The Price of Consent

The idea

This inquiry began with a headline. On October 6, 2026, the Financial Times published Martha Muir's report "NextEra's $67bn Dominion takeover faces political backlash in Virginia". Within two hours I had put a question to Vista's Desk: when a buyer pays a state's customers for the state's consent, who really pays, and can the growth the buyer promises its own shareholders survive the bill? The Desk challenged the idea over three rounds before this inquiry was written. Everything that follows is built from the primary record: the merger agreement and the joint proxy statement, the Virginia commission's docket, the companies' filings and market prices, and, for the history, the statutes, court opinions, letters and archives listed at the end.

A proposal at a nuclear conference

On November 4, 2025, at a gathering of the people who run America's nuclear power plants, one lawyer handed another a letter.

The first was John Ketchum, the chief executive of NextEra Energy. He had begun his career as a tax lawyer in Kansas City, where he took his law degrees, the second of them in taxation, and then practiced in Tampa, at a law firm and as corporate counsel to the parent of Tampa's electric utility, before NextEra hired him in 2002. He rose through its lawyers and its finance staff, became chief financial officer in 2016 and chief executive in 2022. There was a fitness in that. NextEra had become the largest generator of power from the wind and the sun on earth, in a country where that business runs on federal tax credits, and it was led by a tax lawyer.

The second was Robert Blue, the chief executive of Dominion Energy, and he was as Virginian as a utility executive can be. He studied at the University of Virginia and then at Yale Law School, practiced law at Hogan & Hartson, and served Governor Mark Warner as counselor and director of policy before he joined Dominion in 2005. Three years later he finished an MBA at Virginia's Darden School. He ran Dominion's Virginia utility before he ran the whole company, which he has led since 2020 and chaired since 2021.

The letter was a proposal. NextEra wanted to buy Dominion, all in stock, at a premium of about 20%, and it offered $600 million in cash to Dominion's shareholders to soften the blow to their dividends.

Blue had not asked to be bought. In December his board decided not to look for other buyers because, in the words of the filing that later told the story, Dominion was not being offered for sale. In January it asked NextEra for a meaningful increase in the premium. The courtship had begun.

It lasted six months. The boards signed on Friday, May 15, 2026; Bloomberg reported the deal on Sunday; and on Monday, May 18, before the stock market opened, the companies announced it, a deal worth about $67 billion. Each Dominion share would become 0.8138 of a NextEra share, plus a small slice of a $360 million cash payment. The combined company would be the largest utility in the United States, with Florida Power & Light at one end, Dominion Energy Virginia at the other, and, in Dominion's territory, the largest concentration of data centers in the world. Shareholders of both companies have since voted yes. Only the regulators remain.

A utility cannot simply be sold. Its new owner must win the consent of the people who regulate it: here the State Corporation Commission of Virginia, the commissions of South Carolina and North Carolina, the Federal Energy Regulatory Commission and the Nuclear Regulatory Commission. Virginia matters most, because most of Dominion's customers live there, and Virginia has made the most noise. Its governor has intervened in the case, which no Virginia governor had done before. Its lieutenant governor has come out against the deal. Its attorney general's office wants the clock started over.

What Virginia wants is more for its customers. To win approval, the companies have offered to put credits on Virginia bills, about $10 a month for a typical household, paid for by NextEra's shareholders, along with money for bill assistance and job training and promises about jobs and headquarters. In September they sweetened the offer. When I say that Virginia will get more, I mean that the commission's final order will very likely require more of this kind of thing than the companies have offered so far: larger credits, or tighter conditions on how the Virginia utility is run and how much cash it may send to its new parent.

Since the deal was announced, NextEra's shares have fallen 18%, from $93 to $76 at the close on October 5. Utilities as a group fell 9%. Had NextEra kept pace with them, each share would be worth almost $9 more, about $18 billion across the company. It is natural to read that fall as the price of Virginia's displeasure.

I read the numbers the other way. Everything the companies have offered customers so far comes to about $2.5 billion of shareholder money, about 1.6% of what NextEra is worth. Virginia will get more, and NextEra's own merger agreement, which makes walking away cost $4.83 billion, all but guarantees that NextEra will pay it. But even another $1.5 billion would cost NextEra's holders less than a dollar a share. What has cost them more is the fall in the share price itself. NextEra plans to sell $20 billion to $28 billion of new stock between 2027 and 2032. At $76 instead of $93, raising the same dollars takes about 64 million more shares, and the 9% a year that NextEra promises in earnings per share becomes, at today's price, about 8.7%.

