Part of the reason I think queues aren't too bad is that if you have, say, 10 stations, a person is mostly just filling up then paying, so the stations don't have a lot of downtime, and flow rate is not an issue.

Still, that's a need to maintain and operate 10 pumps, when it sounds like 28 days out of the month it won't need more than 2, maybe 4. They're building out the ability to serve peak demand, despite average demand being so much lower.

But so long as pumps aren't expensive, and tanks aren't expensive, building and maintaining for peak demand isn't hugely costly.

Electrical infrastructure is different, of course. It's much more complex to build out a 10 MW superstation than a single 1 MW station. And then the individual chargers are pretty expensive, too, so building out 10 stalls is gonna be really expensive, too.

If the bottleneck is the "pipe" bringing in 1MW, but each stall actually supports up to 350 kW or something, it could be configured that there isn't any wasted capacity in the aggregate (total charge being delivered is still 1 MW across 10 chargers, even as some drop to zero as the vehicle approaches full charge, or there are vehicles being swapped at that spot). Add in some on-site storage of something like 4 MWh so that you can flexibly add 1 MW of charging capacity (for a total of 2MW, without necessarily needing to ever pull more than 1 MW from the grid) for 4 hours, and that might take care of the surge demand in a reasonable way, too.

Maybe even throw in surge pricing of a per-minute use of the stall, so that people will be incentivized not to leave their vehicle longer than they need and unnecessarily delay the queue.

Still, none of that is cheap. I'd expect a charging station that implements those things to be more expensive than an ordinary city charger, to accommodate that significantly higher need for complex equipment and processes.