Ask the federal freight data system to name the country’s most important corridor and it will decline, because none of its three main instruments is built to answer that question. The Federal Highway Administration designates a network of roads. The Freight Analysis Framework estimates flows between zones. The American Transportation Research Institute ranks the places where trucks lose speed. Those are three different questions, and no product from any of them contains an ordered list of corridors.

Every published ranking of American freight corridors was therefore assembled by whoever published it, out of components that were not designed to be ordered. That is worth knowing before reading one. It is also worth knowing which instrument answers which question, because each of the three is genuinely authoritative inside its own scope and misleading outside it.

The National Highway Freight Network is the closest thing the country has to an official statement of which roads carry freight importance, and it is built from four subsystems rather than one. FHWA’s current accounting puts the Primary Highway Freight System at about 41,799 centerline miles following the 2022 congressional redesignation, of which 38,014 miles are Interstate and 3,785 are not. Add the remaining Interstate mileage outside the primary system, estimated at 10,265 centerline miles, and the critical corridors that states and metropolitan planning organizations designate themselves, and the whole network reaches roughly 60,110 centerline miles.

The total is the least interesting figure in that paragraph. What shapes the map is a growth cap written into statute: each redesignation may raise the primary system’s total mileage by no more than 3 percent, and redesignation happens every five years. A corridor whose flows doubled between census years cannot enter the primary system on that evidence alone if the 3 percent envelope is already committed elsewhere. The initial designation was the 41,518-mile network inherited from the earlier primary freight network process, and the 2022 exercise moved it by fewer than 300 miles.

The state-designated pieces are capped by formula rather than by need. FHWA’s current network page states the ceilings as 300 miles of critical rural freight corridors per state, or 20 percent of that state’s primary system mileage, whichever is greater, with 600 miles available to states of below-average population density, and 150 miles of critical urban freight corridors, or 10 percent. Those ceilings have moved. FHWA’s February 2016 guidance implementing the freight program set them at 150 rural miles and 75 urban miles. The allowances doubled, which means the current map of state-designated freight corridors reflects an administrative allowance as much as it reflects freight.

Where the statute does get specific, it is instructive. A rural principal arterial qualifies as a critical rural freight corridor if at least 25 percent of its annual average daily traffic, measured in passenger vehicle equivalent units, is trucks in FHWA vehicle classes 8 through 13. Alternatively it qualifies by connecting the primary system to a facility handling 50,000 twenty-foot equivalent units a year, or 500,000 tons a year of bulk commodities. Those are real thresholds, and they are the only numeric tests of freight importance anywhere in the designation. Everything else on the list is a connectivity or judgment criterion, including a clause for roads a state simply determines to be vital.

As of January 2023 the designated critical rural mileage stood at about 5,390 centerline miles and the critical urban mileage at about 2,656. Against 41,799 miles of primary system, the state-designated layer is small, and it is the only layer where the 25 percent truck test applies.

What the Freight Analysis Framework measures, and what it will not rank

The Freight Analysis Framework is the flow model behind almost every corridor argument made in the United States, and its structure explains why it cannot settle one. FHWA’s description of the current version fixes a base year of 2017, benchmarked to that year’s Commodity Flow Survey, with forecasts at five-year increments to 2050 under three growth scenarios.

The mode split at the base year is the figure most often quoted from it, and the version 5.2 briefing FHWA gave in July 2022 carries exactly one table of it, headed as baseline growth for domestic modes. Its 2017 column reads: truck 11,848,259 thousand tons, pipeline 3,132,993, rail 1,202,016, water 662,453, multiple modes and mail 536,088, air including truck-air 2,136, and other and unknown 93,634, for a total of 17,477,579. Trucks carry just under 68 percent of that.

The heading on that table does more work than it looks like doing. The same briefing states that FAF5 forecasts cover domestic, import and export flows, and it charts total tonnage by flow type on a slide of its own, so a domestic-mode total leaves out trade flows the framework also estimates. Two analysts can quote a 2017 FAF tonnage, each accurately, and disagree because one of them included the imports. Naming the flow set is the difference between a figure another reader can check and one nobody can, a discipline the freight and transportation section applies to every tonnage it prints.

Inside the same table the trajectory matters more than the base-year share. Truck tonnage rises to 17,545,223 thousand tons by 2050, a compound rate of 1.2 percent, and pipeline grows at 1.0 percent against 1.1 percent for the domestic total. Rail tonnage falls, from 1,202,016 to 1,146,624 thousand tons, a compound rate of negative 0.1 percent. Trucks end the forecast at about 70 percent of domestic tonnage. A forecast in which the rail number declines in absolute terms over thirty-three years is a forecast about highway infrastructure whether or not it is read that way.

Three caveats travel with those numbers and none of them is decorative. The forecasts are driven by a commercial econometric system and incorporate economic data as of April 2021, so the vintage of the economics is now several years behind the vintage of the pavement decisions being justified with it. The scenario band is wide: relative to baseline, FHWA’s high growth case runs about 6 percent above at 2050 and the low growth case about 9.6 percent below. And FHWA’s own presentation of the framework’s limitations states that local roadways are not fully captured and that local-scale analysis is likely to require supplemental data.

