I frankly don't see any evidence that Anderson has any depth of understanding of DMUs and their operating characteristics at all. He seems to have read a few rail rags from Europe and come to the conclusion that DMUs are the future. He talked about getting DMUs to replace Amfleet Is?
He seemed to forget that most Amfleet 1s are used on the NEC where his outfit just acquired 75 new electric engines to pull trailer cars, and diesel anything would not work in the most heavily used anchor station of the NEC.
After he had stopped smoking whatever potent stuff he uses, and come down to earth with a coherent plan, instead of the equivalent of the one line tweets, it would be worth having a serious conversation.
There is a work around for that, in the airline business and some extent railroad business, one negotiates a trade in of present asset. Amtrak did it with the SD40Fs on new F40s. Siemens would take that deal in a heart beat for 5 years or more guaranteed work and bigger contract.
How did we arrive at a conclusion that DMUs are cheaper to purchase? Seems not to really align with the realities we have seen so far.
I stated cheaper to operate, it might be cheaper to purchase if Amtrak bulk buys and replaces the Amfleets and some Genesis with one order. And Anderson missed his opportunity to convince the states to buy DMUs instead which could have made the purchasing block for Amtrak even cheaper.
DMUs are cheaper to operate as long your trains are shorter than 5 or 6 cars. After that it becomes quite questionable. They will work in many light corridors that we have today outside of the two coasts. They are not a panacea for all situations.
Then run two DMUs together. Before that though it allows Amtrak to have better yield management, just as Amtrak does with the frequently sold out Acela. You are right, DMUs would not be right for all circumstances but having one or fewer types of equipment yields lower costs. Again out of the airline playbook.