Set as preferred source
Insights from Forged in Disruption: State of Logistics Report.
Global logistics is not moving through another conventional freight cycle. It is settling into an operating environment in which trade policy, geopolitical disruption, energy exposure, regulation, and technology continuously reshape the economics of moving goods.
That distinction matters. A cyclical downturn can be managed through pricing discipline, capacity reductions, and patience. Structural volatility requires something different: networks capable of changing without accumulating excessive cost, inventory, or operational complexity. The source of advantage is moving from owning the largest network to operating the most adaptable one.
A stable headline is hiding an unstable system
US business logistics costs declined 1% in 2025 to approximately $2.41 trillion. On the surface, that suggests a market moving toward normalization. Beneath the headline, however, costs are being redistributed rather than removed.
Parcel expenditure increased 5.2%, storage rose 4.2%, and motor-carrier costs increased even as water transportation expenditure fell 36%. Financial inventory costs declined, but total inventory carrying costs remained above $730 billion. Logistics inflation is no longer moving uniformly through the system. It is concentrating in particular services, corridors, shipment profiles, and inventory positions.
This weakens the value of national averages. A market can appear soft while a critical lane is tightening, an essential gateway is congested, or a regulatory change is making a previously viable flow uneconomic. The relevant question is increasingly not whether capacity exists, but whether usable capacity exists at the required location, cost, and risk level.
Ocean freight illustrates the contradiction. Global shipping remains structurally oversupplied, with container demand expected to grow only 1–2% in 2026 while additional vessels enter the market. Industry EBIT could fall from approximately $32 billion in 2025 to roughly $1 billion in 2026. Yet Red Sea diversions, port congestion, slower sailing speeds, and other chokepoints can reduce effective capacity by 8–10%.
Excess supply and operational scarcity can therefore coexist. This is why market timing is becoming less valuable than contractual flexibility and corridor-level intelligence.
Trade policy has become a decision-frequency problem
The reorganization of trade is no longer centered only on labor costs or manufacturing efficiency. Supply chains are increasingly being designed around political alignment, tariff exposure, sanctions risk, and access to strategically important markets.
US tariff policy changed on average every 1.5 weeks during 2025. Companies that began the year with models based on historical tariff structures found those models outdated within weeks. The challenge was not simply the level of tariffs. It was the frequency with which sourcing, classification, inventory, and routing assumptions had to be reconsidered.
China+1 strategies are now visible in regional growth and freight patterns. Vietnam is projected to grow 7.2% in 2026, India 6.2%, and Indonesia 5%, while European growth remains around 1%. Mexico has become the largest US goods trading partner, and Southeast Asian hubs are absorbing manufacturing and logistics activity previously concentrated in China.
This diversification improves resilience to a single-country shock, but it does not recreate China’s historical combination of manufacturing scale, port density, supplier concentration, and transport efficiency. More origins mean more border crossings, more suppliers, more documentation, and more opportunities for inventory to become trapped between policy regimes.
A second risk follows from this adjustment: network drift. A new supplier is added to reduce tariff exposure. A temporary warehouse becomes permanent. Additional inventory is positioned near a border. Each decision may be rational in isolation, yet collectively they can produce a network that is harder to understand and more expensive to operate.
The cost of inaction remains significant. But the cost of repeated, disconnected action is now equally material.
Energy exposure is moving onto the balance sheet
The Strait of Hormuz carries approximately 20 million barrels of oil per day and around 20% of global LNG trade. Its vulnerability compounds disruption in the Red Sea and adds pressure to tanker rates, insurance premiums, fuel surcharges, and alternative transport corridors.
The consequences extend beyond shipping. A 10% increase in oil prices could add about 0.4 percentage points to global inflation. Higher fuel costs affect parcel pricing, trucking margins, aviation surcharges, inventory carrying costs, and consumer demand at the same time.
Energy is therefore no longer only an operating expense managed through procurement. It increasingly shapes facility location, supplier selection, contract terms, working-capital requirements, and the economic case for automation or electrification.
This is also changing the logic of sustainability investment. The prior model often depended on paying a green premium in anticipation of future policy or customer expectations. That approach is becoming harder to sustain under tighter capital conditions.
The more durable model connects emissions reduction to productivity: fuller loads, fewer empty miles, lower idle time, better routing, slower steaming where commercially viable, and more precise inventory placement. These measures reduce cost and emissions simultaneously. Sustainability becomes less dependent on a separate investment case because it is embedded in asset utilization and operating resilience.
Infrastructure constraints are narrowing the investment field
Global government debt is expected to exceed 100% of GDP by 2029. Public budgets are becoming less capable of funding the roads, ports, rail systems, power infrastructure, and charging networks required by more regionalized and automated supply chains.
Private investment is being repriced at the same time. Fleet renewal, new warehouses, robotics, and terminal modernization must clear higher return thresholds while demand remains uneven. The result is not an end to logistics investment, but a shift toward assets with measurable payback, high utilization, and the ability to support several operating scenarios.
The warehouse market shows this separation. US industrial vacancy reached 7.4% in early 2026, while new construction fell to its lowest level since 2017. That might suggest broadly weaker demand. In practice, occupiers continue to favor large, modern, automation-ready facilities with reliable access to power.
Seventy-six percent of surveyed organizations expect warehouse power requirements to increase by 10–50% over the next five years. Energy capacity is already outranking labor cost as a location consideration in parts of the market. The emerging infrastructure bottleneck is not simply a shortage of square footage. It is a shortage of facilities that can support automation, distributed inventory, and more demanding energy loads.