That is the story in brief. The rest of this inquiry tells it properly: where these two companies came from, how a 1940 Virginia law gives the state its say, how two economists' models predict the price of that say, and what it all means for anyone who owns either stock, and for anyone who wonders who will pay for the electricity that artificial intelligence needs.

Edison's grandchildren

Both companies took their modern shape in 1925, and both were possessions of distant holding companies.

Florida Power & Light was created on December 28, 1925, in the fever of the Florida land boom, by American Power & Light, a holding company in the Electric Bond and Share system that General Electric had organized twenty years before. Its first assets were a jumble of small-town utilities and their odd belongings, among them an ice cream factory, a limestone quarry, a sponge-fishing boat and 35 mules. The man who ran Electric Bond and Share, Sidney Mitchell, had been hired by Thomas Edison in 1885.

Edison himself wintered in Fort Myers, on Florida's Gulf coast. A bridge there was named for him and dedicated on February 11, 1931, his eighty-fourth birthday; he died that October. No lamps had ever been planned for it, and for years it stood dark, until in 1935 Ripley's Believe It or Not made fun of the unlit Edison Bridge. Florida Power & Light gave $10,000 for 54 lamps, and at midnight on November 25, 1937, a telegraph signal from Edison's son Charles switched them on.

Virginia's utility took its form in 1909, when Frank Jay Gould, son of the railroad speculator Jay Gould, merged three Richmond streetcar lines into the Virginia Railway and Power Company. Richmond's streetcars had their own pedigree. In 1888 Frank Sprague, who had worked for Edison, built there the first successful large electric street railway in the world, and Gould's lines descended from it. Gould's Richmond-to-Ashland cars, it was said, could reach 90 miles an hour; Gould himself went on to the French Riviera, where he developed Juan-les-Pins and died in 1956. In 1925 a syndicate led by the Boston engineers Stone & Webster bought the company, set up Engineers Public Service Company to hold it, and renamed it the Virginia Electric and Power Company, VEPCO for short. Dominion traces its ancestry further back still, to a 1787 act of the Virginia General Assembly that named trustees to clear the Appomattox River for boats "capable of carrying six hogsheads of tobacco." The company histories say George Washington and James Madison were among the trustees. The statute names seven other men.

So both utilities spent their youth inside the pyramids of the holding-company age, built by Edison's men and owned by distant financiers. The Supreme Court would later describe some of those holding companies as mere sets of books in Bond and Share's office, and it worked an example of their leverage: "An investment of $10,000,000 thus controls $729,000,000, a ratio of 1 to 73."

Insull's throne

The age had an emblem. His name was Samuel Insull.

He came from London in 1881, a 21-year-old clerk, to be Thomas Edison's private secretary, and he never forgot his first sight of the great man: shabby and careless in his dress, and magnetic, with very bright eyes. When General Electric was formed in 1892 and Insull was passed over for its presidency, he went to Chicago to run a small electric company, and he made it the model for an industry. By 1930 the utilities he controlled produced a tenth of America's electricity, in some 5,000 communities in 32 states. He paid for them partly by selling stocks and bonds to his own customers and employees, which raised money and, he hoped, would quiet the movement for public ownership. Others had done it first. Insull made it famous.

His wife, Gladys Wallis, had been an actress before their marriage, and in 1925 she returned to the stage in The School for Scandal. A young drama critic at the New York Times, Herman Mankiewicz, went to review her opening, came back drunk, and, as his biographer tells it, passed out over his typewriter. Fifteen years later Mankiewicz wrote Citizen Kane, and he gave the scene to Jed Leland, the critic who passes out over his review of the opera debut of Kane's second wife.

On November 4, 1929, six days after the crash, and ninety-six years to the day before John Ketchum handed Robert Blue his letter, Insull opened the Chicago Civic Opera House, a 45-story tower with two wings, shaped like a throne and facing the river. Chicagoans called it Insull's Throne. (The legend that he built it with its back to New York, in a pique, is only a legend.) Two and a half years later it all fell. His investment companies went into receivership, and some 600,000 shareholders and 500,000 bondholders lost their savings.

Insull fled to Paris and then to Greece. On March 16, 1934, the American minister in Athens cabled Washington: "The Foreign Minister informed me this morning that Insull left Piraeus Wednesday at 4 p.m. for Port Said in small Greek freighter named Maiotis." For two weeks the freighter wandered the eastern Mediterranean. When it turned north through the Dardanelles, Turkey agreed to arrest him, held him in Istanbul, and handed him to the Americans, who brought him home on the Exilona. He was tried three times and acquitted three times; at the first trial he counted the selling of his securities to his own customers among his honest errors. On July 16, 1938, he collapsed in the Place de la Concorde station of the Paris Métro as he stepped toward the ticket taker, and the police identified him by a hotel laundry bill. He was seventy-eight. The newspapers could not agree on what was in his pocket: 84 cents by the Associated Press's count, 20 cents by Time's. He was not quite the pauper of the legend. He was living on a pension of $21,000 a year from the companies he had built.

