Newsletter · · Ashutosh Agarwal

Freight Turned and Now It All Rides on AI - The Freight Cycle - Week of July 21, 2026

The Freight Cycle for the week of July 21, 2026. J.B. Hunt kicked off transport earnings with a blowout and spot truckload rates sit in the 97th percentile of five years, but underlying freight demand is up just 1.5% and almost all of the marginal growth is tied to building AI data centers, so the turn everyone waited three years for is balancing on a single, very tall stilt.

The Freight Cycle

Week of July 21, 2026: Freight Turned and Now It All Rides on AI

Well, that escalated. Two months ago the debate was whether trucking had finally bottomed. This week the podcasts stopped debating the bottom and started arguing about the top. J.B. Hunt kicked off transport earnings with a blowout, spot truckload rates are sitting in the 97th percentile of the last five years, and that includes the COVID melt-up, and the freight-conditions gauges are at all-time highs. On paper, the turn everyone waited three-plus years for is here.

But read past the headline and it gets weird. Underlying freight demand is barely growing, about 1.5% year over year. Almost every extra truckload moving in America right now is tied to one thing: building data centers for artificial intelligence. Strip that out and the consumer side of freight is still shrinking. So the cycle "turned," but it's balancing on a single, very tall stilt. That's the story this week, and it's the story I'd be underwriting into rail and truckload earnings over the next fortnight.

Let's get into it.

TL;DR

  • J.B. Hunt (JBHT) blew the doors off and dragged the whole transport complex up 2.8% on the day. Intermodal revenue +22%, volumes +10%, huge for intermodal, and management has not yet raised intermodal prices while peers push 15%. That's stored-up earnings for later.

  • Spot truckload rates are extreme, 97th percentile of the last five years, ~$353/day, up roughly 48–57% year over year once you adjust for fuel. FTR's carrier-conditions index hit an all-time high in May. Capacity is genuinely tight.

  • …but the strength is narrow and partly borrowed. Total freight demand is only +1.5% YoY, and the marginal growth is "almost 100%" AI data-center build-out. Consumer freight (beverages, appliances) is still falling. A chunk of the rate strength is a diesel spike, not demand.

  • Capacity is leaving for regulatory reasons, not economic ones. The government has pulled ~24,000 drivers over English-language rules and canceled ~28,000 fraudulent licenses, with 75 CDL schools now under investigation. Sign-on bonuses are back for the first time in years.

  • Brokerage margins are getting crushed even as revenue soars, J.B. Hunt's brokerage arm grew revenue 49% and still made almost no money. Bad omen for the pure-play brokers (CHRW, LSTR, RXO, HUBG) reporting soon.

  • The Union Pacific–Norfolk Southern merger got a rare on-the-record handicap from a former CSX strategy chief: probably approved, but the conditions could be nasty enough that UP walks.

  • Class 8 trucks: EPA is keeping the 2027 emissions deadline but softening the edges, and 2026 build slots are basically sold out. Rails and railcar lessors were quiet this week, worth noting given three big rail earnings land next week.

What's new

J.B. Hunt turned the transports green, and the tell was intermodal. On the July 16 episode of FreightWaves Today, analyst JP Hampstead (now at Westboro Capital, formerly FreightWaves) walked through the quarter: "J.B. Hunt intermodal revenue up 22% year over year. Intermodal volumes up 10%, which is a huge number for intermodal… The fact that they were able to beat so strongly against already very high expectations this early in the cycle… is what really impressed people." He flagged where the growth is coming from, not the long transcontinental lanes, but the East: "a shift from the transcontinental to really an eastern side of the United States… the eastern portion of intermodal is doing exceptionally well." (FreightWaves Today | July 16)

Here's the part that matters for the stock. FreightWaves co-founder Craig Fuller explained why intermodal volumes are surging without intermodal prices doing much: "Intermodal rates really haven't risen that much. But market share has risen because truckload rates have risen so much that more has converted to intermodal." In other words, truck got so expensive that freight is fleeing to the rails, and J.B. Hunt is grabbing share now, with pricing power it hasn't spent yet. As Hampstead put it, other operators are already raising rates 15% while Hunt sits still. Translation: the price hikes are a future earnings lever, not a past one. (FreightWaves Today | July 16)

The market noticed. On Stock Market Today With IBD (July 16), Ken Shreve summed up the mood: "It was J.B. Hunt last night… They had a good earnings report and a big bottom line and top line beat. And the fundamentals in the trucking area has been well documented in recent months." The transport ETF closed up 2.8%, and railroads "broke out to new highs." (Stock Market Today With IBD)

