North American intermodal up 5.1% YTD as truckload rates forecast +30% in 2026

The Analysis

North American intermodal volumes are running about 5.1% above last year year-to-date, with the week of 15 August alone at roughly 291,000 units, up 2.7% year on year. That is not a seasonal blip: the rail-intermodal book has been running ahead of its five-year average through the first half while the wider freight market stayed soft.

The reason sits on the road side of the equation. C.H. Robinson now forecasts spot truckload rates rising 30% in 2026 as over-the-road capacity tightens again. Union Pacific is already managing bookings in constrained markets, and new long-term rail deals are unlikely to be signed before 2027, which means the window for locking 2026 capacity is open now and closing.

The conversion is most visible on 550-to-1,500-mile lanes, the band where rail can replace a long-haul truck move without wrecking transit time. The five briefs below translate that into specific moves for exporters, cross-border e-commerce sellers, factories, brands and professional buyers, and spell out the arithmetic behind each one.

For Exporters

• Re-price all open DDP quotes to a $2.86-per-mile truck benchmark by 25 September and flag any lane where gross margin falls below 12%.
• Add a separate inland-leg line with a 14-day validity to every new quotation from 1 October.
• Sign a 2026 rail-intermodal capacity agreement by 15 October covering 550-to-1,500-mile lanes, with 70% contracted and 30% held for spot.
• Book rail slots two to four weeks ahead; treat a missed cut-off as a $1,060-per-box penalty and escalate it internally.
• Verify inland point codes and delivery orders on every intermodal booking to avoid ramp storage charges.

If you sell on DDP or DAP terms into US inland points, the 30% truckload increase forecast for 2026 lands on your profit and loss, not your buyer's. Chinese exporters booking 40-foot containers into Los Angeles or Long Beach and then arranging delivery to Chicago, Dallas or Kansas City are paying for the 550-to-1,500-mile leg themselves. At the same time, the 5.1% year-to-date intermodal growth shows the rail alternative is already absorbing volume, so space on the good services is not unlimited.

Assume a 40-foot container moving 1,000 miles inland. Take a truckload benchmark of $2.20 per mile excluding fuel, so $2,200 per box today. The forecast 30% increase takes it to $2,860. Take an intermodal benchmark of $1.80 per mile, or $1,800 per box, with fuel and drayage billed separately but broadly comparable. The saving against the post-increase truck rate is $1,060 per box. On 40 containers a month, of which 60% move past the port, that is 24 boxes and $25,440 a month, or just over $305,000 a year. If only half converts, the saving still lands near $152,000.

Quote with a 14-day validity and break the US inland leg out as a separately indexed line rather than burying it in one DDP number. By 25 September, re-price every open DDP quote at the new truck benchmark and flag the ones where gross margin drops below 12%. By 15 October, sign a 2026 rail-intermodal capacity agreement covering the 550-to-1,500-mile band, targeting 70% of inland volume on contracted rail and holding 30% on the road for urgent orders. Put one named person, your logistics manager, on the booking calendar so rail slots are confirmed two to four weeks ahead.

The alternative is to keep everything on the road and pass the 30% through in your price. That works only if the buyer accepts an annual escalation clause, which most US importers will not. A second fallback is switching Incoterms to FOB port of loading so the inland cost sits with the buyer, but that usually costs you two to three points of price. The traps are practical: rail booking cut-offs are strict, so a missed slot becomes a spot truck move at the highest rate of the year; rail needs an accurate inland point code and a clean delivery order, or the box sits at the ramp and racks up storage; and intermodal adds roughly 1.5 days in transit plus ramp handling, so a DDP delivery date promised on the old truck schedule will slip.

For Cross-Border E-commerce

• Recompute landed cost for the top 100 SKUs at a 30% higher inland truck rate by 30 September; flag any SKU whose gross margin falls below 30%.
• Keep the top 20% of SKUs on truck and move the long tail to rail from 1 October, targeting 60% of inland volume on rail.
• Raise safety stock on rail lanes from 21 to 28 days to cover 1.5 days of extra transit plus ramp handling.
• Reserve air freight for confirmed stock-outs only, capped at 5% of monthly volume.
• Reset customer-facing delivery windows to the rail schedule before the 1 November peak.

For a cross-border importer, the 30% truckload forecast hits the part of the cost stack that is easiest to lose track of: moving stock from a US port to an inland fulfilment centre. That is a 550-to-1,500-mile move, exactly the band where rail-intermodal is replacing road, and the 5.1% year-to-date intermodal growth confirms the capacity is being taken up. Your landed unit cost will move even if your ocean rate does not.

