Load Growth Changed Who Buys Power

For roughly two decades, US electricity demand was flat. Efficiency offset growth, utilities planned for replacement rather than expansion, and the energy transition was largely a story about changing what generated power, not how much was needed.

That assumption broke. Demand is growing again — driven by data centre buildout, electrification and returning industrial load — and the system was not planned for it.

The buyer changed

The consequential shift is not the megawatts. It is who is procuring them.

Historically utilities planned supply and regulators approved it. Now a substantial share of new demand comes from corporate buyers with their own deadlines, their own capital, and no patience for a planning cycle measured in decades.

A company that needs power by a specific date and cannot get it through the normal process will do something else: build on site, sign a long-term contract with a specific generator, restart an asset, put storage behind the meter, or site the facility somewhere with available capacity and accept the inconvenience.

That is a genuinely new class of customer — well-capitalised, time-constrained, and willing to pay a premium for certainty rather than the lowest price.

What that customer actually values

Speed to power. The dominant variable. A more expensive solution available in eighteen months frequently beats a cheaper one available in five years.

Firmness. Intermittent supply does not serve a load that runs continuously. This is why firm generation, storage, and hybrid arrangements have re-entered conversations that had moved on.

Not needing permission. Anything that reduces dependence on a queue or a rate case has value that is hard to see in a levelised cost comparison.

Note that none of these is "cheapest per megawatt-hour." The industry's default metric is a poor guide to what this buyer will pay for.

Where that creates openings

Behind-the-meter generation and storage — avoiding the queue rather than waiting in it.

Grid-enhancing technologies — getting more out of existing transmission is faster than building new transmission, and the payback is measured against a constraint that is now urgent.

Flexible load — the ability to curtail on demand turns a data centre from a grid problem into a grid asset, and can shorten the interconnection conversation considerably.

Brownfield sites with existing interconnection — retired industrial and generation sites where the hard permission already exists. Some of the most valuable assets in this market are old.

The risk in the thesis

Two things worth holding in mind.

Load forecasts are requests, not commitments. A large share of announced demand represents the same project shopping several regions. Capacity planned against speculative announcements can strand.

This buyer is price-insensitive until it is not. Willingness to pay a premium for speed is a function of a build race. If the race slows, procurement reverts to cost.

A venture underwritten on today's urgency should be able to survive its normalisation.

Questions worth asking

Does this sell speed, firmness, or cheapness — and which one is the customer actually buying?

Is the customer contracted, or forecast?

Does the model require a grid connection nobody has secured?

If demand growth moderates, what remains?

The read

Energy ventures spent twenty years optimising for cost in a market that was not growing. The market is growing now, and the scarce thing is not cheap electrons. It is electrons available on a date, reliably, without waiting for permission.

That is a different product, sold to a different buyer, and most of the sector is still priced for the old one.