Congress answered him with the Public Utility Holding Company Act of 1935. Its section 11 limited each holding company to "a single integrated public-utility system," with others allowed only in the same or adjoining states. The Supreme Court upheld the breakup of American Power & Light in 1946, and in 1950 Florida Power & Light became independent; overnight, NextEra's own history says, it acquired about 14,000 stockholders. The Securities and Exchange Commission's orders of 1941 and 1942 left Engineers Public Service with VEPCO alone, and in 1947, when Engineers was liquidated, VEPCO, with 450,000 gas and electric customers, became independent too.

Under that act a Florida holding company would have had to explain why a system in Virginia belonged with one in Florida, two states that do not touch. Congress repealed the act in 2005, and the repeal took effect in February 2006. The law that set both companies free is the law whose absence now lets them marry.

The courtship

The two companies grew up very differently. In 1984 Florida Power & Light acquired a parent, FPL Group, which renamed itself NextEra Energy in 2010, and its wind and solar business made it the world's largest generator of power from the wind and the sun; in October 2020 its market value briefly passed ExxonMobil's, $145 billion to $142 billion, as Bloomberg reported. It has also been the most persistent suitor in American utilities. By my count it has pursued nine of them since 2000 and married one, Gulf Power, in 2019.

Dominion took the other road. It bought gas companies, a South Carolina utility and a share of the country's pipelines, then sold most of the gas business, cancelled the Atlantic Coast Pipeline twenty days after winning its case 7 to 2 in the Supreme Court, and put its future into Virginia, where the data centers came. In 2025 they bought 28% of the electricity Dominion's Virginia utility sold. By July 2026 Dominion counted over 53 gigawatts of data-center demand at some stage of contracting.

The joint proxy statement the companies filed in July tells how the letter of November became the deal of May, and it reads like a courtship conducted by lawyers, which it was.

Blue would not be hurried. When NextEra wanted to announce in March, he called it premature. He asked Ketchum whether it was the exchange ratio that was fixed or the premium, the question of a lawyer who knows that share prices move. In April he demanded a make-whole dividend of at least $360 million for his shareholders, eighteen months of protection for his employees' jobs, and a premium of at least 22% whatever happened to the share prices. And he had a second suitor. In March another company, which the filing calls only Party A, approached Dominion and said it was prepared to sign a merger by the end of the month, at a premium materially below NextEra's. Blue dined with Party A's chief executive in April, told Party A in May that it needed a premium in line with a market control premium, and met its chief executive again in Richmond on May 14, the day before Dominion signed with NextEra. Party A never closed the gap. It offered nothing for customers.

Meanwhile Ketchum courted. In February, on a call about the regulators, NextEra's team walked Dominion's through its own past deals and why some of them had not closed; it is not often that a suitor reviews old broken engagements with the next family. On April 30 he dined with Blue and both companies' lead independent directors. One of them, Dominion's Susan Story, had once run Gulf Power, the one utility NextEra had managed to buy. On May 5 he had what the filing calls an introductory lunch with Dominion's board.

Two moves in that negotiation matter for everything that follows. On March 16 NextEra replaced its $600 million payment to Dominion's shareholders with $1.35 billion of bill credits for customers, and by April 1 the credits had grown to $2.25 billion. The money had changed address, from Dominion's shareholders to Virginia's customers, before Virginia saw a page: the companies had begun bargaining over Virginia's consent between themselves. And on March 26 Dominion asked for a regulatory termination fee equal to 7% of its equity value, to be paid by NextEra if regulators killed the deal. NextEra agreed. That fee is $4.83 billion. Keep it in mind.

Why Virginia gets a say

Virginia's suspicion of monopolies is older than the republic. Section 4 of George Mason's Virginia Declaration of Rights, adopted on June 12, 1776, reads: "That no Man, or Set of Men are entitled to exclusive or separate Emoluments or Privileges from the Community, but in Consideration of public Services." That is as good a definition of a utility franchise as any written since. On December 20, 1787, Thomas Jefferson, then the American minister in Paris, wrote to James Madison about the new Constitution and listed what he missed in it, among them a "restriction against monopolies." The next summer he put it more sharply: "The benefit even of limited monopolies is too doubtful to be opposed to that of their general suppression." Madison wrote back that "Monopolies are sacrifices of the many to the few."

