1Executive summary

In September, the first paper in this series, Capturing the Merchandise Carload Market, documented two decades of lost share. Merchandise carload, all non-intermodal carload traffic other than coal, was about 8.4 million carloads in 2024 and has not grown since 2008, while truck tonnage grew by about a third. That paper argued that the Class I railroads are configured to hook and haul, moving large blocks and unit trains between major terminals, and that the gathering work merchandise carload requires belongs to shortlines and terminal operators. It also argued that the gathering pays only if the Class I supports it with a reliable interchange and rates set for lanes where truck competes.¹

This paper asks the next question: as conditions turn in rail's favor, who captures the value? Four developments bear on the answer. The proposed Union Pacific–Norfolk Southern merger has moved to a full review and would extend the hook-and-haul model across the continent. Diesel has risen sharply, improving rail's cost position against truck for the freight Part 1 identified as recoverable. New heavy industry is being built on rail. And transaction multiples for North American railroads have risen from 7.5x to 11.6x EBITDA for shortline and financial buyers before 2013 to roughly 13x to 22x since 2017.

Our argument is that the gathering function described in Part 1 has become the scarce and valuable part of the network. We call it access capacity: the local service and transload capacity that gather freight, the ready sites that create new freight, and the interline terms that connect both to the Class I network (Exhibit 1). Consolidation will make the Class I railroads less likely to supply it themselves. Whoever builds and controls access capacity will capture most of the value created as merchandise carload returns to rail. Owners who acquired existing traffic at today's transaction values, without the ability to add access capacity, will struggle to earn their cost of capital.

The evidence points the same way across the network. On the main line, the merger and the conditions other railroads are seeking will decide the terms on which traffic moves between systems, and for shortlines the rates charged at gateways will matter as much as whether gateways stay open. At the first and last mile, the rise in diesel since last fall has added about $5.40 a ton more to the fuel cost of a 500-mile truck move than to the same move by rail and transload, roughly the cost of handling a ton of dry bulk through a transload. On the land side, the Class I railroads list thousands of properties near their track but certify only a few dozen as ready to build on, and closed industrial plants with track in place have become the largest reservoir of land that can be made ready quickly. In the capital markets, a buyer paying 17 times EBITDA needs EBITDA to grow about 6% a year to earn a 10% unlevered return under our assumptions, more than twice the 2.7% growth in U.S. carloads in what has been a strong year.

The implications extend Part 1's division of labor. All shortline and terminal operators should treat access capacity as their core business. Those that connect with Union Pacific or Norfolk Southern, or that depend on traffic interchanged with either railroad at a gateway, should also secure terms before the November 18 deadline for requests for conditions in the merger proceeding. The Class I railroads already rely on their shortline and terminal partners as their gathering arm; consolidation makes it more important that the interchange terms behind that relationship carry through the merger. Shippers should lock in rate and volume commitments before a line they depend on changes hands. The public sector and its development partners should invest in readiness, not acreage. Capital providers should underwrite an operator's ability to add carloads rather than the traffic it already has, and recognize that terminals and land can add EBITDA at lower transaction values than another railroad.

Evidence and judgment. As in Part 1, every figure is drawn from the public record and cited, with one exception: the aggregate of 17 transactions our team reviewed, advised on or led, which we report without naming the deals. Figures are as published by each source as of early October 2026. The interpretation and recommendations are the authors' judgment, based on experience owning and operating shortline railroads and terminals, and are not investment, legal or financial advice.

2Building on Part 1

Part 1 made the operating case for recovering merchandise carload and set out how an operator would execute it. This paper carries each of its central findings into the market as it stands in late 2026: the merger, fuel prices, new industrial projects and the prices paid for railroads. Exhibit 2 shows how each finding is extended and where this paper takes it up.

Two of Part 1's findings carry through this paper unchanged. Service reliability, as much as price, decides whether freight stays on rail. And no amount of gathering pays unless the Class I hauls the block on a schedule and at a rate that leaves the shortline, the terminal and the shipper a saving over truck.

3Consolidation raises the stakes at the interface

3.1 Consolidation will extend hook and haul across the continent

Part 1 described how the Class I railroads, under precision scheduled railroading, ran longer trains through fewer terminals and reduced the work of gathering individual cars; four of the seven told the Government Accountability Office they reduced how often they serve some smaller customers.² The Union Pacific–Norfolk Southern merger would extend that model across the continent. The Surface Transportation Board accepted the revised application in May 2026, and on September 18 turned down motions from BNSF, CSX and five shipper associations to deny it without a full review, noting that its decision "does not reflect any determination on the merits".³ The record will close in 2027 (Exhibit 3).

The case for the merger rests on single-line service between the coasts. The applicants project about $2.75 billion a year in synergies⁴ and say they will move 2.1 million truckloads a year to rail.⁵ The other large railroads have made clear how they read it. BNSF, CSX and CPKC oppose the merger and are seeking access in the corridors where they would face a single-line competitor; CN settled with Union Pacific in July in exchange for terminal stakes and new rights.^6,7 Exhibit 4 summarizes their positions.

Whatever the Board decides, the direction of the industry is clear. Based on 2024 results, a combined UP–NS would have about 56% more revenue than BNSF.⁸ BNSF and CSX already run joint intermodal service from coast to coast,⁹ although Union Pacific's chief executive has said such voluntary agreements "always break down".¹⁰ Berkshire Hathaway has said it is not looking to buy another railroad,¹¹ while CSX's chief executive has said he is open to "any path that can create shareholder value".¹² The Board itself asked the applicants whether conditions imposed in this case would still work after a later merger.¹³ If the merger is approved, owners and shippers should plan for an industry of two large systems within several years, each even more focused on long-haul density than today's railroads, and each with less reason to do the gathering itself.

3.2 The terms at the interchange are now in play

Part 1 concluded that the work of recovering merchandise carload is the same whatever the merger outcome, because removing an interchange helps freight already on rail and does nothing for the plant with no siding or the shipper with no cars. That remains true of the work. It is not true of the terms on which the work is paid. As the Class I network consolidates, the rates and divisions on traffic passing from a shortline to a Class I railroad become a larger share of the value of that traffic.