The truck flow maps built on the framework are the closest FAF comes to a corridor product, and their own note defines what they show: freight moving by truck on highway segments carrying more than 25 FAF trucks per day, between places typically more than fifty miles apart. That is a map of long-haul corridors by construction. Short-haul movement, which is where most drayage and most urban freight sit, is outside the frame rather than absent from the road.

Where trucks actually stop moving

The one instrument that measures observed corridor performance rather than modeled flow is ATRI’s bottleneck ranking, which sits on a truck GPS dataset the institute states it has collected and processed for nearly 25 years in support of the U.S. Department of Transportation’s Freight Mobility Initiative. The 2026 edition, released in February 2026, evaluates more than 325 freight-significant locations using 2025 data and a speed and volume algorithm, then publishes the top 100 with average, peak and non-peak speeds attached.

Its top ten is a list of interchanges, not corridors. Chicago’s I-294 junction with I-290 and I-88 ranks first and Fort Lee, New Jersey, where I-95 meets State Route 4, ranks second. Then come Atlanta’s I-285 interchange with I-85 north, Houston’s I-45 at I-69 and US 59, Atlanta’s I-75 at I-285 north, Atlanta’s I-20 at I-285 west, Nashville’s I-24 and I-40 at I-440 east, Houston’s I-10 at I-69 and US 59, Cincinnati’s I-71 at I-75, and I-75 at McDonough, Georgia.

Three of the top six are on the same beltway. That is the most useful single fact in the ranking, and it is invisible in any corridor-level treatment, because the constraint is not a length of I-285 but a set of junctions on it. A ranking of locations and a ranking of corridors do not convert into one another, and the conversion is where most published lists go wrong.

Where the three instruments do agree

Read the three together and the same few places recur. The Texas Interstates carry designated primary system mileage and appear twice in ATRI’s top ten through Houston. Chicago sits at the top of the observed delay ranking and at the center of the FAF truck flow map. Atlanta’s beltway is a designated freight route, a rail and port relay point, and the most repeatedly congested single facility in the GPS record. The I-95 corridor through northern New Jersey has the second-worst measured bottleneck in the country attached to it.

What that reading cannot borrow from the July 2022 briefing is an order of states, because the briefing does not print one. It demonstrates its state-level output with a single state, giving Texas a 2017 total of 3,308,058,120 tons against 4,967,119,536 in 2050, and even that number arrives with a footnote recording that the total displayed does not include the mode category logged as other. The six plotted mode bars sum to roughly 3,216 million tons, so a reader cannot close the gap between the stated total and the modes behind it. Beyond that one chart the briefing’s other state-level examples are a through-flow map for Missouri and two tables of Kentucky flows. The geography FAF publishes is zones and states rather than corridors, so converting it into a corridor ranking is an act of interpretation the data does not supervise, and converting it into a state ranking requires a product that publishes one.

The ranked list exists, one state at a time

National corridor rankings are scarce because the federal instruments do not produce them. State rankings are not scarce, because state departments of transportation have both the traffic inventory and the reason to produce them.

Texas is the clearest case. TxDOT’s corridor plan for I-35 reports that the route holds 2,931 lane miles, about 18 percent of Texas Interstate lane miles, and carried approximately 24 percent of all vehicle miles traveled on Texas Interstates in 2014, more than any other Interstate in the state. It also reports that 13 of the 100 most congested roadway segments in Texas in 2015 were sections of I-35. Those are corridor-level comparisons within a defined system, computed from a single inventory, and they support a ranking claim that no federal product does. The corridor’s own case is set out in why I-35 matters to national freight.

The limitation is equally clear. A state ranking is bounded by the state, so a corridor that matters because of what happens 900 miles away ranks by its in-state mileage. Cross-state importance is exactly what a national instrument would capture and exactly what none of them orders.

Which instrument answers which question

For asking whether a road is eligible for federal freight money, the designation is the only instrument that matters, and it should be read as an administrative status with a mileage ceiling rather than as a finding about traffic. For asking what will move over the next twenty years and by which mode, the Freight Analysis Framework is the only national estimate available, and it should be read with its flow filters, its April 2021 economics and a scenario band roughly 16 percentage points wide at 2050 stated alongside the number. For asking where freight is losing time today, the GPS bottleneck ranking is the only measurement of the three, and it identifies junctions rather than corridors.

The practical consequence for anyone writing a project justification is that these three cannot be stacked into a single claim of importance. The evidence base for the intermodal legs is separate again, and is examined in how freight moves from port to highway, while the constraint that shows up on every long corridor once the driving hours run out is treated in the national truck parking shortfall. A corridor can be on the primary system, forecast to grow, and free-flowing. It can be congested at one interchange and unremarkable for 300 miles either side.