Bottleneck record, bottleneck_assessment_v1
Power transformers
Long-lead grid equipment with a concentrated supplier base. Pressure from the copper input price and grid capacity growth; transformer lead times, imports and electrical steel have no registered source. Response is the grid investment capex pulse, with signals from transformer-related companies listed.
Reading
In plain words
Pressure is 86 on a 0 to 100 scale (the weighted 86th percentile of its components' own histories) with 2 of 5 components available, 50 percent of the weight, as of 2026-07. Status: constrained, response underway. The response signal is capital expenditure growth of +35.2 percent year over year across the tagged companies, above the 10 percent threshold. Adding capacity takes 2 to 5 years by the cited source, so today's response is read against that horizon.
rules fired: pressure 86.13 >= 75; capex pulse 0.35 > 0.1
Is the problem being solved?
Pressure and response
- pressure0 to 100, left
- capex growth (fraction)fraction yoy, right
86
0 to 100, coverage 50%, 2 of 5 components, as of 2026-07
35.2%
capex growth yoy, threshold 10%, as of 2025-12-31
Pressure components
Percentiles, weights and what is missing
| Component | Weight | Series and transform | Percentile | Raw | Period | History | Status |
|---|---|---|---|---|---|---|---|
Copper input price pressure 12-month percent change of the global copper price (FRED PCOPPUSDM), the winding conductor | 0.3 | copper-price WLD pct_change_12 | 85.48 | 38.6082 | 2026-07-01 | 403 | present obs 156367, 156368, 156369 ... |
Grid capacity growth 12-month percent change of US net summer capacity (EIA ELGBPUS): every added generator passes through a transformer | 0.2 | us-electric-net-summer-capacity USA pct_change_12 | 87.09 | 4.7032 | 2026-05-01 | 244 | present obs 241165, 241166, 241167 ... |
Transformer lead time Delivery lead time for large power and distribution transformers. | 0.3 | no registered series | -- | -- | -- | -- | missing no transformer lead-time series is registered candidates: DOE Large Power Transformer Resilience Report (2024); NIAC transformer report (2024); producer surveys |
Transformer imports US imports of transformers (HS 8504) as a read on domestic capacity. | 0.1 | no registered series | -- | -- | -- | -- | missing no trade series is registered candidates: US Census USA Trade Online, HS 8504 |
Electrical steel supply Grain-oriented electrical steel output or price. | 0.1 | no registered series | -- | -- | -- | -- | missing no electrical steel series is registered candidates: AISI; USGS iron and steel |
Response detail
What the response signal rests on
Pulse grid_investment, row 2, pulses.v1, 9 of 10 members contribute.
| Company | Latest capex | Prior | Period end |
|---|---|---|---|
| Duke Energy CORP | 14.024B | 12.28B | 2025-12-31 |
| SOUTHERN CO | 12.737B | 8.955B | 2025-12-31 |
| American Electric Power Company, Inc. | 3.453B | 0.399B | 2025-12-31 |
| Vistra Corp. | 2.752B | 2.078B | 2025-12-31 |
| CONSTELLATION ENERGY CORPORATION | 2.949B | 2.565B | 2025-12-31 |
| EATON CORPORATION plc | 0.919B | 0.808B | 2025-12-31 |
| GE Vernova Inc. | 1.277B | 0.883B | 2025-12-31 |
| HUBBELL INC | 0.1551B | 0.1804B | 2025-12-31 |
| QUANTA SERVICES, INC. | 0.6092B | 0.6041B | 2025-12-31 |
No extracted company signals stored for the tagged companies.
Downstream
Systems that depend on it
| Entity | Type | Hops | Centrality |
|---|---|---|---|
| Transmission and distribution grid | infrastructure_system | 1 | 6.7 |
| Electricity system | infrastructure_system | 2 | 4.6 |
| Data center | facility | 2 | 1.75 |
| Semiconductor fabrication plant | facility | 3 | 3.75 |
| AI compute | technology | 3 | 0.75 |
| Economic output | entity | 3 | 0.5 |
| Human health and longevity | entity | 3 | 0 |
Substitutes
Named on the dependency edges
No substitute is named on any stored edge.
Recycling and recovery: Transformer copper and steel are recovered at end of life; no series is registered.
Expansion lead time
24 to 60 months
Large power transformer, order to delivery
Delivery of a large power transformer from order, at current order books; a new transformer plant takes longer.
U.S. Department of Energy, Large Power Transformer Resilience Report to Congress (2024)
“Lead times for large power transformers can reach as much as 60 months; the NIAC report of June 2024 puts current lead times at roughly two to four years.”
checked 2026-09-07
Evidence
Every number, by id
- metricCopper input price pressure: copper-price (WLD)pct_change_12, rank 85.48 over 403 periods, period 2026-07-01, observation ids 156364, 156365, 156366, 156367, 156368, 156369 ...
- metricGrid capacity growth: us-electric-net-summer-capacity (USA)pct_change_12, rank 87.09 over 244 periods, period 2026-05-01, observation ids 241162, 241163, 241164, 241165, 241166, 241167 ...
- pulsecorporate_pulses row 2 (grid_investment)9 contributing companies, pulses.v1
- entitygraph entity Power transformer
- edgeTransmission and distribution grid depends on Power transformer (requires_input, established_identity_or_engineering_relation, weight 1)
- documentU.S. Department of Energy, Large Power Transformer Resilience Report to Congress (2024)lead time 24 to 60 months: Lead times for large power transformers can reach as much as 60 months; the NIAC report of June 2024 puts current lead times at roughly two to four years.
Bottleneck and Response monitor, model bottleneck_assessment_v1, thresholds bottleneck_thresholds_v1
A record shows two signals and never combines them.
Pressure (modeled): each available component is transformed (12-month percent change, year-over-year growth with the sign reversed for supply, level for shares and lead times), then replaced by its percentile rank (0 to 100) within its own stored history up to that period, ties counted half, after a minimum history. The score is the weighted mean of those ranks over the components present, renormalised. Coverage is the share of declared weight that has data; components without a registered source are declared with their weight, their role and candidate sources, and count against coverage. Weights are a modelling choice and are listed on every record.
Response: the capacity response ratio where a capacity pipeline exists (US electricity: the stage-weighted 24-month EIA-860M pipeline over the trailing 24 months of additions, modeled), otherwise the capital expenditure pulse of the tagged company group (matched-sample trailing-four-quarter capex growth from stored 10-K and 10-Q facts, derived), otherwise null with the reason.
Status comes only from these thresholds:
- insufficient data: coverage below 0.5, or no pressure score. No label is guessed.
- below watch: pressure below 60.
- watch: pressure at or above 60 and below 75.
- constrained: pressure at or above 75.
- constrained, response underway: constrained and the response signal clears its threshold (capacity response at or above 1.0, or capex pulse above 10 percent).
Lead time is recommended-class metadata from a small sourced table (src/data/lead-times.ts) and is stored with its citation. Substitutes and downstream systems come from the curated knowledge graph (edges where the record's entity is the dependency; entities within three hops that depend on it, with centrality_v1). Every number on a record traces to stored observation ids, pulse row ids or edge ids in the row's JSON.
History
Assessments stored
| As of | Pressure | Coverage | Response | Status | Computed |
|---|---|---|---|---|---|
| 2026-07-01 | 86 | 50% | 0.35 | constrained, response underway | 2026-09-07 |