Of all the supply chain disruptions affecting hyperscale data center development, none has proven as structurally persistent — or as consequential — as the shortage of large power transformers. What began as a tightening market has become, by most industry accounts, the defining procurement constraint of the current development cycle.
The Numbers Behind the Bottleneck
According to Wood Mackenzie's industry survey data, standard power transformers now average roughly 128 weeks for delivery, while generator step-up transformers — the units that connect large power generation or industrial loads directly to the grid — average around 144 weeks, with some specialised high-voltage orders extending considerably further, in certain cases to three or four years. Demand for generator step-up transformers has grown by approximately 274% since 2019, while demand for standard power transformers has grown by roughly 119% over the same period — growth rates that have sharply outpaced the expansion of manufacturing capacity.
The underlying material constraints compound the problem. Grain-oriented electrical steel, the specialised material required for transformer cores, has seen prices roughly double since 2020, while copper prices have risen more than 50% over the same period. A significant share of large power transformers used in markets like the United States are imported, exposing already-stretched supply chains to additional geopolitical and trade policy risk.
Three Demand Categories Competing for the Same Capacity
- AI and hyperscale data center construction: Each large facility requires transformers at multiple points in its distribution chain, often totalling hundreds of megawatts of capacity per site
- Industrial electrification: Manufacturing facilities converting from fossil-fuel-powered processes to electric systems require new transformer capacity independent of data center demand
- Grid modernisation: Aging transmission and distribution infrastructure requires transformer replacement at a scale that was already behind schedule before AI-driven demand arrived
These three demand sources do not share a common peak cycle — they are all active simultaneously, competing for output from a manufacturing base that cannot expand quickly given the specialised engineering, testing, and factory capacity large transformers require.
When equipment availability becomes the gating factor instead of capital or permitting, the entire project timeline is dictated by supply chains the developer cannot control.
How the Industry Is Responding
Manufacturers are investing significantly in new capacity — well over a billion dollars in some single commitments — though new factories typically take years, not months, to come online, meaning near-term relief remains limited. In the meantime, developers and utilities are adapting procurement practices: ordering equipment years in advance of confirmed need, exploring engineering substitutions where feasible, and in some cases pursuing alternative power delivery architectures — including onsite generation — specifically to reduce dependence on grid-side transformer capacity that is subject to the same constrained supply chain.
Domestic Capacity Expansion Is Underway, but Will Take Years
Major manufacturers including Hitachi Energy and Siemens have announced significant new manufacturing investments specifically targeting transformer production capacity, with new facilities expected to come online over the next several years. While these investments are a meaningful signal that the supply chain is responding to sustained demand growth, the lead time for bringing new heavy-equipment manufacturing capacity online is itself measured in years, meaning the current shortage is likely to persist as a structural feature of the market well into the latter part of this decade, even as conditions gradually improve.
This has led some developers and large equipment buyers to consider longer-term supply agreements or even direct investment in dedicated manufacturing capacity, a strategy more commonly associated with industrial procurement than traditional real estate development — further evidence of how thoroughly AI infrastructure development has come to resemble industrial supply chain management.
What This Means for Project Planning
Given lead times that can exceed the typical construction schedule for the rest of a facility, equipment procurement decisions increasingly need to precede, rather than follow, full financial close and even full design completion in some cases. Developers who treat transformer and switchgear ordering as an early, parallel-path activity — rather than a sequential step following design finalisation — are materially better positioned to hit realistic energisation dates.
DATAPERT incorporates electrical equipment procurement risk directly into the feasibility and programme planning work we conduct with clients, alongside our broader data center development advisory. Start a project to discuss equipment strategy for an active programme.
