White Papers AI Boom Redraws the Air Cargo Landscape

Key Concepts and Definitions

AI-Related Cargo Categories

Semiconductors

In this white paper, this cargo category includes wafers, chips and integrated circuits (ICs), printed circuit boards (PCBs), and semiconductor manufacturing equipment. It supports the production of AI infrastructure and devices and has a relatively stable demand base.

AI Infrastructure

This category covers the equipment needed to build data centers, including servers, storage systems, racks, server components, power and cooling systems, and networking equipment. Expanding data center investment is making it the main driver of volume growth in AI-related air cargo.

AI Devices

These are end-user products with AI capabilities, such as smartphones and laptops (PCs). As memory supply is increasingly directed toward data centers, memory supply constraints are affecting device production and shipments.

Supply Chain and Transportation Strategy Concepts

Big 4 and Capital Expenditure (CapEx)

The Big 4 refers to the four major hyperscalers: Amazon (AWS), Google, Meta, and Microsoft. Their investment in data center facilities influences demand for servers, racks, power and cooling equipment, and air transportation.

China+X

A supply chain restructuring strategy that diversifies production concentrated in China into neighboring countries such as Vietnam. The relocation of production and assembly sites changes the product mix and cargo flows of individual countries.

HBM (High Bandwidth Memory)

High-performance memory essential for AI chip processing and a key component determining AI server performance. It supports Korea's AI semiconductor-related exports and logistics demand.

Lead Time

The time from order placement to actual shipment or delivery. AI servers have lead times of approximately 35-40 weeks, creating a lag before changes in investment are reflected in actual shipments and transportation demand.

Modal Shift

A change in the mode of cargo transportation, such as from air to ocean. If delays in data center construction reduce shipment urgency, shippers may switch to ocean transport for greater cost efficiency.

Long-Term Agreement (LTA)

A contract in which a supplier and buyer agree on volumes and prices several years in advance. The shift toward long-term semiconductor supply agreements is prompting shippers to consider medium- to long-term transportation contracts and capacity planning.

The AI Boom and Air Cargo Market in Key Questions

• Q1.

  How is the AI boom changing the air cargo market?

  The focus of air cargo is expanding beyond traditional semiconductors and electronic products to include data center servers, racks, and power and cooling equipment. Over the past three years, the share of AI-related cargo in total air cargo has increased from 14.7% to 18.1%. AI infrastructure cargo, in particular, is expanding in product variety and driving overall cargo volume growth.

• Q2.

  Are all AI-related product categories growing at the same pace?

  The direction and pace differ by category. Semiconductors continue to grow steadily, while AI infrastructure is expanding rapidly on the back of increased data center investment. AI devices face pressure on shipments as memory supply is increasingly directed toward data centers. The white paper forecasts that AI infrastructure growth will peak in 2026 in terms of its growth rate. Cargo volumes may continue to rise afterward, but at a slower pace.

• Q3.

  Why do cargo trends differ among countries within Asia?

  Each country has a different industrial base and product focus. Taiwan is strong in server-led AI infrastructure, Thailand in server assembly and storage devices, and Vietnam in AI infrastructure and the assembly of finished products such as PCs. Malaysia is a hub for power semiconductors and server and rack component assembly, while Korea has a high share of advanced semiconductors and HBM. In China, AI production is increasingly absorbed by the domestic market, contributing to declining export cargo volumes.

• Q4.

  How long will Taiwan and Thailand trade lanes maintain strong growth?

  The white paper forecasts that trade lanes from Taiwan and Thailand, which have a high concentration of AI infrastructure cargo, will remain strong through the first half of 2027 before growth gradually moderates. Key variables include slowing growth in Big 4 capital expenditure and AI server lead times of approximately 35-40 weeks. Shippers need to reflect both current high growth and subsequent demand changes in contract durations and freight rate terms.

• Q5.

  Could cargo shift from air to ocean even as data center investment expands?

  Delays to data center construction or operating schedules can reduce the need to transport equipment urgently. This may encourage a shift to ocean transport for greater cost efficiency. Shippers therefore need to monitor not only investment levels but also actual project schedules and changes in equipment delivery dates.

• Q6.

  How do the medium- to long-term outlooks for Vietnam and Korea differ from China?

  Vietnam is expected to maintain overall cargo volume growth as AI infrastructure growth offsets demand adjustments in devices. In Korea, semiconductor export volumes are expected to remain strong through 2029, supported by semiconductor and memory demand and the expansion of long-term supply agreements. China, by contrast, may see continued structural adjustment in U.S.-bound export cargo as AI device volumes decline and more AI infrastructure output is absorbed domestically.

• Q7.

  What air logistics strategies should shippers prepare?

  Shippers should tailor transportation strategies to their cargo categories and countries of origin. For Taiwan and Thailand trade lanes, flexible contracts can help address slower growth and a potential shift to ocean transport. For Vietnam and Korea, shippers should consider securing medium- to long-term transportation capacity and logistics partnerships. Plans for China-related cargo should reflect changes in export demand as more output is absorbed domestically. For data center equipment, shippers should regularly check for delays to customers' construction and operating schedules.