Amtrak's Long-Distance Trains: What It Costs to Keep Ru...

There is a train that runs three days a week between New Orleans and Los Angeles, and in a recent year the federal government spent roughly 578 dollars in subsidy for every passenger who rode it. There is another network of trains, in the crowded corridor between Washington and Boston, that actually turns an operating profit and hands a few hundred million dollars a year back toward the rest of the system. Both belong to the same company. The gap between them is the whole story of how America pays for passenger rail, because the profitable trains and the money-losing ones answer two different questions, and Amtrak is legally required to run both.

What it is, and who actually runs it

Amtrak is the trade name of the National Railroad Passenger Corporation, and it exists because Congress decided in 1970 that intercity passenger trains should not be allowed to disappear. By the late 1960s the private freight railroads that had always carried passengers were losing money on every ticket and wanted out. Rather than let the service vanish, Congress passed the Rail Passenger Service Act of 1970, which created Amtrak to take over intercity passenger operations while relieving the freight railroads of the obligation, according to the Congressional Research Service. Amtrak began running trains on May 1, 1971.

That founding bargain still shapes everything. Amtrak was never designed to be a profit-making company; it was designed to preserve a service the market had already voted to abandon. Its statutory goals, set out in 49 U.S.C. 24101 and summarized by the Congressional Research Service, include operating a national rail passenger system, and the exact contours of that mandate have been rewritten by Congress several times since. The general requirement to run a national network is the load-bearing fact here. It is the reason the money-losing trains exist at all.

The system splits into three pieces that are worth keeping straight. The Northeast Corridor is the electrified spine from Washington through New York to Boston. The state-supported corridors are shorter regional routes whose operating costs the states now pay, a shift Congress ordered in the Passenger Rail Investment and Improvement Act of 2008. And the long-distance network is the set of roughly 15 routes over 750 miles, as defined by the Federal Railroad Administration, running anywhere from about 760 miles to the roughly 2,438 miles of the California Zephyr between Chicago and the San Francisco Bay. These are the cross-country trains, and they are where the subsidy argument lives.

Where the money goes

The headline numbers are healthier than they used to be. Amtrak carried a record 32.8 million customer trips in fiscal 2024, up 15 percent over the prior year, by its own accounting, and then broke the record again in fiscal 2025 with 34.5 million trips, up about 5 percent, along with 6.9 billion passenger-miles, also an all-time high. Ridership is climbing and revenue is climbing with it.

The company still loses money every year, and here the accounting needs care, because Amtrak reports its loss two different ways and they are easy to confuse. On an adjusted operating basis, the measure Amtrak leads with, the systemwide loss was about 705.2 million dollars in fiscal 2024 and narrowed to about 598.4 million dollars in fiscal 2025. On a full-accounting basis that includes depreciation and interest, the same years show a total operating loss of about 1.81 billion dollars and 1.76 billion dollars respectively, in the analysis of the Eno Center for Transportation, a transportation think tank that works through Amtrak's own monthly performance reports. Those are two measures of the same year, not two years, and the larger one simply counts more of the true cost. Both are systemwide figures covering every train Amtrak runs.

That last point is the one most commentary gets wrong. Amtrak's published loss is not a long-distance loss. To see what the cross-country trains cost on their own, you have to go down to the route level, and that is where the picture sharpens.

The long-distance network carries only a small slice of the riders. Long-distance trains accounted for about 4.2 million trips in fiscal 2024, roughly 13 percent of the total. But because each of those trips is long, the network's share of passenger-miles is far larger than its share of trips, and its share of the deficit is larger still. A minority of the riders drives a majority of the loss.

Why it looks like a loss

Set the long-distance trains against ordinary efficiency and they fail, badly, on every measure.

Start with the loss per passenger-mile, the cleanest way to compare routes of different lengths. In fiscal 2025 the long-distance service line as a whole earned about negative 0.266 dollars per passenger-mile, in Eno's analysis of Amtrak's route data. The worst routes were far deeper in the red:

Sunset Limited: about negative 0.752 dollars per passenger-mile, the worst in the system. It runs only three days a week.
Cardinal: about negative 0.577 dollars per passenger-mile, also a tri-weekly train.
Southwest Chief: about negative 0.325 dollars per passenger-mile.

On the very worst routes the total cost approaches a full dollar per passenger-mile, against roughly 18 cents for the airlines, by Eno's comparison. That is the efficiency case in a single line: a mode that costs several times what flying costs to move a person the same distance.

Switch to the per-passenger measure and the numbers get more visceral, though they are a different metric and belong to a different year. Working from Amtrak's fiscal 2023 data, Eno found the long-distance routes lost about 150 dollars per passenger, up from about 104 dollars in fiscal 2019. The Sunset Limited carried a federal subsidy of about 578 dollars per passenger and the Southwest Chief about 322 dollars. For scale, the systemwide loss that year worked out to about 27 dollars per passenger. The long-distance routes are not a little worse than the average. They are several multiples worse.

