Methodology

Source hierarchy

When two sources give the same figure, we cite the one higher on this list. When they disagree, we keep both rows and say so on the conflicts list.

1. Primary survey and index data (22 rows): Wood Mackenzie’s lead-time benchmark as published in its own executive summary, and the BLS producer price index via FRED.
2. Government, labs and advisory bodies (6 rows): CRS, NREL, NIAC.
3. Industry associations (6 rows): APPA surveys of public power utilities, NEMA.
4. Manufacturers and utilities, named (6 rows): Hitachi Energy, ERMCO, Giga Energy and utility executives quoted on the record.
5. Analysts (1 rows): independent research such as Thunder Said Energy.
6. Trade press relaying the above (19 rows): POWER, Utility Dive, Reuters and pv magazine. We use them when the original is paywalled or blocks downloads, and the row names the original report.

Aggregator indexes and vendor blog claims sit at the bottom. We keep them when they add a newer date, mark them “weak source” in the table and draw them as dashed hollow marks.

Verbatim quotes

Each row stores the source’s own sentence in the quote field, copied exactly. Where the figure comes from a chart label rather than a sentence, the row says so (“chart value”). Click any mark on the chart, or the quote button in the table, to read it.

As-of dates

“As of” is the period the number describes, which is often months before it was published: Wood Mackenzie’s Q2 2025 survey appeared in September 2025. We store it at the precision the source gives (a year, a month or a quarter) and plot quarters at mid-quarter, years at mid-year and months at mid-month. Every source also carries the date we read it, and every row its verified date.

How we chart ranges

• Weeks only. The y-axis is weeks from order to delivery. A figure stated in months or years (“about 12 months”, “three to six years”) stays in those units and is shown as stated (months or years), as a small diamond on the rail under the chart.
• Ranges are bars. “120 to 130 weeks” is a vertical bar from 120 to 130; the class line passes through its midpoint.
• Open-ended figures are arrows. “More than 160 weeks” is a mark at 160 with an arrow up, and the line segment into it is dotted.
• Gaps are dotted. Where a line spans more than 18 months with no reading in between (large power from 2019 to Q4 2023), the segment is dotted: we know both ends, not the path.
• Lines join comparable readings. A class line connects readings that measure the same thing the same way, mostly Wood Mackenzie’s survey. Conflicting press reports, aggregator reads and vendor field notes are plotted as hollow marks off the line.
• Averages that mix all sizes (NIAC’s 120-week figure, NEMA’s 4 to 6 weeks) are hollow squares: they are real readings, but not of one class.
• ≠ marks a disagreement. A reading flagged ≠ has another source giving a different figure for the same class and period; both are on the conflicts list.
• Every mark is a link. Click, tap or focus a mark to read the source’s own sentence under the chart; open it for the full record.

Classes

We assign every row to one class by hand, from the type the source names. The chart and table use the same assignment.

• Large power (17 rows): Large power and substation transformers.
• GSU (4 rows): Generator step-up (GSU) transformers.
• Pad-mount 3-phase (4 rows): Three-phase pad-mount distribution transformers.
• Pad-mount 1-phase (2 rows): Single-phase pad-mount distribution transformers.
• Distribution, all (7 rows): Distribution transformers, all types including pole-mount.
• All sizes (6 rows): Surveys that average all transformer sizes or do not name a type.
• Prices (20 rows): Price indexes and price changes.

Wood Mackenzie classifies power transformers by type, not voltage, and reports GSUs separately. Some older sources give one figure for large power and GSU units together; those sit on the large-power line and the row says so.

The price index

The headline price change is our calculation from the BLS producer price index for Electric Power and Specialty Transformer Manufacturing, series PCU335311335311, not seasonally adjusted, read on FRED: (475.065 ÷ 255.5 − 1) × 100 = 85.9% from January 2020 to September 2026. Each January reading uses the same January 2020 base. We refresh it monthly when BLS publishes.

Known gaps

• Europe. No ENTSO-E, Eurelectric, T&D Europe or IEA source gives a week figure yet. We hold a 2024 “lead times have doubled” quote and a vendor’s 200-week claim, which we mark weak.
• Pole-mount on its own. Utilities and labs report distribution transformers as one group; no source we have gives a pole-mount lead time.
• 2022 and 2023 for large units. No primary source gives a large power transformer week figure for those years; the Q4 2023 reading is the first.
• Reports read through the press. The NIAC report, the DOE large power transformer report to Congress and some Wood Mackenzie notes block automated downloads, so those figures come via Utility Dive, POWER and Reuters. The row names the original either way.
• Manufacturer backlogs. Order and backlog data from Hitachi Energy, Siemens Energy and GE Vernova earnings calls is not in the dataset yet.

Have a dated figure that fills one of these? Send it with the source to desk@transformerleadtimes.com.

Corrections and sponsors

Send corrections with a public source and we update the row and its verified date. Sponsors buy a labelled slot on the page. Readings, classes and chart positions come from the sources alone.

Reuse

The dataset is open under CC BY 4.0: CSV, JSON, manufacturers CSV. Credit Transformer Lead Times and link transformerleadtimes.com. Field definitions are on the data page.