The stages of an energy transition that have to come first are the ones the national record does not count

An energy transition is a sequencing problem before it is anything else. Wires have to exist before the load they carry. Plant has to exist before the product. A workforce has to exist before the plant it staffs, and a price at which the new thing beats the old one has to exist before anyone commits capital to either. The order is fixed by how long each stage takes to build, and no appropriation reorders it. The national statistical record does not run in that order. It runs in the order of what is easiest to count, which is close to the reverse.

Where this is written from

I write as an analyst inside The Cefa Brief, working from published statistics and public documents and from nothing else. The figures below are annual series redistributed by FRED and retrieved for this site on 19 September 2026, together with the publishing agencies' own account of what each series counts. I hold no position in a utility, a ministry or a project. That vantage sees one thing well, which is what the statistical agencies have decided to count, at the frequency they publish it, on a national basis. It sees very little else. It cannot observe an interconnection queue, a permit, a power purchase agreement, a substation, a transformer order or a single megawatt of anything. Whether a project reached financial close, whether a line was energised, whether an engineer was available in the county where a plant was going up: none of that reaches a national annual series, and I will not read it out of one. What follows is an argument about the order of stages, built from the traces the record does hold and explicit about the ones it does not.

Lead time fixes the order

The ordering is a matter of construction time, not of preference. This publication carries no series of lead times, so the ranking that follows is a judgement rather than a measurement. A vehicle is bought in a single transaction. A factory takes years. A transmission line, a trained workforce and a permitting authority take longer than a factory, and in the case of the workforce the interval is a career. Whatever a government decides to fund first, the stages complete in the order their lead times allow. A programme that funds the fast stages and assumes the slow ones will keep pace has not changed the sequence. It has changed where the queue forms.

Both terms of the first calculation moved

Before any physical stage, a transition is an arithmetic comparison: the cost of building the new thing, discounted across its life, against the cost of continuing to run the old one. The discount rate is the first term, and it did not hold still. The ten-year Treasury constant maturity rate ran 0.89 percent in 2020, 1.45 in 2021, 2.95 in 2022, 3.96 in 2023, 4.21 in 2024 and 4.29 in 2025. That is a rise of 3.4 points, and 4.29 divided by 0.89 is 4.8, so the long rate ends the window at 4.8 times where it began. Generation, transmission and storage are capital paid up front and recovered across decades, which is the asset shape most exposed to that term. A project appraised in 2020 and refinanced in 2025 is not the same project.

The second term is the price of the thing being replaced, and it held still even less. West Texas Intermediate crude averaged 39 dollars a barrel in 2020, 68 in 2021, 95 in 2022, 78 in 2023, 77 in 2024 and 65 in 2025. The high stands 56 dollars above the low, and 95 divided by 39 is 2.4, so the top of the window is 2.4 times its bottom. A substitution that clears at 95 dollars need not clear at 65. Neither term is set by an energy programme, and neither repeated a single reading across the six years: an appraisal made in any one of them rested on a discount rate and a fuel price that had both moved again by the next.

Plant arrives before the things plant depends on

Construction is the fastest of the physical stages, and it is the one that moved. Spending on manufacturing plant ran 75 billion dollars in 2020, 82 in 2021, 125 in 2022, 202 in 2023, 244 in 2024 and 215 in 2025, every figure at a seasonally adjusted annual rate. Each point is already a yearly rate and approximates the year itself;none of them is a month's outlay. The rise from 2020 to 2024 is 169 billion, and 244 divided by 75 is 3.3. The final year gives back 29 billion. Manufacturing capacity in use ran 72.9 percent in 2020, 77.3 in 2021, 77.6 in 2022, 76.6 in 2023, 75.4 in 2024 and 75.3 in 2025. It peaks in 2022 and falls 2.3 points from there to 2025, and it still ends the window 2.4 points above where it started. The two series rise together to 2022 and diverge over the next two years: between 2022 and 2024 plant spending rose 119 billion while capacity in use fell 2.2 points. In 2025 both fall. Neither series is specific to energy, and that is part of the point: the record can show building running ahead of use for two years, and cannot say for which industry.

Counted, but not in the unit the question needs

Two series here sit close to a transition's demand side and its input costs, and neither is stated in the unit the question requires. Total vehicle sales ran 14.9 million units in 2020, 15.4 in 2021, 14.2 in 2022, 16 in 2023, 16.3 in 2024 and 16.6 in 2025, each at a seasonally adjusted annual rate and therefore a yearly rate of sales rather than a month's. The low of the window is 2022 at 14.2 million and the 2025 reading stands 2.4 million above it. What the series does not carry is a split by powertrain. It counts how many vehicles were sold and is silent on which kind, which is the only part of it a transition turns on. Producer prices in chemical manufacturing say the same thing from the cost side. The index ran 289 in 2020, 322.8 in 2021, 359.6 in 2022, 356.5 in 2023, 352.2 in 2024 and 356.5 in 2025, on a December 1984 base of 100. The 2025 reading is identical to the 2023 reading and sits 3.1 points below the 2022 peak; it is 67.5 points above 2020, which is 23.4 percent. It is an index for chemical manufacturing as a whole and isolates no individual material.

The record runs in the reverse of the working order

Set the stages in the order the work requires, then ask which of them this record can see. Generation capacity, transmission, interconnection, storage and the permitting that gates all four come first, and not one of them is recorded by any series carried on this site. There is no megawatt in the file, no mile of line, no queue position, no permit and no connection date. The stages that do appear are the later ones: the price of money, the price of the incumbent fuel, the buildings, the vehicles, the input price index. Completeness of measurement runs close to the inverse of urgency of stage. A reader who judges a transition by the numbers that are published will be judging its last stages and calling the result a verdict on the whole.

What this does not establish should be said exactly. None of the six series named here is specific to energy. Manufacturing plant spending covers factories of every description, capacity in use is a manufacturing measure, and vehicle sales count every vehicle. No figure above is adjusted for prices. Six annual observations cannot date a turn inside a year, cannot separate one cause from another, and cannot span the interval between a stage being funded and a stage being finished, which for the slowest stages is longer than the window itself. The claim made here is about ordering rather than about outcome, and the absence of the early stages from the record is not a reason to withhold it. That absence is the argument's principal piece of evidence.