Export, or not

Selling power sounds like upside. For most industrial sites it is a second business, with its own regulator, its own risks, and compensation that has moved against sellers for a decade. The fork deserves a decision, not a default.

Section 01The fork, precisely stated

A non-export project generates only for the load behind its own meter. Protection settings and, where required, certified control schemes physically prevent power from flowing to the grid; the grid sees a smaller customer, not a supplier. An export-capable project can push surplus power onto the system, which makes it, in the grid's eyes and the regulator's, a source of supply with everything that status entails.

The fork is not about virtue or ambition. It is about which body of rules governs you, how much interconnection engineering you buy, and whose price you depend on. California's Rule 21 interconnection framework treats export behavior as a first-order property of the application: non-export and limited-export configurations pass through screens differently than full-export projects, because they present different risks to the circuit.¹

Section 02What each side of the fork buys you

Section 03The compensation lesson California already taught

The clearest evidence that export revenue is policy, not property, is the state's own record. California's net energy metering framework, which for years credited exports at rates close to retail, was succeeded by a net billing structure that values exports on a different and generally lower basis tied to avoided costs.³ Existing customers were treated under transition rules; new customers face the new math. Whatever one thinks of the policy, the planning lesson is neutral and permanent: export compensation regimes are revised by regulators on regulatory timelines, and a project whose economics require a particular export price has bound its returns to a docket it does not control.

Self-supply value has revision risk too; tariffs change for buyers as well as sellers. But the avoided-purchase side of the ledger is anchored by a stubborn fact: you will keep needing the energy. The export side has no such anchor. Demand for your surplus at yesterday's price is nobody's obligation.

Avoided purchases are anchored by your own demand. Export revenue is anchored by a docket. Build on the anchor you hold.

Section 04The honest cases for export

Neutrality cuts both ways, and there are sites where export logic genuinely applies:

• On-site fuel that does not stop. Facilities whose processes produce continuous fuel, digesters at water and food operations being the classic case, sometimes generate most economically at a steady rate that exceeds low-load hours. Export capability, or storage, is how that surplus avoids being flared away.
• Structured offtake, not merchant hope. Where a creditworthy counterparty contracts for the output on defined terms, export is a business decision with a signature on it, entirely different from building against an open-ended tariff credit. Federal qualifying-facility rules provide one long-standing structural lane for such arrangements.²
• Deliberate oversizing with eyes open. Occasionally the economics of a larger installation, priced honestly, beat two stages of a smaller one, and limited export is the pressure valve. Legitimate, provided the export value is carried at a defensibly conservative number and the case survives that number going to zero.

What does not belong on this list is the most common real-world motive: a vendor's model that needs export credits to rescue an oversized machine. Sizing should fail gracefully into "buy the smaller unit," not escape into an export assumption.

Section 05The two lines every model must carry either way

Whichever side of the fork a project takes, two tariff realities follow it and are routinely omitted from first models. The first is standby service: a customer relying on its own generation typically still wants the grid behind it, and that backup is a priced service under filed schedules, not a courtesy.⁴ The second is the family of nonbypassable and departing-load charges that can continue to apply to load you now serve yourself. Both belong in the paths comparison from the first draft, because both reduce the spread between grid and self-supply that headline models advertise. A study that shows you the generation price without the bill that remains has shown you half a bill.

Sizing to the meter: the minimum-load discipline, worked

The non-export choice carries its own sizing rule, and it is worth working once with illustrative numbers. Picture a facility whose daytime load runs near four megawatts and whose overnight minimum settles near two and a half; every figure here is illustrative, and your interval data supplies the real ones. A non-export configuration must never push power past the meter, so continuous generation is bounded not by the impressive daytime number but by the quiet overnight one, less an engineering margin for the nights that dip lower and the maintenance weeks when a line shuts down. The honest continuous rating for self-supply at that site is therefore near two megawatts, and the remaining daytime demand stays with the grid or with equipment that can follow load downward.

Notice what this rule does to a seller's instincts. The larger machine carries the better unit price and the bigger invoice, and the only way to justify it at this site is either an export assumption, which is the previous section's cautionary tale, or a load-growth assumption, which deserves its own evidence rather than a hopeful line in someone's model. The minimum-load discipline is thus one of the sharpest neutrality tests in the field: an analyst who has actually studied your interval data will talk about your overnight valleys within the first ten minutes; an analyst who talks only about your peaks has studied your bill's headline and their own price list.

The configuration details matter too, and they are decisions, not defaults. A certified power-control scheme that dynamically limits output to site consumption can hold generation closer to the load line than a simple relay-enforced cap, at the cost of controls scope; reverse-power protection settings, metering arrangements, and the interconnection application's stated export status all have to tell the same story the economics tell. Owners do not need to engineer these themselves. They need the study to state which scheme the numbers assume, because a model built on one scheme and a project built on another will meet, unhappily, at commissioning.

Section 06The decision test

Run one test before the fork is chosen: delete every dollar of export revenue from the model and see what remains. If the project still clears your bar serving only your own load, export capability is an option to price, not a premise. If deleting export kills the project, you are not evaluating an energy project for your facility; you are evaluating entry into the power-sales business, and that decision deserves its own memo, its own counsel, and its own appetite for regulatory risk. Most industrial owners, seeing it framed that way, choose their own meter. That is not timidity. It is knowing which business you are in.