The applicants have proposed Committed Gateway Pricing for interline carload moves between facilities served only by UP or NS and facilities served only by BNSF or CSX, including shortlines that connect only with those carriers.¹⁴ Opponents, including the Stop the Rail Merger Coalition, said the proposal covers less than 1% of traffic and expires in five years, and CN, before its July settlement, cited the Board's finding that it applies to "only a tiny fraction of rail traffic".¹⁵

The risk for shortlines is less that gateways close than that the rates across them change. BNSF's chief executive told shortline owners that Union Pacific's volumes have fallen 13% over the last ten years while its revenue per unit has risen 37% more than at the other Class I railroads.¹⁶ BNSF is an interested party, but the concern is reasonable: a railroad that controls both ends of a long move has less reason to price a shortline's interline traffic for a lane where truck competes. As Peter Swan of Penn State Harrisburg put it, "keeping gateways open is not the same as keeping joint and local rates at current levels".⁸

History also suggests that conditions are easier to win than to enforce. The Union Pacific–Southern Pacific merger gave BNSF about 4,000 miles of trackage rights.¹⁷ In December 2025, BNSF said Union Pacific had denied or delayed 69 of the 200 requests for access to customers it had submitted since that merger,¹⁸ and the Board was still setting terms at Lake Charles this year.¹⁹ Shortlines and shippers that seek conditions should seek reporting and an enforcement mechanism with them.

3.3 Integration will release traffic and assets to the edges

Large mergers also change what the merged railroad chooses to do itself. The UP–SP merger was followed by a service crisis in 1997 and 1998,¹⁷ which congressional inquiries estimated cost about $4 billion, as cited by the American Farm Bureau.²⁰ UP and NS have offered shippers temporary access to another railroad if service declines during integration.²¹ The synergies the applicants project will come partly from consolidating yards and lines, and the truckloads they plan to win will need transloads, local switching and first- and last-mile service that a coast-to-coast railroad is unlikely to provide itself. In our experience, large mergers are followed by sales of branch lines and terminals. Operators with the capability to run them should expect opportunities of that kind in 2027 and 2028 if the merger is approved.

4Returning demand runs through the gathering layer

Part 1 identified three groups of merchandise freight that rail can recover: rail-served plants that truck part of their own outbound, truck shippers within dray range of a rail terminal, and new industrial sites now being planned. Each is growing for a different reason, and each reaches a Class I train only through the access layer.

4.1 The recovery is broadening, with a long lag from new sites to carloads

Part 1 reported carloads excluding coal up 4.5% in the first quarter of 2026, the highest first-quarter level since 2015.¹ The upturn has continued into the second half, although total carload growth has eased slightly, from about 3% after 26 weeks¹ to 2.7% through late September. U.S. rail traffic is up 3.5% this year; intermodal leads, and in recent weeks the carload gains have come from industrial freight such as chemicals, metals and minerals.²²

The third group, new industrial sites, is the largest in the long run and the slowest to arrive, and the pipeline is broad. The largest projects now under way span steel, chemicals and vehicles, and each is designed around rail. Mesabi Metallics announced a $15 billion integrated steel mill in Lee County, Iowa, in September, designed for up to 10 million tons a year when fully operational, with first steel targeted for 2030 and iron ore moving to the site by rail from Minnesota.²³ Hyundai Steel and its partners broke ground this year on a $5.8 billion, 2.7-million-ton electric arc furnace mill at Donaldsonville, Louisiana, on a site with access to a Union Pacific main line; finished coils will ship by rail and truck, and production is targeted for 2029.^24,25 Chevron Phillips Chemical and QatarEnergy's $8.5 billion Golden Triangle Polymers plant in Orange, Texas, which entered commissioning this year, will ship its polyethylene and receive feedstock by rail through a new on-site rail yard.^26,27 In November 2025 Norfolk Southern completed a rail bridge connecting its network to Scout Motors' new vehicle plant in Blythewood, South Carolina.²⁸ Behind these are the Class I railroads' broader pipelines: Norfolk Southern's customers committed $7.7 billion to more than 60 rail-served projects in 2025, and BNSF completed 117 rail-served projects with more than $5.3 billion of customer investment, its most in six years.^29,30 Announced investment takes years to become carloads: Norfolk Southern's volume fell 4% in the same year its customers committed $7.7 billion.²⁹ The owners who capture this demand will be those who prepare sites and capacity before it arrives, not those who wait for it to appear in the traffic data.

4.2 Diesel has reset the economics for the first two groups

The first two groups are the ones Part 1 expected to be won fastest, and both turn on the cost of the truck alternative. EIA's weekly diesel average, the index most fuel surcharges use, was $6.38 a gallon on September 28, up from $3.75 a year earlier.³¹ Railroads move a ton of freight nearly 500 miles per gallon on average and are three to four times more fuel-efficient than trucks,³² so a rise in diesel adds more to a truck move than to a rail move of the same length.

Fuel surcharges show how the increase reaches shippers. Van fuel surcharges averaged about 40 cents a mile through most of 2025;³³ in the week of September 20–26, fuel accounted for about 84 cents of DAT's all-in van spot rate.³⁴ Union Pacific's carload surcharge, set from the same index, was 33 cents a car-mile in October 2025, and its published formula puts November 2026 at about 84 cents.³⁵ Because a railcar carries about four truckloads,³⁶ the same per-mile increase is spread over four times the tonnage. On a 500-mile move with a 30-mile dray at destination, the fuel surcharge rose about $9.20 a ton by truck and about $3.80 a ton by rail and transload (Exhibit 5). The difference, about $5.40 a ton or roughly $500 a carload, is in the range a transload typically charges to handle a ton of dry bulk, $3 to $8.³⁷

For the second group, truck shippers within dray range of a terminal, the effect is direct. Because transload costs are mostly fixed per ton while linehaul savings grow with distance, higher diesel shortens the distance at which a transload pays, and lanes that did not work last year should be priced again. Because the dray is the part of the move that still pays truck fuel prices, a transload close to its customers is worth more than one farther away. For the first group, rail-served plants that truck part of their outbound, the same arithmetic applies to the share they now send by truck, which makes Part 1's customer audit worth repeating at current fuel prices. In both cases, because rail surcharges lag truck surcharges by about two months, rail looks cheapest at the peak; with EIA forecasting retail diesel to average $5.07 a gallon in 2026 and $4.40 in 2027,³⁸ operators should price contracts that still work at the lower figure.

