On Thursday, October 22, Louisiana’s Public Service Commission voted 4-1 to adopt its “Louisiana Advanced Nuclear Exploration” program, requiring the state’s three investor-owned utilities — Entergy, Cleco Power, and Southwestern Electric Power Company — to study advanced nuclear options for their systems. The program is small by design: utilities may spend up to $2 million over its three-year life, and the framework as adopted does not raise customer rates for the work. Commissioner Davante Lewis, the lone dissenter, argued that ratepayer money should not fund research for private companies without a guarantee the public benefits.
Read it as a financing story first and a nuclear story second. The exploration program is the cheap part. Studying reactors costs millions; building them costs billions, and the question the commission’s vote leaves untouched is now the central question in American nuclear power: who pays for the construction. Entergy’s chairman and chief executive, Drew Marsh, has answered that question about as plainly as a utility chief executive ever does. New nuclear, he has said, requires help from “either the federal government or big data center customers with their trillion-dollar balance sheets.” Without one of those two backers, Entergy is not building.
That position, stated on the record in May and repeated on earnings calls since, reframes what the nuclear revival actually is. The wave of nuclear announcements over the past two years is not, for the most part, a wave of utilities building reactors for their ratepayers. It is a wave of the largest electricity buyers in the country underwriting the existing fleet — and, increasingly, being asked to underwrite the next one. If Entergy will not put a reactor in rate base on its own balance sheet, then new-build economics in the Gulf South are being priced by the load, not by the grid.
The $2 million program and the billion-dollar question
Start with what the commission actually approved, because the gap between it and a reactor is the story. Under the program, the three utilities will partner with nuclear technology firms on feasibility studies, economic analyses, and other early-stage work on integrating nuclear energy into their systems. Electric cooperatives may opt in. The commission had originally explored funding the work through a 20-cent monthly charge on every meter; at Thursday’s meeting it dropped that surcharge and instead allowed utilities to recover the costs within base rates, so long as rates do not increase. Any future rate increase tied to the work would require a separate commission application.
Commission Chair Eric Skrmetta said he expects that by the end of the three-year period the utilities will have identified the right technology for their systems and a way to move forward. Entergy Louisiana spokesman Brandon Scardigli said the company was glad to see efforts “that ensure our region is in a good position to take advantage” of new nuclear opportunities, and stressed that the framework “ensures that customer rates will not increase as a result of those efforts.”
Note the careful choreography. Nobody is promising a reactor. The program buys three years of studies for $2 million — a rounding error against a single reactor — with an explicit promise that customers will not see higher bills for the privilege. That is politically necessary in a state where industrial customers and consumer advocates are already fighting over who bears the cost of the data-center build-out. It is also, from an investor’s perspective, an admission: the exploration is affordable precisely because the construction is not yet anyone’s commitment.
Marsh’s line, on the record
Marsh’s position is not new, which is what makes it credible. Speaking in Little Rock on May 4, he said new nuclear would be part of Arkansas’s future — a company study projects regional electricity sales tripling or quadrupling by 2050, which he said cannot be served by natural gas and solar alone — but that the cost of a plant would be daunting. The overruns alone on the Vogtle reactors in Georgia, he said, are bigger than Entergy Arkansas itself. Construction, he said, “would require help from either the federal government or big data center customers with their trillion-dollar balance sheets,” adding that the latter “would need to see lower costs.”
On earnings calls, he has been blunter about the mechanism. “The key issue for us is our ability to manage the construction risk,” he said on one call. “And of course, we need a customer to want to pay for that.” On another, the company said it will not take on existential risk around new large-scale nuclear. The through-line is consistent: Entergy wants the nuclear option, it will not bet the company on it, and the party that wants the power must carry the risk.
That inverts the traditional utility model. In the classic regulated structure, the utility builds, the commission approves, the cost enters rate base, and customers pay over decades while shareholders earn an allowed return. Marsh is saying that model does not work for new nuclear at Entergy’s scale, and that the alternative is a customer-funded model. The load, in other words, becomes the rate base.
Louisiana’s load is arriving faster than its nuclear answer
Entergy’s data-center pipeline stands at 7 to 12 gigawatts, with another 3 to 5 gigawatts of other industrial load behind it, Marsh has said on earnings calls. The company has raised its four-year capital plan to $43 billion, concentrated on natural-gas generation and transmission to serve that demand. The flagship is Meta’s Hyperion campus in Richland Parish, backed by a $12.9 billion Entergy proposal — more than 5 gigawatts of new gas-fired generation, battery storage, and transmission — that regulators are expected to consider on December 16.
That proposal is the near-term answer to Louisiana’s load growth: the power being built for the hyperscalers right now is gas, not nuclear, because gas can be permitted and built on a timeline the developers will accept. Nuclear is the longer game. The state released a nuclear strategic framework in March, and Meta has separately solicited proposals for 1 to 4 gigawatts of new nuclear generation that could begin supplying power in the early 2030s. The exploration program fits inside that timeline — a three-year study program for a technology whose first Louisiana electrons, on the most optimistic reading, are a decade away.
For investors, the sequencing matters. Entergy’s $43 billion plan does not need a nuclear breakthrough to pencil out — it is a gas and wires plan, backed by electric service agreements the company only counts once signed, with early-termination penalties and minimum bills designed to keep other customers from carrying stranded costs. Nuclear is an option the company is keeping alive, not a commitment it is funding.
The template the hyperscalers already proved
Here is where the story connects to the series this site has been building. Over the past year, the hyperscalers have demonstrated, deal by deal, that they will underwrite nuclear power — but so far only the existing fleet, not new construction.