Spot rates are at levels we basically never see. On the same July 16 show, FreightWaves' Zach Strickland gave the number that stopped me: "Spot rates are in the 97th percentile of the last five years. And that includes COVID mind you. That's insane." He pegged the spot index at the 96.9th percentile and tender rejections (the share of loads carriers turn down, a tightness gauge) around 16%. FTR's own carrier-conditions index, discussed by FTR trucking VP Avery Weiss on the July 14 State of Freight, "jump[ed]… in May to an all-time high of 20.4," a reading that "indicates market conditions greatly favoring carriers." (FreightWaves Today | July 16; FTR | State of Freight)

But, and this is important, the very same week those rates fell hard. Weiss: "The total market broker posted rate fell just under $0.14 a mile week over week for the largest decrease in a single week since the first week of 2021, and the fourth largest weekly drop ever." Dry van and flatbed each dropped about 14.6 cents; reefer plunged 26 cents. He's careful to note these follow record-high prints the prior week, and year-over-year the numbers are still enormous (dry van up ~48% all-in, ~57% fuel-adjusted). By July 20, Fuller described the market as simply "quiet": "not a ton of really firmness and activity. Rejection rates are the 15s. Spot rates are… about 353. Again, these are high rates compared to where they have been… but not as strong and as robust as they have been." (FTR | State of Freight; FreightWaves Today | July 20)

Now the twist: almost none of this is being driven by the consumer. This was the single most important data point of the week, from Dr. Jason Miller of Michigan State (the supply-chain economist whose freight-volume work is widely followed), on the July 20 FreightWaves Today:

"Freight demand is up a call it a modest one and a half percent year over year… Major bifurcation based on whether shipper verticals provide to the CapEx portion of this sort of physical ecosystem for AI. So electrical good wholesalers, their sales are up 28% year over year… Battery shipments, large storage batteries production is up substantially… But then if you shift over more on consumer centric freight, beverage production is down about 4% year over year. Major appliance production down 8 to 10% year over year."

And then the line that should worry anyone long the group at these multiples: "Right now the fate of the trucking sector in terms of marginal increase in freight volumes is almost 100% tied to this physical ecosystem for AI computing. And so anything that would start to show cracks in that ecosystem would be a huge concern." (FreightWaves Today | July 20)

You can see the same signal in the odd corners of the network. On State of Freight, Weiss noted flatbed, usually a housing-construction read, and housing is weak, is holding up anyway "benefiting from massive development of data centers." Even a cross-border carrier owner in El Paso, Salvador of Vest Express, told The FreightCaviar Podcast that Juárez factories are shifting from autos to "a lot of the servers… going to these huge giga centers, data centers… Meta… Apple… Amazon… Tesla goods." The AI build-out isn't an abstraction in the freight data; it's showing up on flatbeds and at the border. (FTR | State of Freight; The FreightCaviar Podcast)

A lot of the rate "strength" is actually a fuel bill. The Iran conflict and the on-again-off-again Strait of Hormuz threat sent diesel vertical. Weiss: "The national average retail price of on-highway diesel jumped 21.8 cents to $4.796 a gallon during the week ended July 13. That is the fifth largest increase in diesel prices this year, but it is the 13th largest in the entire history of the data set, which goes back to 1993." Crude jumped $7.70 in four sessions after the ceasefire "essentially collapsed." On Logistics Matters (July 17), DC Velocity editor Ben Ames, citing the AFS Logistics/TD Cowen quarterly report, connected the dots on why headline rates look so strong: "Diesel prices rose about 51% [in Q2] compared to January and February levels… Truckload rates jumped to their highest level in the second quarter for 15 quarters. That came despite modest cargo demand… LTL fuel surcharge… rose over 60% above June 2025 levels." Read that again: highest truckload rates in nearly four years, on modest demand, because of fuel and driver scarcity. (FTR | State of Freight; Logistics Matters with DC VELOCITY)

That same Logistics Matters read had the tell-tale LTL detail: weight per shipment is falling. Less freight in each less-than-truckload shipment "reflects continued softness in industrial and manufacturing demand." So the industrial economy that LTL carriers like Old Dominion and Saia serve is still soft, the rate line is being held up by the fuel surcharge, not the box getting heavier.

