S13.2 · Transportation, Logistics & Mobility
Asset-light load-matching intermediaries whose take-rate margin is the one revenue model in the sector agentic AI can directly displace.
Freight brokerage is the asset-light intermediary layer between truck capacity and shippers, and it carries a distinction no other transportation segment does: agentic AI here goes after the revenue model itself, not the cost base. Sizing it cleanly is difficult — no single figure isolates brokerage alone — but US third-party logistics gross revenue of roughly $213B (Armstrong & Associates, most recent estimate) is a reasonable proxy, most of it non-asset transportation management. Digital freight brokerage specifically ran $19-20B in 2024, with one forecast reaching $78.3B by 2035 (Precedence Research) — a ~27% CAGR (Market.us) that captures two things at once: underlying volume growth and traditional brokerage migrating onto digital rails.
The economics start with the take rate. Brokers earn a gross margin of 12-16% of freight spend for matching shippers with carrier capacity and negotiating rate — a high-volume, transactional, asset-light business with almost no regulatory gate. A $75K FMCSA bond is the principal barrier to entry, which is why the market supports both a set of players with meaningful scale (C.H. Robinson, Uber Freight, RXO, Total Quality Logistics, Echo) and a long tail of thousands of small brokers. Volumes do not move independently; they track the trucking freight cycle, so traditional brokerage was down through 2023-2024 alongside the broader freight recession and is recovering with it. The segment sits downstream of carrier capacity and upstream of shippers across manufacturing, retail and other freight-generating sectors. It is a coordination layer, not a physical-asset business, and its hubs — Chicago, Chattanooga, Kansas City — are logistics-network hubs rather than manufacturing centers.
The take rate is compensation for exactly two functions: matching a load to available capacity, and negotiating the price. Agentic systems now perform both directly — automated quoting, automated carrier sourcing, load-matching bots that never put a human broker on the phone. What historically justified the margin was information asymmetry: brokers held relationships and load-board visibility that neither shippers nor carriers could access directly. Once AI commoditizes matching, the asymmetry collapses — and what goes with it is the economic basis for the margin, not merely its size. A 2023 shutdown of a prominent asset-light freight-tech operator was an early casualty of this dynamic; the pattern it signaled is a structural attack on the brokering function itself, distinct from the tooling that supports it.
That is what separates the disruption case here from the rest of the sector. Elsewhere in transportation, AI compresses a cost line — fuel, maintenance, labor — inside a revenue model that stays intact. In brokerage, AI substitutes for the revenue-generating function directly. Brokers integrating TMS and visibility software to protect margin are mitigating, not reversing. The more durable response is expansion into managed transportation and cross-border brokerage, which moves revenue toward services less exposed to pure load-matching automation.
For spot and digital lanes the shift is already under way. The broader expectation is that within 2-5 years the majority of spot freight clears with minimal human touch. Take rates compress industry-wide at that point, and the sorting variable becomes proprietary automation — brokers who own it and brokers who do not will not be running the same business.