The nuclear tariff is the new power contract

Google has agreed to buy 96 megawatts of nuclear power the way a corporate account adds lines to a phone plan, and the structure Georgia Power filed to sell it may matter more than the megawatts. PWD's tracking shows Georgia Power submitted a subscription tariff with the Georgia Public Service Commission for the output of a nuclear uprate; if approved, it gives Southern Co. a copyable way to sell uprate megawatts to large loads and hands every utility with an uprate candidate a filing to imitate.

The same week's renewable transactions did not print a price. Navaris, a five-day-old fund, announced 304.5 megawatts of German onshore wind; Zelestra broke ground on 38.5 megawatts of German solar and disclosed none of the economics; ACEN units dropped 456 megawatts of Philippine solar with no counterparty, transfer price, or stated reason; Endra entered Finland with no capacity, offtaker, or connection named. Across those announcements the common field was the blank where the price should have been, and Georgia's filing is different because it puts the price discovery inside a regulatory docket.

Power scarcity is moving into the regulated rate base, and the instrument for pricing scarce megawatts is becoming the utility tariff rather than the merchant power purchase agreement. A tariff is not a deal announcement; it is a product with a rate schedule, a service territory, and a commission that must approve or reject it. By converting a nuclear uprate into a subscription, Georgia Power is selling a regulatory claim on the energy, and the claim is copyable in a way a negotiated PPA is not.

The subscription structure separates the uprate from the bespoke negotiation, and that separation matters more than the 96 megawatts. A tariff on a nuclear uprate puts the sale into the rate base and the terms into the public record, which lets Southern Co. take the same form to a paper mill, a factory, or another data center in its service territory. Every other utility with an uprate candidate can take Southern's filing and adapt it, and what they get is a structure that works without renegotiating the world.

The week's other policy moves point the same way. New York's battery bottleneck has moved from the cost of the asset to the interconnection contract that grants permission to connect it. Texas cut the cost of holding an ERCOT queue position and added a two-year energization grace period while the state is still counting how many of its 474 gigawatts of queued projects are real. The queue, the interconnection contract, and the tariff are all the same instrument in different forums: a regulated device for allocating the grid's scarce permissions. Merchant developers who can only print megawatts are competing for those permissions with less than utilities have to sell.

For merchant transactions, the blank price column is an underwriting statement. When a fund buys 304.5 megawatts of German onshore wind with no price announced, the seller transfers merchant risk to the buyer, who must satisfy lenders without a public revenue figure. The same blank runs through Zelestra's 38.5-megawatt German solar and ACEN's 456-megawatt Philippine solar, which name capacity without a counterparty, a transfer price, or a reason. These projects may be sound, but they are unpriced, and in a capital market where a data center can sign a public tariff, unpriced renewable capacity gets marked down.

Endra's Finnish entry takes that pattern to its end: the announcement names no capacity, no offtaker, and no connection, which makes the project a capital-raising instrument before it is an infrastructure fact. A renewable developer that cannot name a buyer is asking capital to price energy policy, construction risk, and power-market spread without the numbers an underwriter would need. A utility tariff names the load, the megawatts, and the regulatory forum, which is why it is becoming the instrument for large loads while unpriced project announcements migrate to smaller merchant deals.

The Georgia filing does not, by itself, mark the end of the PPA: large loads will still sign them, and merchant developers will still sell project rights without a price column, because the alternative is not always a utility with an uprate to sell. Where an uprate or a utility asset exists, though, the tariff gives the utility a standard form the PPA market lacks—negotiated case by case, with terms that leak only if someone chooses to disclose them. The tariff is filed, heard, and approved, and every subsequent customer gets the benefit of the form. That difference in transaction cost will matter more as data center demand pushes utilities to harvest every megawatt they already own.

The next move belongs to the other utilities with uprate candidates, now that Georgia Power has given them the filing; the first utility to copy it will confirm that the nuclear tariff is a product class rather than a single transaction, and the second will make it a market.