching and maintaining a satellite in low-earth orbit? (Real number of satellites probably will be lower, initially, but so will the number of paying customers)
Edit: thinking a bit more about it. These satellites weigh about 260kg. That would be 10-ish kg/customer. They fly, mostly, at 340km, or 30-ish as high as jets. Not trying to model this precisely (to say it mildly), let’s guess, with highly reusable rockets, it’s as expensive to move a kg up a km as with a commercial jet. If so, it’s 30 times as expensive per kg to bring such a satellite in its orbit as it is to use a jet airplane to take a kg to cruising height. So, each customer would have to pay the equivalent of moving 300-ish kg by long-distance jet. That’s 3 personal flights with luggage. Guessing at a satellite lifetime of 6 years, that’s ½ a flight a year. Seems affordable for many in the USA, even taking into account that the satellites have to be built and managed.