Capacity is leaving the market, but for reasons that have nothing to do with the economy. This is the quiet structural story, and it's the strongest argument that the tightness is durable. On the July 20 FreightWaves Today, FreightWaves' Phil Brink laid out the enforcement wave: "The Department of Homeland Security is teaming up with the Department of Transportation to investigate roughly 75 CDL training schools suspected of fraud… DOT says it's already pulled more than 24,000 drivers for failing English language requirements and canceled more than 28,000 fraudulently issued licenses." NFI CEO Sid Brown, on the July 14 show, described what that feels like on the ground: "With all the driver schools closing down, with the government cracking down on all the English requirement languages… It's just cut down the supply of available drivers… for the first time in a couple years, we're actually starting to see people pay sign-on bonuses." A separate driver-pay index cited on that episode (baseline 100 in January 2020) has climbed to 170, up 70%, and bottomed only back in 2023. (FreightWaves Today | July 20; FreightWaves Today | July 14)

Why this matters: in a normal cycle, high rates lure capacity back and kill the rally. If the drivers are being removed by regulation and can't easily come back, the supply side stays tight even if demand is only so-so. That's the bull case in one sentence.

Class 8 trucks: the 2027 emissions cliff got sanded down, and there's nothing left to build this year. Weiss walked through the EPA's proposed changes on State of Freight: the agency is keeping the January 1, 2027 NOx effective date but "plans to roll back the expanded warranty obligations on truck and engine manufacturers" and is adding "non-conformance penalties or NCPs, whereby manufacturers who need additional time… can continue to build and sell current technology trucks into 2027 by paying a substantial penalty for each truck." EPA's math: "a net of $12 billion over 10 years, which it says translates into about $6,000 a truck." Just as notable for the OEMs (Paccar, Cummins), "truck manufacturers have basically run out of build slots for the current year," so incremental demand pushes into 2027. And at least one major maker "does not plan to start building its new technology heavy-duty engines until well into next year." (FTR | State of Freight)

The rail merger got a real handicap. Talking Transports (July 14) hosted Farouk Bezar, a senior operating partner at Little John & Company and, tellingly, a former Chief Strategy Officer at CSX, on the proposed Union Pacific–Norfolk Southern tie-up:

"My gut tells me it is going to get approved. What I don't know is will there be conditions imposed?… Will it create a situation where the current participants, and I guess it's really driven by UP, no longer wants to do it. Like, if it's some type of rate cap, inter-switching, open gateways… it'll be up to them to decide if it's so severe that they don't want to move forward with it."

Bezar also poured a little cold water on the "everyone else has to merge too" reflex, noting the second transcontinental piece, BNSF, sits inside Berkshire Hathaway, "not a traditional public company… a business that has a much different opportunity cost." Don't assume BNSF is forced to respond on Wall Street's timetable. He tied the whole thing back to the cycle: capacity "is coming out of the market, like out of the truckload market. And that's going to help brokerage, and ultimately that's going to help intermodal as well." (Talking Transports)

Private equity is circling again. Same episode, Bezar: after "over 40 months" of freight recession, "at least over the last 60 to 90 days, things seem to be improving… if this continues, I think you'll see more deal flow." Where's the money going? "Definitely brokerage… private equity gravitates, at least in transportation logistics, more towards lower capital, asset-light, asset-based companies", not asset-heavy truckload, where consolidation is a strategic-buyer game (think Knight). (Talking Transports)

The import wave keeps building, and it's landing on the East Coast and the border. International container spot rates to Europe are "approaching $8,000… up 400%" from a couple months ago, per Fuller (July 20), and the trans-Pacific benchmark is above $7,700/TEU as shippers front-load ahead of tariff and Hormuz risk, pulling the traditional October peak into July. Mexico stayed the top U.S. trading partner in May at $87.23 billion two-way (+17% YoY), with Laredo up 19.4%. And a quirky read on the same disruption: United Airlines' Q2 cargo revenue jumped 22.6% to $527 million, its biggest cargo haul since the pandemic, as Middle East air capacity fell 12% and yields spiked. The freight is out there; the question is whether it reaches the truckers' hands or stays consumer-soft. (FreightWaves Today | July 20; FreightWaves Today | July 14)

The debate: real turn, or false start?

For once, the week's podcasts genuinely argue both sides, so here they are, steel-manned.

The bull case (this is the real turn). Spot rates at the 97th percentile of five years. Carrier-conditions index at an all-time high. Tender rejections stuck in the mid-teens, a tight-market reading. Capacity is being structurally removed by immigration and licensing enforcement that rates can't quickly reverse. Sign-on bonuses and a driver-pay index up 70% since 2020 confirm the labor squeeze is real, not seasonal. J.B. Hunt beat hard "this early in the cycle" and hasn't even used its intermodal pricing lever yet. Smart money, private equity, is starting to move after a 40-month drought. If you believe capacity stays constrained, operating leverage does the rest, and 2027 estimates are too low.

