Alright, I did it, based on OpenFlights route data (https://openflights.org/data.html). The list is:

Albuquerque, Atlanta, Austin, Boston, Charlotte, Chicago, Dallas-Fort Worth, Denver, Detroit, Houston, Kansas City, Las Vegas, Los Angeles, Miami, Milwaukee, Minneapolis, Orlando, Philadelphia, Phoenix, Portland, Sacramento, Salt Lake City, San Antonio, San Diego, St. Louis, Toronto.

Both Chicago airports are on the list, as are Fort Lauderdale (FLL) and Long Beach (LGB), but this is about metro areas, not airports. I also manually filtered out Honolulu.

So long as airport infrastructure is in place, it's trivial for airlines to add routes based on demand. There are a lot of airports, such as former Air Force bases, that could handle significantly increased demand. You can find a lot of fairly large city pairs without non-stops but that's usually because neither are hubs and there isn't quite enough demand.

That's true. But is Amazon going to be big enough to create enough demand that airlines would make those additions? (Not a rhetorical question.) Sure, Pittsburgh could have a flight to Seattle. But it doesn't right now. Is the increased demand from an Amazon HQ2 in Pittsburgh enough to convince Alaska or Delta to start running those flights?