Take a 40-high-cube container with $60,000 of goods, $5,000 of ocean freight and clearance, and $2,200 of inland trucking, giving $67,200 landed. The 30% truck increase adds $660 and takes it to $67,860, a 0.98% rise in landed cost. If you sell at a 35% gross margin, revenue per container is about $103,385 and gross profit falls from $36,185 to $35,525, so the margin rate slips from 35.0% to 34.3%. That is 0.7 percentage points gone from one line item. Switching 60% of that inland volume to intermodal at the assumed $1.80 per mile saves $636 per box and claws back nearly all of it without touching your selling price.

Move to a tiered replenishment model before the October peak buying window. Keep the top 20% of SKUs, the ones carrying 70% or more of revenue, on expedited truck capacity into the port-adjacent warehouse, and route the long tail by rail. Lift safety stock from 21 to 28 days on intermodal lanes to absorb the extra 1.5 days of transit plus two days of ramp handling. From 1 October, re-quote every SKU whose inland cost moved more than 1% of landed cost, and adjust price or bundle rather than absorbing it.

The tempting shortcut is to divert everything to air to protect delivery promises; at current cross-border air rates that costs several times the inland truck increase, so reserve air for confirmed stock-outs only. A second option is holding more inventory on the West Coast and skipping the inland move, which trades trucking cost for higher West Coast storage and slower delivery. Watch three traps: intermodal containers often need a different chassis and a booked appointment at the ramp, and missed pickups trigger per-day storage that can exceed the freight saving; returns and exchanges get slower when restock moves by rail, so set customer-facing delivery windows from the rail schedule rather than the old truck schedule; and if you sell on a marketplace with a fulfilment service level, a two-day slip can push you out of the programme.

For Manufacturing Plants

• Classify all inbound material as line-stopping or non-critical by 30 September; roughly 50% should be eligible for rail.
• Lock a named 2026 rail allocation, not just a rate, in the first half of October for the 550-to-1,500-mile band.
• Hold a 10-day safety stock on truck-supplied line-stopping items and a 21-day buffer on rail-supplied material.
• Target $63,600 a month of savings by converting 60 of 120 monthly inbound loads to rail.
• Keep high-value components on truck because rail carrier liability is lower.
• Pre-clear inland point codes and delivery orders to avoid ramp storage after free time.

A plant importing raw material or components through a US port and hauling them inland is exposed twice: once to the 30% truckload forecast for 2026, and once to the risk that a rail conversion delays production feed. Intermodal is already up 5.1% year-to-date with the week of 15 August at about 291,000 units, so the rail network is carrying more volume and ramp congestion is a real variable. Union Pacific is managing bookings in constrained markets, which means availability, not price, is the binding constraint on the lanes where rail is the substitute.

Assume the plant takes 120 truckloads a month of inbound material on a 1,000-mile lane. At the $2.20-per-mile benchmark that is $2,200 per load, or $264,000 a month. A 30% increase takes each load to $2,860 and the monthly bill to $343,200, which is $79,200 more, or about $950,000 annualised. Intermodal at the assumed $1.80 per mile costs $1,800 per load, a $1,060 saving against the post-increase truck rate. Converting the 50% of inbound volume that is packaging, steel coil and other non-perishable material, so 60 loads, saves $63,600 a month and covers about 80% of the increase.

Split inbound material into two classes by 30 September: line-stopping items and everything else. Keep line-stopping items on truck with a contracted 2026 rate and a 10-day safety stock; move the rest to rail with a 21-day buffer. Lock the rail allocation with your forwarder in the first half of October, before Union Pacific finalises its 2026 book, and ask for a stated allocation number rather than a rate. Re-sequence the production plan so that rail-received material feeds builds that sit three weeks out on the schedule, and hold a two-week buffer on any single-source component whose lead time already exceeds 45 days.

Two alternatives exist and both carry a price. You can pre-buy and hold more raw material to absorb slower transit, but that consumes working capital and warehouse space and does nothing about the rate. You can dual-source to a domestic supplier, which removes the inland leg entirely but usually carries a 15-to-25% unit-price penalty. The traps sit at the ramp and in the paperwork: rail ramps charge per day after free time, so a missed appointment turns a $1,060 saving into a loss; intermodal movements need the material's inland point code and a delivery order that matches the purchase order exactly, or the box is held for a customs or carrier exception; and carrier liability on rail is lower than on road, so keep high-value components on truck regardless of the saving.

For Brand Owners

• Publish a two-tier delivery promise by 15 October: the original window for the top 20% of SKUs on truck, a longer window for rail stock.
• Give retail partners allocation priority over direct channels whenever a rail delay occurs.
• Set a weekly exception report with the 3PL and forwarder covering ramp dwell and missed rail cut-offs.
• Re-forecast November and December delivery windows on rail transit times rather than truck.
• Never split a single SKU family across modes; keep one order on one mode end to end.