4.3 Price will win the freight; service will keep it

Fuel alone will not hold the freight. During the 2008 fuel spike, U.S. carload originations rose 0.2% in the first half and fell 2.2% for the year as the recession took hold, and about half of carload non-users named service problems as the obstacle while only 14% named price.³⁹ Union Pacific's chief financial officer said in September that the railroad was beginning to see customers shift to rail because of its fuel-efficiency advantage.⁴⁰ The freight that stays after diesel falls will be freight that received what Part 1 described as the core of the gathering offer: a published local schedule, car supply and one price for the whole move.

Operators and the Class I railroads are already investing in that offer. OmniTRAX launched OmniTRAX Connect, a transload program across its 39 railroads, naming customers including The Home Depot, Clorox and CMC Steel.⁴¹ CN opened a second Battle Lumber transload in Wisconsin after rebuilding track, crossings and utilities at a former rail yard.⁴² Union Pacific added 15 Focus Sites in 2025, 12 of them on shortlines,⁴³ and BNSF's one disclosed new Certified Site of 2025 was the first served by one of its partner shortlines.³⁰ The Class I railroads are increasingly relying on others to deliver the first and last mile, consistent with the division of labor Part 1 described.

5Access capacity is scarce

5.1 Ready sites, not land, are the constraint

Part 1 named the loss of rail-served industrial sites as one of the causes of lost share, and rail-served industrial development as one of the services an operator can provide. The constraint today is not land beside track but sites that are ready to build on. Union Pacific lists more than 2,000 properties within 800 meters of its lines,⁴³ Norfolk Southern more than 800 rail-served properties,²⁹ and CSX serves more than 1,110 sites.⁴⁴ The number each has certified as ready is much smaller: 6% of CSX's sites carry its Select Site designation, Union Pacific has 39 Focus Sites, and Norfolk Southern earned 15 REDI designations in 2025 (Exhibit 6). The other Class I railroads report certified sites but not a count of all properties near their lines. In May 2026 CN added five U.S. Certified Rail-Ready sites and recertified six, about 3,280 acres in all,⁴⁵ and in March CPKC announced 14 Site Ready locations in the U.S., Canada and Mexico with more than 6,000 acres.⁴⁶

The difference between land and a ready site is time, cost and a service commitment. BNSF's standard process for a new industry track takes about 57 weeks from first review to service.⁴⁷ Industrial track costs upward of $1 million a mile,⁴⁸ and upgrading a signalized road crossing can take up to two years.⁴⁹ A developer cannot promise a tenant service that the railroad has not committed to provide, so many sites beside track are marketed as truck-served or not at all. Our 2025 paper, Rail-Served Industrial Sites & Greenfield Development,⁵⁰ found little new rail-served construction over the past two decades.

This is where the shortline's position is strongest. A shortline decides its own switching schedule and can commit to serving a new customer, which a Class I railroad often will not do for a small shipper. The value of a rail-served site lies in its readiness: zoning, utilities, track design and that service commitment. Whoever completes them captures most of the value of the land. Large developers are building rail parks in the biggest markets, such as Hillwood's nearly 1,400-acre Alliance Logistics District with BNSF in Fort Worth⁵¹ and Union Pacific's Mainline Texas Industrial Park near Houston.⁵² The shortage will last longest in smaller markets and for single-user sites, which is where most shortlines operate.

5.2 Closed industrial plants are the largest reservoir of ready land

The fastest way to add ready sites is often to reuse an industrial site that already has them in all but name. Closed mills and plants are already zoned for heavy industry, have large utility connections and have track on site. At least 81 North American paper mills, with 20.7 million tons of capacity, have shut since 2016, including 16 in 2025.⁵³ Tradepoint Atlantic, the former Bethlehem Steel plant at Sparrows Point, Maryland, shows what reuse can achieve: 3,300 acres with 70 miles of track and access to both CSX and Norfolk Southern,⁵⁴ where construction of a $1.2 billion container terminal with on-dock rail began in 2026.⁵⁵ Norfolk Southern bought the 1,800-acre former International Paper mill at Courtland, Alabama, in January 2026 to recruit rail customers, and is working with utilities on power, gas and water.⁵⁶ At River Ridge Commerce Center in Indiana, a 6,000-acre former Army ammunition plant, OmniTRAX operates the railroad and develops sites.⁵⁷

Reuse carries its own risks, and they decide whether a project succeeds. Environmental liability sets the schedule: at Sparrows Point, the buyer was required to provide $48 million in financial assurance to the state and agreed to pay EPA $3 million under a prospective purchaser agreement before redevelopment began,⁵⁸ and in September the sale of the former Liberty Steel mill in Georgetown, South Carolina, fell through because the buyer could not complete soil and water testing and permits by the seller's deadline.⁵⁹ Track that has been idle for years usually needs new ties, switches and crossings, and a layout built for one mill may not suit several tenants. The track's legal status also matters: spur, industrial and switching track can be built or removed without STB approval, but track that is part of a rail line cannot be removed without abandonment authority under 49 U.S.C. 10903.⁶⁰

The reservoir is also contested. The large power connections and water supply that make a closed plant attractive to industry also attract data centers, which ship little by rail. In Kentucky, data center campuses are planned at the former Paducah uranium enrichment plant, the Century Aluminum smelter in Hancock County and the former AK Steel mill in Ashland.⁶¹ A rail operator that waits for these sites to come to market will often lose them to the bidder that values the power connection most. One that can price the cleanup, the track work and the utilities, and approach the owner early, can often acquire a closed plant for less than an equivalent greenfield site, as Norfolk Southern's purchase at Courtland suggests.