Virginia built its guardian against them a century later. The State Corporation Commission was created by the Virginia Constitution of 1902 and opened in March 1903, to regulate the railroads, the telephone and the telegraph. A. Caperton Braxton, who led the fight for it at the constitutional convention, called railroad rates the greatest and most important economic question before the civilized world, and promised to protect consumers from the rapacity of the common carriers. The commission was given the powers of a court as well as of a regulator, and to this day, it says, no other state has placed so broad an array of regulatory duties in one agency. In 1908, ruling on one of its first rate orders, Justice Oliver Wendell Holmes wrote: "The establishment of a rate is the making of a rule for the future, and therefore is an act legislative, not judicial, in kind." There is an irony at the root of it. The constitution that created Virginia's guardian of consent was never put to the voters, and the same convention wrote the poll tax that shut most Black Virginians out of the vote.

The law that governs this sale is younger, from 1940. It asks one question. The commission must approve when it is "satisfied that adequate service to the public at just and reasonable rates will not be impaired or jeopardized." If the commission does not act in time, the application is "deemed approved." The penalty for buying a Virginia utility without permission is a fine of "not more than $1,000," a figure that has not changed since 1940.

Notice what the law does not say. It does not require a buyer to give customers anything; it asks only that they not be harmed. Mark Christie, a former Virginia commissioner who went on to chair the federal regulator, has called the statute obviously receptive to mergers. So why offer $2.25 billion? Because the commission may attach conditions to its approval, and has: in 2000 it approved Dominion's own purchase of Consolidated Natural Gas only on condition that Virginia Natural Gas be sold or spun off. Because the Carolinas apply tougher tests. And because the politics are hot.

The credits are not a tax. A tax is set by law and paid to the government. These are offered by the buyer and paid to the customers, as a separate line on the bill, about a cent for every kilowatt-hour, scaled to how much each customer uses. They bind only if the commission writes them into its order, and the companies have promised that no part of them will come back to customers later. For a typical household, $10 a month is about 6% of a bill of some $169; over four years it would come to $480. The other money has purposes too. EnergyShare, Dominion's bill-assistance program, began in 1982 with the United Way and the Salvation Army; NextEra would add $100 million to it through 2038. A $100 million workforce fund would train Virginians for utility jobs. The promised Virginia Supplier Program, up to $1 billion a year for five years, is a purchasing commitment and no gift, and its costs could still reach customers if the commission finds them prudent.

Politics, Florida and Washington

Dominion has been the largest corporate political donor in Virginia for decades; in 2018 it gave to 121 of the 140 members of the General Assembly. In 2015 a law froze its base rates and suspended the commission's reviews of its earnings, and the commission's staff later found that from 2017 to 2020 Dominion had collected $1.143 billion more than a fair return would have allowed.

Its most determined opponent is a hedge fund manager in Charlottesville. Michael Bills runs Bluestem Asset Management, about $1.5 billion, in the home town of the University of Virginia, his alma mater, where he once helped manage the endowment and has taught finance at the McIntire School of Commerce. In 1992 he helped found the Sorensen Institute for Political Leadership. In 2018 he founded Clean Virginia to back candidates who refuse utility money, and by the 2022 and 2023 cycle each side was spending more than $10 million.

Electricity bills and data centers were central to the 2025 election. Abigail Spanberger, a former CIA officer and member of Congress, won the governorship with almost 58% of the vote, and that November the commission raised the typical bill by $11.24 a month for 2026. This August the new governor wrote in The Washington Post of being deeply skeptical that selling the state's primary utility to an out-of-state company is good for the commonwealth, and on August 17 entered the case, the first Virginia governor ever to intervene before the commission. The filing warned that the deal, as proposed, "risks saddling all businesses and consumers ... with unaffordable rates." Robert Blue once wrote policy for a Virginia governor. Now a Virginia governor is a party to his case, and a skeptical one.

The Speaker of the House of Delegates, Don Scott, wrote on September 1: "If this merger is not focused on affordability, it should not go forward." The lieutenant governor, Ghazala Hashmi, sent the commission 64 questions in July and on October 6 came out against the deal. The attorney general's office, arguing that the September offer changed everything, told the commission: "The Joint Petitioners have put a different deal on the table." And the chair of the three commissioners who will decide, Kelsey Bagot, once worked for NextEra, disclosed it, and has not stepped aside.