For a brand, the 30% truckload forecast is a delivery-promise problem before it is a cost problem. If you convert inland moves to rail to dodge the increase, and the 5.1% year-to-date intermodal growth says the market is already doing exactly that, you add transit time on 550-to-1,500-mile lanes. Every promise you have published to retail partners and direct customers was built on a truck schedule, and that schedule is the thing that changes.

Model the trade-off. Assume 50,000 orders a month at a $120 average order value and a published five-day delivery promise. Moving part of restock to rail pushes order-to-delivery to about 6.5 days on affected lanes. If the 1.5-day slip causes 2% of orders to be refunded or returned, at an assumed 30% cost of order value, that is 1,000 orders and about $36,000 a month. Against that, converting 200 truckloads a month at a $1,060 saving each is $212,000 a month. The saving is nearly six times the service cost, so the answer is not to refuse rail. It is to stop promising one uniform window.

Split your delivery promise into two tiers by 15 October. Guarantee the original window only on the 20% of SKUs that carry the bulk of revenue and stay on truck; publish a longer, honest window for everything that moves by rail. Give retail partners allocation priority ahead of direct channels when a rail delay hits, because chargebacks on retail cost more than a consumer's one-day slip. Put your 3PL and your forwarder on a shared weekly exception report so a ramp delay is visible before it becomes a customer-facing miss, and re-forecast the November and December windows using rail transit rather than truck transit. Give the change one named owner, your customer operations lead, so the new windows reach partners and the website on the same day.

The obvious fallback, absorbing the 30% and keeping the truck promise, protects the experience but takes roughly a full point of margin if the inland leg is 3% of landed cost. The opposite fallback, moving everything to rail, is cheaper but guarantees a service dip across the whole catalogue. The traps sit in the middle: do not mix rail and truck lead times inside one SKU family, because partial arrivals create phantom stock-outs on the website; do not let a marketplace fulfilment service level be measured against a truck-based promise you can no longer keep; and do not let promotional calendars assume rail-transit inventory will land in time. Fix the promise first, then the promotion.

For Procurement Teams

• Model a 70/30 contract-to-spot split on your 550-to-1,500-mile volume; the modelled annual saving against all-spot is about $1.78m on 2,400 loads.
• Open negotiations in the last week of September and target a signed rail capacity agreement by 15 November.
• Require a firm allocation number, a named fuel index formula, and a 30% cap on spot exposure.
• Name a backup rail provider and a qualified truck carrier for every core lane.
• Reject any force-majeure clause that does not name specific ramps or allow substitution between them.

The single most actionable fact here is that Union Pacific is managing bookings in constrained markets and new long-term rail deals are unlikely before 2027. Against a forecast 30% rise in spot truckload rates for 2026, that means the contracted rail window is open for a few more weeks and then closes. Intermodal volumes are already 5.1% above last year year-to-date, so you are negotiating against a book that is filling, not an empty one.

Run the portfolio arithmetic. Assume 2,400 load-equivalents a year on a 1,000-mile lane, so 200 a month. All-spot at the post-increase truck rate of $2,860 costs $6.864m a year. All rail at the assumed $1.80 per mile costs $4.32m. A 70/30 contract-to-spot split, with the contracted volume on rail at $1,800 and the spot balance on road at $2,860, costs $5.083m: 1,680 loads at $1,800 plus 720 loads at $2,860. That is $1.781m cheaper than going all spot, and it leaves a spot tail to cover demand spikes instead of buying the whole year at the top of the market.

Go to market in the last week of September, not October. Bundle your 550-to-1,500-mile convertible volume into one package and offer a 12-month volume commitment in exchange for a firm capacity allocation and a rate cap. Ask for three clauses: a defined fuel index with a stated pass-through formula, a cap on spot exposure at 30% of committed volume, and a mutual force-majeure and re-routing clause that names specific ramps and allows substitution between them. Target a signed agreement by 15 November so it is in place before the 2026 truck book is set, and nominate a backup rail provider for each core lane.

The alternative to contracting early is to stay flexible and buy spot as rates move; that is only rational if your volumes are genuinely uncertain by more than 30% either way. The other alternative is multi-sourcing the mode rather than the supplier, keeping a qualified truck carrier for every rail lane so you can switch within a week. The traps are contractual rather than operational: an allocation clause without a service-level commitment and remedies is worth little in a tight market; a fuel clause written as a percentage of the base rate quietly re-inflates when the base rate rises; and a force-majeure clause that names no specific ramp gives the carrier an easy exit exactly when you need it. Read the exit terms as carefully as the rate. Send the clause list in writing a week before the meeting, so the carrier has to answer your terms rather than headline a rate.

↑ Back to the key points