6Transaction values already assume the edges will grow

6.1 Transaction multiples have risen

Part 1 noted that capital providers have no accepted basis for valuing growth at shortlines and terminals. They have nonetheless been paying for it. Infrastructure funds continue to buy shortlines and rail terminals, and there is no shortage of targets: of the roughly 610 U.S. shortlines, about two-thirds are independent.⁶² Transaction values have risen with that demand (Exhibit 7). When shortline holding companies or financial buyers were the acquirers, transactions in BMO's 2005–2012 set were valued at 7.5x to 11.6x trailing EBITDA; Class I buyers paid more, 15.0x for the EJ&E and 24.2x for the DM&E including contingent payments.⁶³ Since 2017, most railroad transactions with a public price and EBITDA have been valued at roughly 13x to 22x.

The exceptions were valued lower for identifiable reasons. Transtar had one steel customer under a long-term contract, and terminals with contracted cash flow, such as the Port Arthur crude oil terminal FTAI Infrastructure agreed to buy for about $255 million on about $50 million of expected EBITDA, can trade at about 5x.⁶⁴ Exhibit 8 lists the main transactions with a public price and EBITDA figure.

Notes: The 2025, 2021 and 2016 multiples are our calculations from the cited figures. Each multiple uses a different EBITDA basis (trailing, annualized or forward), so they are comparable only roughly. Genesee & Wyoming also owned railroads in Australia and the U.K. at the time of its sale.

Two features of these transactions bear on our argument. First, buyers are paying for growth and savings they expect to create rather than for the cash flow they acquire: Genesee & Wyoming paid about 10.5 times the EBITDA it expected in its first year of owning Providence & Worcester, and that figure already counted $8 million of cost savings. Second, strategic buyers can pay more than financial ones because they capture value beyond the target's own cash flow. A Class I railroad that buys a connecting shortline captures the long-haul revenue on its traffic and absorbs its overhead; CN's US$230 million purchase of Iowa Northern in 2023 is a recent example, and CN did not disclose the railroad's EBITDA.⁷⁶ Public railroad stocks generally trade at or below the lower end of these transaction multiples; Union Pacific was at about 14.6 times EBITDA in early October, against a 10-year median of 13.5x.⁷⁷

6.2 What our own transaction reviews show

The public record leaves out most single-shortline sales, so we also drew on 17 North American shortline and regional transactions from 2011 to 2021 that our team reviewed, advised on or led. We report them only in aggregate, because several of the underlying figures came from confidential offering materials. Across the 17 transactions, the transaction value averaged about 12.9 times at-closing EBITDA, with a median of 12.1x and a range of 5.0x to 15.3x.

The headline multiple understates how much the outcome depended on what the buyer did after closing. In several of these transactions the buyer removed the seller's corporate overhead, related-party costs or duplicate administration soon after closing, which lowered the effective multiple by three to eight turns. Deferred capital spending worked in the other direction: track, bridges and equipment the seller had not maintained became the buyer's cost on day one, and in one transaction a few million dollars of deferred work added more than half a turn to the effective multiple. Railroads sold with long-term volume commitments from their main shippers traded toward the top of the range, because the buyer was paying for contracted rather than forecast cash flow.

6.3 The growth embedded in today's transaction values

A higher transaction value changes where an investor's return must come from. To illustrate, we assumed that 60% of a railroad's EBITDA is left as cash after maintenance capital and taxes, a ten-year hold, no debt and a target return of 10% a year. At 8x EBITDA, EBITDA needs to grow only about 2% a year. At 13x, close to the average in our reviews, it needs to grow about 5%. At 17x, the multiple paid for the Wheeling & Lake Erie, it needs about 6%, and nearly 8% if the railroad is later sold three turns lower (Exhibit 9). U.S. carloads are up 2.7% this year, a strong year by recent standards.²² Different assumptions change the figures but not the conclusion: a higher cash share or a lower target return reduces the growth required, but at current multiples the gap between required growth and market growth remains wide.

This is where the capital markets connect to the rest of our argument. At today's transaction values, the return has to come from growth the owner creates, and the main sources of that growth are the elements of access capacity: new customers on ready sites, transloads that bring truck freight to rail, and the service and pricing that keep it there. An owner that holds the railroad, the terminals and the land beside the track can capture the growth each creates for the others. Capital that simply holds the asset will find it difficult to earn its return. Part 1's execution sequence commits capital to a terminal or a site only after shipper volume and Class I terms are in writing. At today's transaction values, the same discipline belongs in the acquisition model: the growth a buyer pays for should be tested lane by lane, and against the interchange terms, before the price is agreed.

The same arithmetic has consequences beyond the buyer. Some of the growth a buyer needs will come from rates, switching charges and demurrage, so shippers on lines acquired at high multiples should expect them to be reviewed. A buyer that has stretched on value has a reason to defer track and bridge work to protect cash yield, which is how the deferred capital we found in our reviews accumulates. Exit values are uncertain, because a railroad bought at 15x to 17x needs the next buyer to pay a similar multiple, and the likely buyers of a large shortline group are few; if UP–NS is approved and another combination follows, the Class I railroads may be less active as buyers for some years. Small railroads are worth more to a buyer that already owns a group, because that buyer can remove the overhead of the one it acquires. And for a railroad owner, building a terminal or buying land on its own line often adds EBITDA at a lower multiple than buying another railroad.

7What each party has to do now

Part 1 concluded that merchandise carload will not be recovered by any one party acting alone, and set out what five parties have to do. Those requirements stand. Exhibit 10 adds what is now urgent for each party, and what each should build over the next several years if access capacity is where value is moving.