This is widening the valuation gap between modern logistics infrastructure and legacy buildings that are difficult to retrofit.
AI is progressing from observation to execution
The most useful way to understand AI in logistics is not by model type, but by the function it performs: interpret, predict, recommend, and execute.
Interpretation converts emails, documents, telematics, sensor readings, and shipment events into usable information. Prediction identifies delays, demand changes, equipment failure, or delivery risk. Recommendation determines the best response. Execution completes the transaction through rerouting, booking, invoice validation, appointment scheduling, or inventory movement.
Most organizations have made progress in the first two stages. The more significant operating leverage is now appearing in recommendation and execution, where AI moves from providing information to closing routine decisions.
FedEx estimates that its predictive maintenance platform has prevented approximately 17,000 hours of potential downtime and generated about $10 million in annual savings. C.H. Robinson has automated millions of shipping tasks, including quotes, load tenders, and appointments, while attributing part of a 30% productivity increase across 2023 and 2024 to generative AI. Penske expects 30–40% productivity improvements from an agentic system that validates shipment status across roughly 600,000 loads.
The pattern is consistent. Returns are strongest in repetitive workflows with clear economics, accessible data, and a defined operational response. AI produces less value when added to fragmented processes without changing how work moves through the organization.
The constraint is increasingly data legibility. The average company has mapped only 15–20% of its extended value chain with enough detail to assess ownership, tariff, material, and sourcing exposure. More sophisticated models cannot compensate for a network that remains largely opaque.
The labor bottleneck is becoming more specialized
Warehouse employment has stabilized at approximately 1.8 million to 1.9 million workers, but the underlying shortage has changed. The most difficult positions to fill are increasingly automation technicians, maintenance specialists, technically capable supervisors, and staff able to manage exceptions across integrated systems.
Turnover still exceeds 40–50% annually in many facilities. Automation reduces the dependence on labor for repetitive movement, but it increases dependence on a smaller group of people capable of maintaining and improving automated operations.
This creates a less visible constraint on scaling. Capital may be available for robotics or AI, yet productivity gains remain limited when technical expertise is concentrated at a few sites or held by external vendors. The workforce question is shifting from broad hiring capacity toward the retention and distribution of operational knowledge.
Market boundaries are beginning to collapse
Amazon’s expansion into full-service logistics demonstrates how quickly conventional industry categories are losing relevance. Through Amazon Shipping and Amazon Supply Chain Services, the company is extending its fulfillment, transportation, and delivery infrastructure to external merchants, including businesses outside its traditional fulfillment ecosystem.
This is not simply another parcel service. Amazon combines merchant demand data, inventory visibility, warehouse capacity, transportation assets, routing intelligence, and direct access to the customer interface. It competes across the order-to-delivery chain rather than within a single logistics mode.
Traditional 3PLs are moving in the same direction from a different starting point. Customs, warehousing, freight forwarding, compliance, financing, and transport management are converging into integrated orchestration. As basic transportation pricing becomes more transparent, margin increasingly sits in coordinating the system rather than selling access to one part of it.
Rail consolidation could further alter the structure of US logistics. The proposed Union Pacific–Norfolk Southern combination would create the first coast-to-coast single-line railroad, covering more than 50,000 route miles across 43 states. Removing interchange dwell could improve transcontinental transit times and make rail more competitive with long-haul trucking.
But the transaction remains a regulatory thesis, not an operating reality. The Surface Transportation Board rejected the initial application as incomplete in January 2026, citing missing forward-looking market-share projections and parts of the merger agreement. A revised application followed in April, making the regulatory calendar the most consequential variable for the rail market.
Approval could improve service and shift significant freight from road to rail. Conditions, delays, or rejection could produce a very different competitive structure. Shipper concerns over pricing, service concentration, and integration disruption remain part of the economic equation.
Regulation is also changing access to capacity
The US Supreme Court’s ruling in Montgomery v. Caribe Transport II, LLC adds another structural pressure. State-law negligent-hiring claims against freight brokers are not pre-empted when they relate to motor-vehicle safety.
The ruling does not make brokers automatically responsible for every carrier accident. It does, however, increase the commercial value of documented carrier selection, insurance verification, safety performance, and auditable compliance processes.
This is likely to affect smaller brokers and carriers most. Larger platforms can spread compliance costs across higher transaction volumes and invest in more sophisticated vetting systems. Smaller providers may find access to freight increasingly dependent on the quality and traceability of their safety records.
Carrier selection is therefore becoming more than a rate and service decision. It is becoming a liability-allocation mechanism.
Adaptability is becoming the economic center of logistics
The defining divide in logistics is no longer between asset-heavy and asset-light models. It is between systems designed for periodic optimization and those capable of continuous adjustment.
Physical scale still matters. It creates density, purchasing leverage, and service reach. But scale becomes less defensible when it carries rigid infrastructure, fragmented data, or slow decision processes. Smaller providers can compete where lower technical debt allows faster integration, while platforms can control freight flows without owning every asset involved.
The capital premium will increasingly accrue to infrastructure that remains useful across several demand environments. The operating premium will accrue to networks that can detect change, understand exposure, and act before local adjustments accumulate into systemic cost.
Volatility is no longer the exceptional condition around which logistics networks must build contingency plans. It is becoming the environment in which those networks are expected to perform.
