S10.11 · Food, Agriculture & Beverage
An asset-heavy cold-storage and logistics network where AI-driven routing and warehouse automation is displacing manual regional expertise.
Global cold chain logistics runs roughly $300-350B (2024), with forecasts reaching somewhere between $1.0-1.5T by the mid-2030s depending on methodology (Precedence Research, Grand View Research, 2025). Food and agriculture account for the majority of that volume, ahead of pharmaceutical cold chain. This is an asset-heavy infrastructure business — the pallet positions and refrigerated fleets are the product — and AI's clearest current effect is on the labor cost of routing and warehouse operations, not on the core cold-storage asset itself.
Growth runs high-single-digit, driven by rising fresh and frozen food demand and by cold-chain infrastructure buildout in emerging markets. The densest existing infrastructure sits in the US, EU, China and Japan; the fastest capacity growth is in India, Southeast Asia and Latin America, where buildout is still catching up to demand. Globally the industry remains fragmented: Americold and Lineage Logistics are the largest players following a period of consolidation, operating alongside regional third-party logistics providers and grower-owned cooperatives, and the folding of that fragmented regional layer into scaled networks continues.
The segment sits downstream of production and processing and upstream of retail and foodservice distribution. The economics are utilization economics: warehousing and refrigerated trucking drive throughput- and utilization-based margins on a heavy asset base. Some revenue is contracted on a take-or-pay basis, which insulates part of the book from spot-market volatility, but fuel and energy cost pass-through to customers is imperfect — operators wear input-cost swings the contract structure does not fully offset.
For the largest players, rolling fragmented regional capacity into scaled networks is where the growth is: a network runs at better utilization than the sum of its parts, and utilization is the margin driver in this business. Cold-storage investment by grower cooperatives serves a different purpose — protection against spoilage risk, not a growth bet.
AI concentrates on two cost lines: route and network optimization, and warehouse labor. AI-driven routing software and automated storage-and-retrieval systems are collapsing the cost of the planning and physical-handling work that regional logistics operators have historically optimized manually, on accumulated local knowledge of routes, customers and seasonal patterns. That local knowledge was a real moat — a well-run regional operator won simply by knowing its territory better than anyone else — and software can now approximate or exceed it for routing and slotting decisions, though it does not replace the physical infrastructure itself. The systems are already running at the largest scaled players, Lineage and Americold among them. For mid-market operators, who face a real fixed-cost hurdle in deploying them, broad adoption is more likely 2-5 years out.