Shortline and terminal operators should treat access capacity as their core business rather than an adjunct to running trains. In the near term, operators that connect with Union Pacific or Norfolk Southern, or that depend on traffic interchanged with either railroad at a gateway, should identify which gateways and routings their traffic depends on, estimate what a change in routing or rates would cost, and file for conditions, with reporting and enforcement, before the November 18 deadline if they need them; we take no position on whether the merger should be approved. All operators should repeat the customer audit and requote lanes at current fuel prices, with quotes that still work at $4.40 diesel. Over the next several years it means running terminals as a separate line of business, building only behind signed volume, placing transloads close to clusters of customers, and identifying the plants on their lines most likely to close so they can approach the owners before the sites are marketed.

Class I railroads already treat their shortline and terminal partners as their gathering arm: Union Pacific placed 12 of its 15 new Focus Sites in 2025 on shortlines, and BNSF's one disclosed new Certified Site of 2025 was the first served by a partner shortline (Section 4). The question consolidation raises is whether the terms that support that relationship, namely interchange windows, divisions and car supply, carry through a merger and into a combined network's pricing. Part 1 set out what that support requires. The merger proceeding is the occasion for the applicants to state, in their filings and contracts, how shortline interline traffic will be priced and handled, and for every Class I railroad to keep the terms it offers its gathering partners competitive with truck as the network consolidates.

Shippers and manufacturers should secure rate and volume commitments in writing before a line they depend on changes hands, estimate what losing a routing option would cost them, and locate new plants on sites that are served, zoned and committed.

The public sector and its development partners should invest in readiness rather than acreage: zoning, utilities, lead-track design and a committed serving railroad before a tenant arrives. Closed plants should be acquired only after environmental testing, track inspection and a check of the track's STB status. Public owners of rail lines that cannot fund rehabilitation should consider long-term operating concessions that bring private operators and capital to lines they cannot sell; our August paper, Buying Cash Flow Without Buying the Railroad,⁷⁸ describes that structure.

Capital providers should underwrite an operator's demonstrated ability to add carloads rather than the traffic it already has, assume exit multiples below today's, and compare the cost of adding EBITDA through terminals and land on lines already owned with the cost of acquiring another railroad. Owners considering a sale should compare a full sale with a partial recapitalization or a concession, and should know which overhead a buyer will remove and which capital work is overdue before they negotiate.

All parties should underwrite projects without assuming a federal grant, while applying to state freight rail programs and supporting the 45G bills (Section 8).

8What would change our view

Our argument depends on conditions that could change, and on public support that is less certain than it was. Public funding has paid for much of the track rehabilitation and transload capacity at the edges of the network. About $2.6 billion of Consolidated Rail Infrastructure and Safety Improvements (CRISI) grants had gone to shortline projects as of November 2024.⁷⁹ The 45G track maintenance tax credit equals 40 cents per dollar of qualified spending, capped at $3,500 per mile, a cap unchanged since 2005;⁸⁰ H.R. 516 and S. 1532 would raise it to $6,100 a mile and index it to inflation,⁸¹ and as of late April 2026 the bills had 165 House and 42 Senate cosponsors.⁸² The authorization in the federal surface transportation law (IIJA) ran out on September 30, 2026. The House Transportation and Infrastructure Committee approved a five-year, $580 billion replacement by a 62–2 vote in May, but the Senate has not produced its own bill.⁸³ A continuing resolution funds programs through December 11 but did not carry forward IIJA's advance appropriations; the American Public Transportation Association estimates the resulting cut to rail funding at 82%, as cited by the Union of Concerned Scientists.⁸⁴ Unlike highways, rail programs have no trust fund and depend on annual appropriations.⁸⁵ If federal rail funding is not restored, access capacity will be built more slowly and more of its cost will fall on private capital.

Regulation is, on balance, moving in a direction that helps the edges of the network. In July 2025 the Seventh Circuit vacated the STB's 2024 reciprocal switching rule but did not rule on the requirement that Class I railroads report expanded service metrics every week, so it remains in effect.⁸⁶ Those reports give shortlines and terminals a public benchmark against which to sell their own service. Crew costs are moving the other way: recent Class I settlements are raising pay for the train crews that shortlines and terminal operators compete for, and a tentative CPKC agreement on the former Dakota, Minnesota & Eastern would raise engineer pay 41.42% by 2029.⁸⁷ Because local switching and terminal work are labor-intensive, that would limit how fast the access layer can grow.

Part 1 answered the objections most often raised to its operating case. Exhibit 11 answers those most often raised to this one, in the same form.

Two developments would most change our view. If the merger is denied or heavily conditioned, there would be fewer line sales and terminal openings, although the gateway risk to shortlines would also recede. And if diesel falls faster than expected while industrial projects slip, both near-term and long-term demand for access capacity would arrive later, and higher borrowing costs would compound the effect on transaction values. Exhibit 12 lists the signposts we will track, including the two measures Part 1 proposed for judging whether merchandise carload share is being recovered.

About the author

Tim Eklund is Managing Partner of Rail-Industrial Partners, LLC. He has led the acquisition, operation and growth of 11 shortline railroad operations and the development of rail-industrial parks in North America.

His experience includes owning and operating shortline railroads and rail-served terminals, and working with Class I railroads on pricing and divisions, contracts, marketing and industrial development, interchange and blocking, car supply, local service design, and network planning, which is the basis of the judgments in this paper.

About Rail-Industrial Partners

Rail-Industrial Partners is a specialized investment and strategic advisory platform focused on the North American freight rail, rail services and rail-industrial sectors. It invests in and operates shortline railroads, terminals and rail-served industrial sites, and advises owners of those assets.

The firm's partners and advisors have owned and operated shortline, regional and Class I railroads throughout North America.

