Transformer lead times and the 2026 supply gap

Wood Mackenzie's Q2 2025 transformer market data puts average lead times at 128 weeks for power transformers and 144 weeks for generator step-up (GSU) transformers — between two and a half and three years from order to delivery. The same data set estimates a supply gap of roughly 30% for power transformers in 2025, and about 10% for distribution transformers.

Two things follow. A lead time of 128 weeks is an average, not a ceiling: high-voltage units with unusual specifications, or units placed into an order book that is already full, run longer. And a 30% gap is not a scheduling inconvenience. It means that in a given year, demand for power transformers exceeds what the industry can deliver, and the shortfall is resolved by price, by queue position, or by projects moving to the right.

Demand rose on four fronts at once. Grid investment increased as utilities in North America, Europe and Australia began replacing assets commissioned in the 1970s and 1980s. Renewable capacity adds more than one large unit per project — a generator step-up or collector transformer at the point of generation, and a main step-up transformer at the point of interconnection — so each gigawatt of new capacity occupies a queue of its own. Data centre load is the third and least deferrable: the IEA's Energy and AI report puts data centre electricity consumption at about 415 TWh in 2024, roughly 1.5% of global demand, rising above 945 TWh by 2030.

Supply did not move at the same speed. Grain-oriented electrical steel, copper winding wire, high-voltage bushings and on-load tap changers are each concentrated in a small number of qualified suppliers. Skilled labour — winding, core stacking, test engineering — cannot be added in a quarter. A new high-voltage transformer plant takes years to build and to qualify its first units, so in the near term the supply curve is close to vertical.

Lead time becomes a financial number at the point where it moves the commercial operation date. Take an illustrative 200 MWac solar plant with a 22% net capacity factor and an offtake price of USD 40/MWh. Annual generation is about 385 GWh; six months of lost production is roughly 193 GWh, or about USD 7.7 million of revenue that is not deferred but never generated — the resource does not wait for the transformer.

Add financing. If USD 120 million of debt has been drawn at 7%, six months of interest is about USD 4.2 million with no offsetting revenue. The two effects together come to roughly USD 12 million on a single project, before liquidated damages, PPA milestone clauses or the cost of re-booking transmission capacity.

The assumptions are illustrative rather than a forecast for any market, but the structure holds. The cost of a delay is dominated by lost generation and carried financing, not by the difference between two bids. A bid that is 4% cheaper and eight weeks later is usually the more expensive bid.

The measures that work are commercial, and they are taken early. Reserve capacity before final investment decision: a reservation agreement with a defined cancellation ladder costs less than a delayed COD, and it fixes a position in the queue rather than a price. Use framework agreements across a portfolio: committing several projects at once gives a manufacturer a reason to hold slots and to buy long-lead components in advance. Split delivery by milestone: a substation with two main transformers can often be energised in stages, and taking the first unit early converts part of the programme into revenue. Accept an alternative specification where it is technically equivalent: a standard impedance instead of a bespoke one, a standard rating instead of an oversize, or a different bushing family can move a unit into a slot that already exists. And keep the transformer off the critical path: sequence civil works, structures and cabling so that a late unit delays as few activities as possible.

From the manufacturer's side, a quoted lead time is a chain of dependent steps. Design and drawing approval comes first, and it cannot start until the specification is frozen. Raw material procurement follows: grain-oriented electrical steel, copper, tank steel, bushings, tap changer, and for high-voltage units the insulation and lead-exit assemblies that are made to order. Manufacturing covers winding, core stacking, tank fabrication, oil processing and vacuum drying. Testing covers routine tests on every unit and, where required, type and special tests — lightning impulse, temperature rise and partial discharge measurement. Shipping is a stage in its own right: a large oil-filled unit is heavy and out-of-gauge, and both the sea freight and the final inland movement need permits and a route survey.

Only part of that chain is factory time. Weeks spent in specification review, in drawing approval cycles, or in changing the impedance after the order is placed, are weeks added to the end of the programme. The most effective single action a buyer can take to protect a delivery date is to freeze the specification before the order, and to schedule the review cycle rather than treat it as an interruption to the schedule.

A delivery date is only as good as the capacity behind it. Five questions separate a commitment from a hope. What is the manufacturer's annual capacity in MVA, and how much of it is already committed? How many units are scheduled in the window being quoted, and where does this order sit in that sequence? Are the critical components — bushings, tap changer, grain-oriented steel — held under framework contracts or bought on the spot market? Is test capacity available, because an impulse generator and a controlled partial discharge background cannot be rented at short notice? And are the milestone dates written into the contract with remedies attached?

Test capability is a useful proxy for the rest. Our two impulse generator systems cover products rated up to 220 kV at 1200 kV and above 220 kV at 4800 kV; against the highest lightning impulse withstand level for the 500 kV class in IEC 60071-1 — 1675 kV at Um = 550 kV — that leaves a margin of about 2.9×. Partial discharge background is ≤10 pC at the Harbin oil-immersed base and ≤1 pC at the Hai'an dry-type base, against the 100 pC limit that IEC 60076-3 applies to units with Um above 72.5 kV. The laboratory is CNAS-accredited under registration L5393.

We do not publish a single standard lead time, because the number depends on voltage class, rating, specification maturity and the order book at the time of enquiry. Any manufacturer that quotes one figure without first reading your specification is quoting an average, not a commitment. What we can state is capacity: 40,000 MVA a year across the two manufacturing bases.