The data-centre construction boom is doing two things to air cargo at once: it is adding freight tonnage to a market that had little spare capacity, and it is deepening the directional imbalances that make capacity planning and rate-setting harder for carriers, forwarders and shippers alike.

That is the core finding reported by The Loadstar, which identifies hyperscale data-centre buildouts as a growing demand driver for air freight. The equipment that fills these facilities — servers, networking hardware, power and cooling infrastructure — moves by air when deployment timelines are tight, and the current wave of artificial-intelligence-driven construction has made those timelines tighter than ever.

For airlines and freighter operators, the demand is welcome but awkward. Data-centre cargo is heavy, dense and often time-critical, which makes it attractive revenue per kilogramme. But it does not spread evenly across a network. It concentrates into lanes serving data-centre hubs and the manufacturing centres that feed them, soaking up freighter and belly capacity on those routes while leaving return legs underutilised. That imbalance is the second half of the story: the same boom that lifts volumes also distorts them, pushing up rates on the constrained headhaul legs and leaving carriers to reposition or discount capacity on the backhaul.

The commercial consequences ripple down the chain. Forwarders competing to move high-value technology hardware face thinner capacity on the lanes that matter most, and shippers of more conventional air freight — electronics, automotive parts, e-commerce — find themselves bidding against AI infrastructure projects for the same belly space. When a single hyperscale deployment can require multiple freighter loads, the capacity squeeze can move spot rates quickly. Shippers with fixed allocations gain; those buying on the spot market carry the risk.

For carriers, the strategic question is how long the surge lasts. Data-centre construction is a project-based business, not a steady baseline like e-commerce. Airlines that add freighter capacity to serve it today may find the directional imbalance reverses or the volumes taper once a buildout wave completes. The Loadstar's framing — growth paired with widening imbalance — captures exactly that dual character: a demand tailwind that carries structural risk.

It also lands at a moment when air cargo capacity is already stretched across many trade lanes by e-commerce volumes out of Asia and by re-routings and disruptions elsewhere in the supply chain. Any incremental demand stream with uneven geographic distribution compounds the scheduling problem for integrators and combination carriers, which must decide where to position aircraft and how to price directional capacity without destabilising contract rates.

Shippers of data-centre equipment itself face their own calculus. The cargo is high-value and often tied to commissioning deadlines at the destination facility, which weakens their hand in rate negotiations. Where air freight is the only mode that meets an energisation date, price sensitivity drops sharply — one reason carriers can monetise this traffic so effectively.

The wider implication for the freight market is that AI infrastructure spending is no longer just a technology story; it is now a visible input into air cargo demand. Procurement teams that model air freight capacity on historical seasonality may need to add a variable for hyperscale construction schedules, because those projects can arrive as sudden, lumpy surges rather than predictable monthly flows.

The Loadstar's report suggests the trend has room to run: as long as data-centre construction continues at pace, air cargo operators should expect the growth to keep arriving unevenly. The lanes that serve the boom will stay tight, the backhauls will stay soft, and the imbalance between the two will remain a pricing problem for the market to solve.