Notes and sources

1. 1Capturing the Merchandise Carload Market, September 2026. https://www.rail-industrial.com/news/recapturing-the-merchandise-carload
2. 2U.S. Government Accountability Office, Freight Rail: Information on Precision-Scheduled Railroading, GAO-23-105420, December 2022. https://www.gao.gov/assets/gao-23-105420.pdf
3. 3STB, Decision on motions for summary denial, PR-26-23, September 18, 2026. https://www.stb.gov/news-communications/latest-news/pr-26-23/
4. 4Union Pacific, Union Pacific and Norfolk Southern create America's first transcontinental railroad, July 29, 2025. https://investor.unionpacific.com/news-releases/news-release-details/union-pacific-and-norfolk-southern-create-americas-first
5. 5FreightWaves, Union Pacific, Norfolk Southern file revised merger application, April 2026. https://www.freightwaves.com/news/union-pacific-norfolk-southern-file-revised-merger-application
6. 6FreightWaves, BNSF, CPKC and CSX to request broad access to a combined UP-NS network, September 2026. https://www.freightwaves.com/news/bnsf-cpkc-and-csx-to-request-broad-access-to-a-combined-up-ns-network
7. 7Union Pacific and CN, Agreement to expand customer opportunities in connection with merger, July 22, 2026. https://www.globenewswire.com/news-release/2026/07/22/3331806/0/en/Union-Pacific-and-CN-Reach-Agreement-to-Expand-Customer-Opportunities-in-Connection-with-Merger.html
8. 8Supply Chain Dive, Union Pacific–Norfolk Southern merger market impact, February 2026. https://www.supplychaindive.com/news/union-pacific-norfolk-southern-merger-market-impact/809390/
9. 9CSX, CSX and BNSF announce new intermodal services, August 2025. https://www.csx.com/index.cfm/about-us/csx-bnsf-announce-new-intermodal-services-offering-seamless-coast-to-coast-rail-solutions/
10. 10FreightWaves, Union Pacific, Norfolk Southern: rival alliances bolster merger case, September 2025. https://www.freightwaves.com/news/union-pacific-norfolk-southern-rival-alliances-bolster-merger-case
11. 11Yahoo Finance, Berkshire Hathaway not looking to buy another railroad, August 2025. https://finance.yahoo.com/news/berkshire-hathaway-not-looking-buy-172401826.html
12. 12TT News, CSX chief on efficiency and rivals' merger, February 2026. https://www.ttnews.com/articles/csx-efficiency-rivals-merger
13. 13Thompson Hine, STB Lifts Abeyance in Union Pacific–Norfolk Southern Merger. https://www.thompsonhine.com/insights/stb-lifts-abeyance-in-union-pacific-norfolk-southern-merger-and-sets-procedural-schedule/
14. 14Federal Register, UP–NS revised application decision, May 29, 2026. https://www.federalregister.gov/documents/2026/05/29/2026-10751/union-pacific-corporation-and-union-pacific-railroad-company-control-norfolk-southern-corporation
15. 15Railway Age, STB Accepts UP-NS Revised Merger Application. https://www.railwayage.com/regulatory/stb-accepts-up-ns-revised-merger-application-delays-proceedings/
16. 16FreightWaves, via Yahoo Finance, BNSF CEO cautions short lines, April 2026. https://finance.yahoo.com/economy/policy/articles/bnsf-ceo-cautions-short-line-171045620.html
17. 17Denis Breen, FTC Bureau of Economics, The Union Pacific/Southern Pacific rail merger: a retrospective on merger benefits, 2004. https://www.ftc.gov/sites/default/files/documents/reports/union-pacific/southern-pacific-rail-merger-retrospective-merger-benefits/wp269_0.pdf
18. 18Railfan & Railroad, BNSF calls on STB to review harm caused by UP-SP merger, December 2025. https://railfan.com/bnsf-calls-on-stb-to-review-harm-caused-by-up-sp-merger/
19. 19STB, PR-26-17, July 21, 2026. https://www.stb.gov/news-communications/latest-news/pr-26-17/
20. 20American Farm Bureau Federation, Proposed rail merger comes at farmers' expense, March 2026. https://www.fb.org/market-intel/proposed-rail-merger-comes-at-farmers-expense
21. 21Norfolk Southern, UP and NS affirm strength of merger application and offer new customer assurances, July 27, 2026. https://www.norfolksouthern.com/en/newsroom/news-releases/union-pacific-and-norfolk-southern-affirm-strength-of-merger-application-and-offer-unprecedented-new-customer-assurances
22. 22AAR weekly rail traffic, week ending September 26, 2026, via AJOT. https://www.ajot.com/news/aar-reports-rail-traffic-for-the-week-ending-september-26-2026
23. 23The Waterways Journal, Iowa steel plant announced on Upper Miss, October 2, 2026. https://www.waterwaysjournal.net/2026/10/02/iowa-steel-plant-announced-on-upper-miss/
24. 24Progressive Railroading, Union Pacific to serve future Hyundai steel plant in Louisiana, March 26, 2025. https://www.progressiverailroading.com/union_pacific/news/Union-Pacific-to-serve-future-Hyundai-steel-plant-in-Louisiana--74185
25. 25Seoul Economic Daily, Hyundai, POSCO break ground on $5.8 billion U.S. steel mill, 2026. https://en.sedaily.com/finance/2026/09/06/hyundai-posco-break-ground-on-58-billion-us-steel-mill
26. 26BIC Magazine, CPChem's $8.5B Golden Triangle Polymers facility begins commissioning phase, May 14, 2026. https://www.bicmagazine.com/departments/operations/cpchem-golden-triangle-polymers-facility-startup-phase/
27. 27Golden Triangle Polymers, Project FAQs. https://www.goldentrianglepolymers.com/project-faqs/
28. 28Railway Track & Structures, NS completes Scout Motors I-77 rail bridge, November 2025. https://www.rtands.com/freight/class-1/ns-completes-scout-motors-i-77-rail-bridge/
29. 29FreightWaves, New industrial development worth $7.7B for Norfolk Southern in 2025, February 3, 2026. https://www.freightwaves.com/news/new-industrial-development-worth-7-7b-for-norfolk-southern-in-2025