The trains are also chronically late, and the tri-weekly schedules of the Sunset Limited and the Cardinal make the fixed overhead per passenger worse than a daily train would carry. No long-distance route met the Federal Railroad Administration's 80 percent on-time standard in fiscal 2024. The largest single cause of delay was freight train interference, which Amtrak says accounted for about 850,000 minutes of delay in 2024. That points to a genuine structural problem, which we will come back to, but on the raw efficiency ledger it is one more entry against the trains.

Then there is the contrast that makes the whole argument concrete. The Northeast Corridor, the same company's flagship, earned an operating profit of about 352.5 million dollars in fiscal 2025, or roughly positive 0.131 dollars per passenger-mile. In fiscal 2023 Acela earned about 34 dollars of operating profit per passenger and the corridor's other trains about 12 dollars. One part of Amtrak covers its operating costs and more; another part loses hundreds of dollars a rider. Same company, same era, opposite economics. If the only question were whether a route pays for itself, the long-distance trains would already be gone.

Ratio-adjusted for the towns the trains serve

Now change the question, and the ledger flips.

The first thing to notice is what the Northeast Corridor's profit does not include. The corridor is operationally profitable, but it sits on aging tunnels, bridges, and catenary with billions of dollars in unmet capital needs, and its operating surplus comes nowhere near covering them. Operational self-sufficiency, which the Congressional Research Service also attributes to the corridor, is a real distinction, but it is not the same as paying its own way in full. Even the profitable part of Amtrak leans on federal capital. So the clean line between the self-funding corridor and the subsidized long-haul is blurrier than the operating numbers suggest.

The second thing is what the long-distance trains are for. They were never meant to compete with the airlines on cost per mile. They were meant to be the national network, the coast-to-coast connective tissue that the 1970 statute required, and by that measure they do a great deal. The 15 long-distance routes serve stations in 39 states and provide service at nearly half of all Amtrak stations, across a system that reaches something over 500 stations in 46 states, the District of Columbia, and several Canadian provinces. For many of the towns on those routes, the daily train is not one option among several. It is the only scheduled intercity public transportation there is: no commercial airport, no intercity bus, nothing else that runs on a timetable.

Advocates put a number on that reach. The Rail Passengers Association, a pro-rail lobbying group, estimates the national network serves roughly 40 percent of the nation's small and rural communities. That figure comes from an advocacy source and is not tied to a specific government census count, so it belongs in the case rather than at its foundation, but it captures the shape of the argument. The riders who depend on these trains skew older, lower-income, and less able to drive long distances, and for them the alternative to the train is often no trip at all.

Measured that way, the per-passenger loss looks less like waste and more like the price of a mandate. Much of what the fully allocated cost assigns to a lightly used route is fixed national-network overhead that would not disappear if the train did. And the routes are not optional in the way a business line is optional; Congress requires the national system to exist. Judging a legally required lifeline by whether it clears a private-sector return is, the defenders argue, using the wrong yardstick on purpose.

The freight-delay problem cuts the same way. Most long-distance miles run on track owned and dispatched by freight railroads, and federal law has given Amtrak's trains preference over freight for more than fifty years. That preference is routinely ignored, which is why freight interference tops the delay charts. The Passenger Rail Investment and Improvement Act of 2008 tried to add teeth: it set the 80 percent on-time standard and, in its Section 213, empowered the Surface Transportation Board to investigate host-railroad preference violations. The point for the ledger is that the long-distance trains' worst operational failing is largely imposed from outside, by the freight owners of the rails, not manufactured by Amtrak.

Congress has continued to treat the network as worth keeping and even worth growing. The Infrastructure Investment and Jobs Act of 2021 provided about 66 billion dollars in guaranteed advance appropriations for rail over fiscal 2022 through 2026, including about 22 billion dollars in grants to Amtrak and about 44 billion dollars for Federal Railroad Administration discretionary rail programs, the largest federal rail commitment in the company's history. One caveat matters enormously here: that money is overwhelmingly restricted to capital, new trains and fixed infrastructure, and by law it cannot substitute for the annual operating grant that actually keeps the daily trains running. The infrastructure law rebuilds the network; it does not pay to operate it. The two accounts are separate on purpose.

The same law ordered the Federal Railroad Administration to study expanding the long-distance network, and the resulting report frames the connectivity case in numbers. Its preferred set of conceptual routes could serve 34 states and give about 39 million people new access to passenger rail, including roughly 7 million in rural communities. Those are figures for potential future routes that are unfunded and conceptual, not the current 15-route network, so they describe an ambition rather than a reality. The study also recommends restoring daily service on the tri-weekly Sunset Limited and Cardinal, which would spread the fixed cost of those routes over more riders and could improve the very per-passenger figures that make them look worst today.

The ledger reading

Amtrak's long-distance trains are, at the same time, indefensible and defensible, because the two verdicts answer two different questions.