On October 6, Google and Constellation announced a 20-year power purchase agreement for 890 megawatts of new uprate capacity across 11 existing reactors, plus a 15-year agreement for 2,700 megawatts of existing output, with Constellation committing more than $4.3 billion of fleet investment. On September 30, Constellation announced a 20-year PPA with Amazon covering 690 megawatts, including 190 megawatts of new capacity at Calvert Cliffs in Maryland, backing more than $3 billion of Maryland infrastructure investment. Microsoft’s 20-year agreement underpins the restart of the 835-megawatt Crane Clean Energy Center — the former Three Mile Island Unit 1 — in Pennsylvania. And on October 2, We Energies announced that Oracle would subscribe to 10 to 20 percent of the output of Wisconsin’s Point Beach plant, 125 to 250 megawatts, for its Port Washington data-center campus, an arrangement that still requires Wisconsin commission approval.
Count the pattern. Every one of these deals attaches a creditworthy buyer to megawatts inside plants that already exist, already hold licenses, and already connect to the grid. The buyer’s contract de-risks the capital spending the way a rate order de-risks a utility’s — but without a regulator, and without asking other customers to pay. That is the template. What Marsh is now saying, in effect, is that the template has to stretch further: from uprates and restarts to actual new reactors. And the stretch is not trivial. An uprate is outage work on a licensed plant; a restart is a licensed plant returning to service. A new reactor is a construction program, a licensing program, and a decade of execution risk — the Vogtle experience, not the uprate experience.
The honest question is whether the hyperscalers’ appetite extends that far. Their revealed preference so far is for speed: uprates deliver in 2028, the Crane restart is targeted for the late 2020s, and Point Beach power is already generated. A new reactor delivers in the 2030s at the earliest, at a cost no American project has yet demonstrated it can control. Meta’s solicitation for 1 to 4 gigawatts of new nuclear is the first real test of whether a hyperscaler will sign something that looks like a new-build commitment rather than a fleet deal. Until one does, Marsh’s position is a statement about the market, not a negotiating posture.
What Entergy actually holds
Entergy is not starting from nothing. The company operates four large nuclear plants across its Gulf South territory, and it is already pursuing the cheap megawatts: it has sought approval for a roughly $68.7 million upgrade at Waterford 3 in Louisiana that would add about 45 megawatts, mostly during a scheduled refueling outage. That is the uprate playbook, applied inside a regulated fleet.
The longer option sits in Mississippi. Entergy holds an NRC early site permit for a potential new unit at its Grand Gulf station, expiring in April 2027; the company applied in September 2025 to renew it for another 20 years, and the NRC accepted the application for review in January 2026. An early site permit is not a construction license — it resolves site safety and environmental questions in advance — but it is the kind of asset that takes years to obtain and that a developer preserves precisely so the option stays alive.
There is also the federal alternative Marsh named. The Vistra deal covered in this series — a conditional Department of Energy loan commitment of up to $4.2 billion for uprates and life extension at three plants — shows Washington willing to act as the cheap senior lender for the existing fleet. Whether federal credit extends to new construction at Entergy’s scale is a different question, and one the current administration’s nuclear push has not yet answered with a closed financing.
The risks worth naming
The investor risks here are specific. Construction risk is the first and largest: the American new-build record since Vogtle is a record of delay and overruns, and Marsh’s own comparison — overruns bigger than Entergy Arkansas — is a reminder of the scale mismatch between a reactor and a mid-sized utility balance sheet. Customer-concentration risk follows: if new nuclear depends on one or two hyperscalers, the project’s fate rides on their demand thesis and their willingness to sign decades-long commitments for power they cannot yet see. Regulatory risk cuts both ways: Louisiana’s commission has blessed exploration without rate impact, but it has not been asked to bless construction cost recovery, and the December 16 decision on the $12.9 billion Meta build-out will show how far the commission will go in assigning data-center costs.
Timing risk is the quiet one. The Grand Gulf early site permit expires in April 2027; if the renewal stalls, one of Entergy’s few tangible new-nuclear assets lapses. The exploration program runs three years, so its findings land around the time Meta’s solicited nuclear proposals are supposed to start delivering power in the early 2030s — a timeline that leaves little slack for a technology never built at commercial scale in Louisiana.
What to watch
Four milestones will tell investors whether the who-pays question gets an answer. First, the December 16 commission decision on the Meta build-out: it will set the precedent for how Louisiana assigns the costs of hyperscaler load, and that precedent will shape any future nuclear cost-recovery fight. Second, the Grand Gulf early site permit renewal — a quiet administrative decision that determines whether Entergy keeps a shovel-ready site. Third, Meta’s 1-to-4-gigawatt nuclear solicitation: the first hyperscaler new-build commitment, if it comes, would be the market-clearing event Marsh is waiting for. Fourth, the exploration program’s findings over the next three years, and whether any utility emerges from them with a technology choice and a customer attached.
None of this is investment advice, and nothing in this analysis recommends buying or selling any security. What the Louisiana vote and Marsh’s position together demonstrate is a structural shift in how new nuclear gets financed, if it gets financed at all. The old model asked ratepayers to fund construction through rates. The emerging model asks the customer who wants the power to fund it through a contract — and so far, that customer has funded the fleet it already has, not the reactors that do not yet exist. Until a hyperscaler or the federal government signs up to carry construction risk for a new plant, Entergy’s nuclear future will remain what the commission just authorized: a well-funded study of a question nobody has yet paid to answer.