30. 30BNSF, BNSF customers invested record $5.3 billion on 2025 rail-served projects, April 2026. https://www.bnsf.com/news-media/railtalk/service/2025-economic-development.html
31. 31EIA, Weekly U.S. No. 2 diesel retail prices. https://www.eia.gov/dnav/pet/hist/LeafHandler.ashx?n=PET&s=EMD_EPD2D_PTE_NUS_DPG&f=W
32. 32AAR, Freight Rail Facts & Figures. https://www.aar.org/freight-rail-facts-figures/
33. 33DAT, Truckload freight rates hit two-year highs as diesel costs surge, April 2026. https://www.dat.com/company/news-events/news-releases/dat-truckload-freight-rates-hit-two-year-highs-as-diesel-costs-surge
34. 34The Trucker, DAT: Load posts down 1% from the prior week, September 29, 2026. https://www.thetrucker.com/trucking-news/business/dat-load-posts-down-1-from-the-prior-week
35. 35Union Pacific, Carload mileage-based fuel surcharge. https://www.up.com/shipping/surcharge/mileage
36. 36RSI Logistics, Comparing the costs of rail shipping vs. truck, 2024. https://www.rsilogistics.com/blog/comparing-the-costs-of-rail-shipping-vs-truck/
37. 37Steel Wheel Logistics, What is transloading?, February 2026. https://steelwheellogistics.com/blog/what-is-transloading
38. 38EIA, Short-Term Energy Outlook, September 9, 2026. https://www.eia.gov/outlooks/steo/
39. 39Progressive Railroading, Study: the impact of higher fuel costs on rail carload and intermodal market share, 2009. https://www.progressiverailroading.com/rail_industry_trends/article/Study-The-Impact-of-Higher-Fuel-Costs-on-Rail-Carload-and-Intermodal-Marketshare--19844
40. 40Reuters, via GV Wire, Union Pacific says high diesel prices shifting freight from trucks to rail, September 16, 2026. https://gvwire.com/2026/09/16/union-pacific-says-high-diesel-prices-shifting-freight-from-trucks-to-rail/
41. 41Railway Age, OmniTRAX Launches Expanded Transload Platform, September 30, 2026. https://www.railwayage.com/freight/short-lines-regionals/omnitrax-launches-expanded-transload-platform/
42. 42Railway Age, CN Focuses on Terminal Efficiency, Expands Transload Operations in Wisconsin, September 29, 2026. https://www.railwayage.com/freight/class-i/cn-focuses-on-terminal-efficiency-expands-transload-operations-in-wisconsin/
43. 43Union Pacific, Union Pacific working with short line railroads to expand businesses, August 25, 2025. https://investor.unionpacific.com/news-releases/news-release-details/union-pacific-working-short-line-railroads-expand-businesses
44. 44Progressive Railroading, CSX identifies 21 new Select Sites, March 16, 2026. https://www.progressiverailroading.com/csx_transportation/news/CSX-identifies-21-new-Select-Sites-for-development--76551
45. 45Railway Age, CN adds new industrial development opportunities to Certified Rail-Ready Sites program, May 2026. https://www.railwayage.com/freight/class-i/cn-adds-new-industrial-development-opportunities-to-certified-rail-ready-sites-program/
46. 46Railway News, CPKC opens 6,000 acres of Site Ready land across North America, March 2026. https://railwaynews.net/cpkc-opens-6000-acres-site-ready-land-north-america.html
47. 47BNSF, Guidelines for industry track projects, July 2023. https://www.bnsf.com/ship-with-bnsf/ways-of-shipping/pdf/indytrkstds.pdf
48. 48Railway Age, Building successful industrial development spaces, March 9, 2026. https://www.railwayage.com/freight/building-successful-industrial-development-spaces/
49. 49Location Georgia, Is your industrial site rail ready?, August 2024. https://www.locationgeorgia.com/is-your-industrial-site-rail-ready-august-2024/
50. 50Rail-Served Industrial Sites & Greenfield Development, October 2025. https://www.rail-industrial.com/news/rail-served-sites-whitepaper
51. 51Railway Age, Hillwood, BNSF and City of Fort Worth launch Alliance Logistics District, November 2025. https://www.railwayage.com/news/hillwood-bnsf-city-of-forth-worth-launch-alliance-logistics-district/
52. 52Union Pacific, Union Pacific unveils new industrial park, December 2025. https://investor.unionpacific.com/news-releases/news-release-details/union-pacific-railroad-unveils-new-industrial-park-offering
53. 53Rivermill Advisors, North American mill closures by year, 2016–2025. https://www.rivermilladv.com/uploads/8/3/5/4/83544922/mill_closures_2016-2025.pdf
54. 54Tradepoint Atlantic, Why Tradepoint. https://www.tradepointatlantic.com/why-tradepoint/
55. 55Yahoo Finance, Tradepoint Atlantic and TiL break ground on Sparrows Point terminal, May 2026. https://finance.yahoo.com/economy/policy/articles/tradepoint-atlantic-til-break-ground-102306118.html
56. 56Moulton Advertiser, Norfolk Southern purchases former International Paper mill site, January 2026. https://www.moultonadvertiser.com/news/article_37770667-04c3-41d0-b3c7-8b5dd6904fbe.html
57. 57River Ridge Commerce Center, OmniTRAX takes over rail operations, November 2021. https://www.riverridgecc.com/news/article/omnitrax-takes-over-rail-operations-within-indianas-river-ridge-commerce-center-eyes-new-park-development-projects
58. 58Maryland Department of the Environment, Actions clear the way for redevelopment at former Sparrows Point, September 2014. https://news.maryland.gov/mde/2014/09/18/actions-clear-the-way-for-job-creating-redevelopment-at-former-sparrows-point-steelmaking-facility/
59. 59WCSC, Sale of former Liberty Steel mill site has failed, September 2026. https://www.live5news.com/2026/09/18/state-representative-confirms-sale-former-liberty-steel-mill-site-has-failed/
60. 60STB staff guidance on spur, industrial and switching track (undated; posted April 2026). https://content.govdelivery.com/attachments/AZMARIC/2026/04/24/file_attachments/3628145/STB%20Response.pdf