Ask whether a cross-country train pays for itself, and the answer is a flat no. On a route where the subsidy runs to hundreds of dollars a passenger and the cost per mile is several times an airline seat, the efficiency case is not close, and the honest way to state it is that these are the least cost-effective trains in the system by a wide margin. Ask instead whether the country should have a scheduled public way to reach the towns these trains serve, most of them with no airport and no bus, and the market has already answered: it will not provide one at any price, which is why Amtrak exists in the first place.

What holds the whole thing together is that the profitable and the unprofitable trains are bound into one statutory network. The Northeast Corridor cannot be the entire company, because the law requires a national system, and the national system is the part that loses money. Amtrak projects that its train operations will reach adjusted operating profitability around 2028, but that is a company projection, it refers to the operating measure rather than full cost including depreciation and interest, and it does not change the underlying structure. The long-distance trains will keep showing up as the largest loss on the ledger and the largest coverage on the map, and which of those two numbers you lead with will tell you, before you ever open the performance report, what you already believed passenger rail is for.

Fact-check notes and sources

Amtrak created by the Rail Passenger Service Act of 1970, began operating May 1, 1971, to preserve intercity rail while relieving freight railroads of passenger service; national-network goals set in 49 U.S.C. 24101; the corridor is operationally self-sufficient while the national network runs a deficit: the Congressional Research Service overview of Amtrak. The exact statutory subsection wording of the national-network mandate was not independently verified against the U.S. Code text and is cited here in general terms.
Roughly 15 long-distance routes over 750 miles, serving 39 states and nearly half of Amtrak's stations, ranging up to about 2,438 miles: the Federal Railroad Administration's long-distance service study page. The broader "500-plus stations in 46 states" figure is standard Amtrak system description; exact station counts vary by year.
Record ridership of 32.8 million trips in fiscal 2024 (up 15 percent), 4.2 million long-distance trips (about 13 percent of the total), and a 705.2 million dollar adjusted operating loss: Amtrak's fiscal 2024 results. Record 34.5 million trips in fiscal 2025 (up about 5 percent), 6.9 billion passenger-miles, and a narrowed 598.4 million dollar adjusted operating loss: Amtrak's fiscal 2025 results. These loss figures are systemwide, not long-distance only.
Full-accounting total operating loss of about 1.81 billion dollars (fiscal 2024) and 1.76 billion dollars (fiscal 2025); long-distance service line at about negative 0.266 dollars per passenger-mile in fiscal 2025, with the Sunset Limited at negative 0.752, the Cardinal at negative 0.577, and the Southwest Chief at negative 0.325; the worst routes near a dollar per passenger-mile against roughly 18 cents for airlines; and the Northeast Corridor at positive 0.131 dollars per passenger-mile with about 352.5 million dollars of operating profit: the Eno Center for Transportation's analysis of Amtrak's fiscal 2025 performance data, a transportation think tank working from Amtrak's own monthly reports. The adjusted and total operating losses are two measures of the same year, and the larger includes depreciation and interest.
Per-passenger figures for fiscal 2023: about 150 dollars lost per long-distance passenger (up from about 104 dollars in fiscal 2019), about 578 dollars of subsidy per passenger on the Sunset Limited, about 322 dollars on the Southwest Chief, and about 27 dollars systemwide; Acela and other Northeast Corridor operating profit per passenger: the Eno Center for Transportation's analysis of Amtrak's fiscal 2023 data. Per-passenger and per-passenger-mile losses are different metrics from different years and are not interchangeable.
No long-distance route met the 80 percent on-time standard in fiscal 2024, and freight train interference caused about 850,000 minutes of delay in 2024 as the single largest delay cause on host-railroad track: Amtrak's on-time performance page. The 80 percent standard and the Surface Transportation Board's preference-violation authority come from the Passenger Rail Investment and Improvement Act of 2008, per the Congressional Research Service.
Infrastructure Investment and Jobs Act rail funding of about 66 billion dollars in guaranteed advance appropriations for fiscal 2022 through 2026, including about 22 billion dollars in Amtrak grants and about 44 billion dollars in Federal Railroad Administration discretionary grants, restricted to capital: the Federal Railroad Administration's infrastructure law page.
Potential expansion serving 34 states and giving about 39 million people new rail access, including roughly 7 million in rural communities, plus a recommendation to restore daily service on the Sunset Limited and Cardinal: the Federal Railroad Administration's Amtrak Daily Long-Distance Service Study final report. These describe conceptual, unfunded future routes, not the current network.
The national network serving roughly 40 percent of small and rural communities: the Rail Passengers Association's rural mobility fact sheet, a pro-rail advocacy group; the figure is not tied to a specific government count and is labeled here as advocacy.

This post is informational and journalistic, describing a public corporation and public records. It is not legal, financial, or policy advice. Figures are drawn from government reports, a congressional research service, a company's own disclosures, and one transportation think tank's analysis of those disclosures; where advocacy sources are cited, they are labeled, and where two accounting measures of the same figure exist, both are named.

Amtrak's Long-Distance Trains: What It Costs to Keep Rural America on the Rails