Amtrak released its fiscal-year-end financial and operating data last month. The passenger railroad posted an overall operating loss of $1.76 billion, which was $49 million less than the $1.81 billion loss it had shown for fiscal 2024. However, Amtrak chooses to ask Congress for money based on its “adjusted” operating loss, which excludes a billion dollars per year of depreciation, $364 million in planning costs for capital projects, and hundreds of millions of dollars in other expenses and revenues which are on the operating side of the budget under normal accounting rules but which Amtrak says are not as relevant to their need for annual federal operating subsidies.

The adjusted operating loss was $598.4 million, which was $106.8 billion better than the fiscal 2024 adjusted operating loss of $705.2 million. A large part of that was the amount of project planning expenses dropping by $83 million.

In more tangible terms, ticket revenue increased by $254 million, which was more than the increase in labor costs, train operations, and fuel/power/utilities combined (+$188 million).

Ridership was up across the board, with Northeast Corridor ridership up 8 percent over the prior year, state-supported route ridership up 2 percent, and long distance route ridership up 4 percent. However, increasing ridership is only good for the bottom line if you are making money on each passenger. If you are losing money per-passenger, then more passengers is not so good. In this case, the overall revenue per available seat-mile increased by 1.3 cents per ASM, while costs only increased by 0.3 cents per ASM. The cost recovery ratio went from 83.7 percent in 2024 to 86.8 percent in 2025, which was welcome news.

The Northeast Corridor trains increased their operating profit to $352.5 million in 2024, offsetting everything but the long-distance trains.

Route-level analysis

A full table of route-level results for fiscal 2025 are on page 6 of the monthly report. That data shows adjusted operating revenues and losses, riders, seat-miles, passenger-miles, and frequencies for each train route.

Using that data, we were able to provide more detailed analytics for most train routes. First of all is the division of the financial data into the seat-mile and passenger-mile data to get a sense of costs, revenues, and profit/loss per available seat-mile (ASM) and per passenger-mile, which are metrics used by the airline industry.

Meanwhile, costs on the Sunset Limited and the Capitol Limited somehow approach $1.00 per passenger-mile. (The airline average is around 18 cents per mile.)

Then we have other derived data – the load factor (how full the average train is), ticket revenue per passenger (in average dollars and per passenger-mile – for NEC and long-distance trains, ticket revenues are very very close to total revenues, but not for state-supported routes, where the state putting in a large subsidy for the operating service is part of the plan), average passengers per train, and average miles per passenger.

The average miles per rider metric helps understand long-distance trains. Yes, the full Southwest Chief route from Los Angeles to Chicago runs 2,256 miles, but the average rider only takes it 855 miles, or 38 percent of the way. (Sort of Chicago to Dodge City.)

Two updates that make the above data make a bit more sense:

• If the numbers for the Capitol Limited and the Silver Star look a bit off, it is because Amtrak eliminated those trains mid-year and replaced them with a new Chicago-Miami train, the Floridian, which stops in D.C. to let Northeast Corridor passengers on. The combined train is supposed to reduce demand on NYC-area tunnels until those replacement tunnels are complete.
• Last month, Amtrak changed the name of the San Joaquins train, and it is now the Gold Runner, so look for that change in future reporting.