61. 61Lexington Times, Data centers are eyeing Kentucky's idle industrial sites, August 2026. https://lexingtonky.news/2026/08/03/data-centers-need-massive-power-they-re-eyeing-kentucky-s-idle-industrial-sites/
62. 62Mergers & Acquisitions, On the Right Track: Infra Funds Fuel Rail M&A Revival, December 22, 2025. https://www.themiddlemarket.com/feature/on-the-right-track-infrastructure-funds-fuel-a-rail-ma-revival
63. 63Providence and Worcester Railroad, proxy supplement with BMO precedent transactions and public company multiples, October 3, 2016. https://www.sec.gov/Archives/edgar/data/0000831968/000091426016000131/pwr-10316.htm
64. 64Offshore Technology, FTAI unit to acquire USDG crude oil logistics assets for $255m, September 29, 2026. https://www.offshore-technology.com/news/ftai-acquire-usdg-crude-oil-logistics-255m/
65. 65FTAI Infrastructure, Q2 2025 earnings call transcript, August 8, 2025. https://equibles.com/stocks/fip/calls/2025-q2
66. 66Union Pacific, Form 8-K announcing the Norfolk Southern merger agreement, July 29, 2025. https://www.sec.gov/Archives/edgar/data/100885/000119312525167154/d64537dex991.htm
67. 67Macrotrends, Norfolk Southern EBITDA 2012–2026. https://www.macrotrends.net/stocks/charts/NSC/norfolk-southern/ebitda
68. 68CPKC, Canadian Pacific and Kansas City Southern Execute Agreement to Combine, September 2021. https://www.cpkcr.com/en/media/Canadian-Pacific-and-Kansas-City-Southern-Execute-Agreement-to-Combine-Creating-First-Single-Line-Rail-Network-Linking-US-Mexico-Canada
69. 69Kansas City Southern, fourth-quarter and full-year 2020 results, January 22, 2021. https://www.sec.gov/Archives/edgar/data/54480/000005448021000014/q42020exhibit991.htm
70. 70U.S. Steel, sale of Transtar announcement, 2021. https://investors.ussteel.com/news-events/news-releases/detail/74/u-s-steel-announces-sale-of-transtar-llc-to-fortress
71. 71Fortress Transportation and Infrastructure (FTAI), fourth-quarter 2021 earnings call, February 25, 2022, transcript via The Motley Fool. https://www.fool.com/earnings/call-transcripts/2022/02/25/fortress-transportation-infrastructure-investors-f/
72. 72Kansas City Southern, Form 425 filing with Morgan Stanley precedent transactions, 2021. https://www.sec.gov/Archives/edgar/data/54480/000119312521240618/d130317d425.htm
73. 73Railway Age, Brookfield, partners to acquire G&W for $8.4B, July 2019. https://www.railwayage.com/freight/short-lines-regionals/brookfield-partners-to-acquire-gw-for-8-4b/
74. 74Railway Gazette International, Genesee & Wyoming to buy Providence & Worcester Railroad, August 2016. https://www.railwaygazette.com/operators/2016/08/16/genesee-wyoming-to-buy-providence-worcester-railroad/
75. 75Supply Chain 24/7, Genesee & Wyoming set to buy RailAmerica, July 2012. https://www.supplychain247.com/article/genesee_wyoming_set_to_buy_railamerica/RailAmerica
76. 76CN, 2023 annual report, management's discussion and analysis. https://www.sec.gov/Archives/edgar/data/16868/000001686824000009/a2023annualstatsmda.htm
77. 77GuruFocus, Union Pacific EV-to-EBITDA, as of October 4, 2026. https://www.gurufocus.com/term/enterprise-value-to-ebitda/UNP
78. 78Buying Cash Flow Without Buying the Railroad, August 2026. https://www.rail-industrial.com/news/concession-whitepaper
79. 79Trains, ASLRRA president discusses CRISI's short line transformation, November 12, 2024. https://www.trains.com/trn/news-reviews/news-wire/aslrra-president-discusses-crisis-short-line-transformation/
80. 80Progressive Railroading, Short-line stakeholders seek to modernize 45G tax credit. https://www.progressiverailroading.com/railPrime/details/Short-line-stakeholders-seek-to-modernize-45G-tax-credit--73014
81. 81ASLRRA, 45G Short Line Tax Credit Modernization. https://www.aslrra.org/advocacy/45g-short-line-tax-credit-modernization/
82. 82ASLRRA, Views & News, April 29, 2026. https://www.aslrra.org/news/views-and-news/views-news-042926/
83. 83IEDC, Update on Surface Transportation Reauthorization. https://www.iedconline.org/news/2026/06/10/federal-policy-updates/update-on-surface-transportation-reauthorization/
84. 84Union of Concerned Scientists, Extra Time, Quiet Cuts, September 24, 2026, citing the American Public Transportation Association's estimate. https://blog.ucs.org/kshen/extra-time-quiet-cuts-surface-transportation-reauthorization-extension/
85. 85Legis1, Rail Reauthorization: Money, Mission, and Hard Choices. https://legis1.com/news/rail-reauthorization-2026-money-mission-and-hard
86. 86FreightWaves, Setback for rail shippers as court vacates switching rule, July 9, 2025. https://www.freightwaves.com/news/federal-court-vacates-rule-to-help-shippers-with-inadequate-rail-service
87. 87Progressive Railroading, CPKC, BLET reach tentative five-year deal, October 1, 2026. https://www.progressiverailroading.com/labor/news/CPKC-BLET-reach-tentative-five-year-deal--77858
88. 88Progressive Railroading, House transportation committee drafts surface transportation reauthorization bill. https://www.progressiverailroading.com/federal_legislation_regulation/news/House-transportation-committee-drafts-surface-transportation-reauthorization-bill--76957

Data notes. Valuation multiples use different EBITDA bases (trailing, annualized or forward) and are comparable only roughly. The 17 transactions from our own reviews are reported only in aggregate, because several of the underlying figures came from confidential offering materials. The fuel-surcharge illustration (Exhibit 5) and the return illustration (Exhibit 9) are the authors' calculations; their assumptions are stated with each exhibit. Transaction value is the price reported for each transaction, on an enterprise-value basis except where Exhibit 8 notes otherwise.