The bear case (false start). Underlying demand is up just 1.5%, and essentially all of the marginal growth is one theme, AI data-center construction. As Dr. Miller warned, "anything that would start to show cracks in that ecosystem would be a huge concern." A meaningful share of the rate strength is a diesel spike tied to a geopolitical flashpoint that could reverse as fast as it appeared. LTL weight-per-shipment is falling, telling you the industrial economy is still contracting. And spot rates just posted their fourth-largest weekly drop ever the same week everyone was celebrating the highs. Pay record multiples for peak-looking rates built on fuel and a single end-market, and you're buying the top.

My read: both are true, which is exactly why this is dangerous. The supply story (bull) is durable; the demand story (bear) is fragile and concentrated. This isn't a normal broad-based recovery, it's a tight-capacity market being fed by one hungry customer. That can run further than skeptics think and break faster than bulls expect. I'd rather own the names with a structural share-gain story (intermodal taking freight from truck) than the ones that simply need the tide to keep rising.

The names in play

  • J.B. Hunt (JBHT), the quarter of the week. Intermodal +22% revenue / +10% volume, eastern lanes leading, and, critically, pricing power still in reserve while peers hike 15%. The bull thesis is a coiled-spring earnings story, not a "beat and it's over" story. Watch whether that unused pricing lever shows up in guidance.

  • Knight-Swift (KNX), teed up as the next read. On Stock Market Today With IBD, Ken Shreve called it "a little bit of a turnaround story… earnings going forward are really expected to turn around," with a clean technical setup, and said flatly, "I think we'll see nice numbers from Knight Swift next week as well." Truckload-heavy, so it's the purest bet on whether the spot strength converts to earnings.

  • Union Pacific (UNP) and the rails, three rail reports land next week (UNP among them), and the group already "broke out to new highs." Layer on the UP/NS merger overhang, where a former CSX strategist thinks approval is likely but the conditions are the swing factor. Rails were otherwise quiet on the podcasts this week, the earnings will fill that vacuum fast.

  • The pure-play brokers (CHRW, LSTR, RXO, HUBG), the warning label. See read-throughs below.

  • Paccar (PCAR), Cummins (CMI), sold-out 2026 build slots plus a softened (not eliminated) 2027 emissions cliff. Demand that can't be filled this year is demand deferred, not lost.

Read-throughs

Freight brokers (CHRW, LSTR, RXO, HUBG): brace for the margin squeeze. J.B. Hunt's brokerage arm is the canary, and it's not singing. Per Hampstead on July 16: "ICS overall revenue up 49%. Volume up 19%, revenue per load up 26%, but margins compressed by 300 basis points year over year… made $1.7 million in operating income on $380 million." Revenue up nearly half, profit rounding to zero. When truckload rates rip, brokers have to buy capacity in a spot market at the 97th percentile and can't reprice the customer fast enough. Fuller's takeaway: "It's going to be interesting to see what the pure play brokerages do… J.B. Hunt is showing the pressure that brokerages as a whole are under." The asset-integrated model is winning; the asset-light middlemen are getting pinched exactly when volumes look great. Don't confuse a broker's rising revenue line with a rising profit line this quarter. (FreightWaves Today | July 16)

Class 8 OEMs and dealers (PCAR, CMI, ALSN, RUSHA): supply-constrained, not demand-constrained. Build slots are gone for 2026 and the 2027 pre-buy math just got clarified by EPA. The nuance from Weiss, at least one major maker won't build its new-tech engines "until well into next year", is a competitive wrinkle worth watching; it could force rivals to move, or leave a gap in new-tech supply. Either way, backlog visibility looks better than the sluggish freight economy would suggest.

Railcar lessors and builders (GATX, TRN, GBX): nothing this week. No substantive podcast coverage of lease rates, utilization, or order activity in the window. Given how loud the intermodal-share story is, the absence is itself worth flagging, the conversation is all about containers and boxes on flatcars, not the cars themselves.

Shippers, coal, grain, chemicals: quiet, with one structural note. No fresh cyclical reads this week. The only durable point came from Bezar, explaining CSX's Quality Carriers logic: converting chemical freight off the highway onto rail, in a business where "coal [is in] secular decline" and "chemicals was a large and profitable business." The modal-shift thesis, truck-to-rail, keeps showing up as the connective tissue between an expensive truckload market and the rails' volume story.

What changed from last week

The frame flipped. A cycle that spent three-plus years as a bottoming call is now, per the data, a turn, spot rates at multi-year extremes, an all-time-high conditions index, and J.B. Hunt confirming it in hard numbers. What's new and genuinely different is the diagnosis of why: this is a supply-starved market feeding a single, AI-shaped source of demand, propped up further by a fuel spike. The bull-bear debate stopped being "has it bottomed" and became "how much of this is real, and how long does the one customer keep buying." That's a more advanced, and more precarious, stage of the cycle than we were discussing a week ago.