As late as 1959, per the World Almanac, the US ran over 450,000,000 rail passenger riders per year. Since the US had about 165 million people at the time- about half what the US has in 2013, adjusting that figure to the current population of the US, gives a population adjusted figure of 900,000,000.
This puts Amtrak's 35 million passenger riders per year into that perspective.
Consider first available rail cars: Amtrak has only 130 long distance train cars ordered. Assumming each of these new cars carried 100 people per day (and that no older cars were retired), 365 days per year, these cars would add about 4.75 million riders per YEAR,
Consider, too, what Amtrak has been saying the last 3 or 4 years ago how much government money would be needed to develope additional lines. Figures in the multiple 10s of millions per mile.
So what is the smarteset short term solution?
Do not add significant new routes to the system, but double the density of coverage per existing route
I live in metro Denver. Here, we have lost the future opportunity to connect via existing freight lines south of Lodo due to terrible long term planning. Fixing the issue, caused by recent, near sighted greed, would take billions of dollars.
So, we should forget south bound connections, and, run 2 Amtrak Zephyrs through Denver per day. Run them 6 to 12 hours apart. No track improvements needed. The convenience of Amtrak in Denver would increase as the square of the number of trains.
As running 2 trains per day along a 3 day route (averaged upward) would double the number of cars, we might be dealing with an adddition 35 - 40 cars, and, 4-5 locomotives (spares) for a cost of perhaps $100 million on passenger cars and $30 million for locomotives.
So, for long distance trains, we might be able to ball park this figure at $130 million or so per route (not including hiring and training of personnel).
On the 4 east/west routes that run from New Orleans-LA, Chicago-LA, Chicago-Oakland, Chicago-Seattle/Portland we are talking about capitol expenses on equipment of about $550 million.
There are other advantages in 'doubling' up. A huge advantage as that each train does not have to follow the same route over the entire distance. A second Empire Builder might always go to Portland via Spokane. A second Cresent might extend to LA along the route of the Southest limited or a portion thereof. A second Southwest Chief might connect through KC to St. Louis, This could be done WITHOUT TRAINS HAVING TO WAIT TO BE RE-ASSEMBLED MID-ROUTE, a problem that doomed splitting trains in early Amtrak days.
There is a huge political benefit to this, too. Seeing multiple frequencies on long distance runs would stimulate, IMO, local political interest. For example, two trains per day between San Antonio and Oklahoma City might encourage one the trains extending to Newton, KS, where the extended train might meet- with smaller delays- one of two Southwest Limiteds.
So, IMO, the best way to get Amtrak to keep growing would be to order 1000 or so passenger cars and 150 locomotives on top of existing orders, and, double longer distance frequencies. This would be 3.5 to 4 billion capital expense spread over 3 to 4 years, possibly financed- in this era of low interest rates- via federally backed bonds.
Then, the network expansion, IMO, would follow rapidly.