Florida has sent its baggage north. On October 2 the commission's chief hearing examiner ordered the companies to produce the so-called Robo memo, a 2021 memorandum reportedly sent to NextEra's chief executive at the time, Jim Robo, about Florida Power & Light officers' alleged use of political consultants. Behind it lies Florida's 2020 ghost-candidate affair, in which a spoiler who shared the incumbent's surname drew 6,382 votes in a state senate race decided by 32. A former state senator was convicted; Florida Power & Light has denied any role and says its own review found no wrongdoing. A federal appeals court, in a shareholder suit over the matter, wrote, "The complaint has it all," while stressing that it was ruling on allegations. The examiner wants the memo to test the companies' claim of historic managerial fitness.

Was anything Trumpian going on? Not about the merger itself: I found no statement on it from the president, the White House, the Energy Department or the Justice Department. But the two companies stand on opposite sides of Washington's energy war. NextEra gave $1 million to the president's 2025 inaugural committee, and in March it announced that the president had approved up to 10 gigawatts of its gas-fired plants under the investment deal with Japan. Dominion's offshore wind farm, 176 turbines going up in the Atlantic off Virginia Beach, was stopped by the Interior Department in December 2025 "for reasons of national security," and restarted only when a federal judge enjoined the order in January; it sent its first power to the grid in March. In the 2026 State of the Union the president said: "We're telling the major tech companies that they have the obligation to provide for their own power needs." On the other side, fourteen members of Congress wrote to the federal regulator on September 29: "Our fundamental concern is that this new entity will be both a gigantic power generator and a gigantic power-providing utility."

What NextEra signed

The merger agreement obliges NextEra to use its reasonable best efforts to win approval and to take "any and all steps necessary," divestitures included. It may refuse only a "Burdensome Condition," and the test is narrow: a condition that would have a material adverse effect on a company the size of Dominion, which is far smaller than the company NextEra would become. The $2.25 billion of credits is written out of that test altogether. If a regulator blocks the deal, or NextEra walks away from conditions it calls burdensome, NextEra pays Dominion $4.83 billion. That is more than $2 for every NextEra share.

Twice before, regulators have refused NextEra a utility on its terms, and twice NextEra let the utility go. In 2016 Hawaii's commission rejected its bid for Hawaiian Electric, unpersuaded that customers would benefit; NextEra's rate credits had never grown past $60 million, and it paid Hawaiian Electric $95 million to part ways. In 2017 Texas rejected its bid for Oncor after NextEra refused the ring fence the commission required, independent directors with a veto over dividends; its chief executive at the time, Jim Robo, called the conditions deal killers. That walk cost NextEra no fee at all. Sempra then signed for Oncor, accepted the same ring fence, and won approval in under seven months.

Walking away cost NextEra $95 million in Hawaii and nothing in Texas. In Virginia it would cost fifty-one times the Hawaii bill. That changes the bargain.

A beautiful mind and a split

Vista's Decision Library carries two models for this situation, and I have not seen either applied to this deal.

The first is John Nash's, of A Beautiful Mind fame. The idea came to him at nineteen, as an undergraduate at Carnegie Tech, in an elective course in international economics taught by Bert Hoselitz. Harold Kuhn, chairing Nash's Nobel seminar in 1994, said he had written it before he knew that game theory existed. The paper, "The Bargaining Problem," appeared in Econometrica in April 1950. Kuhn liked to joke that the things being bargained over in its example, a bat, a ball, a toy and a knife, proved that a teenager had written it, and that neither colleagues nor the journal's editor could talk Nash out of them. Nash shared the economics prize in 1994, and in 2001 Russell Crowe played him on the screen. In May 2015 he and his wife Alicia were killed in a taxi on the New Jersey Turnpike, on the way home from Norway, where he had just received the Abel Prize.

I had the good fortune to study Nash's model in 1992, in fellow Nobel laureate Roger Myerson's game theory class at Northwestern, which is why I was so excited to realize that it can be used in this inquiry.

The Library's name for his model says what it does: the outside option decides the split. Each side's share of a bargain depends on what it gets if the talks fail. Here is this bargain, written out.

Let G be what the deal is worth to NextEra's holders if it closes, before any concessions. Let F be the fee NextEra pays if regulators kill it: $4.83 billion. Let C be what Virginia's customers get as the price of consent. If the deal closes, NextEra's holders keep G minus C. If it fails, they pay F. So agreeing is worth G plus F minus C more to them than failing. Virginia's customers get C if the deal closes and nothing new if it fails, so agreeing is worth C to them.

Nash's solution chooses C to make the product of those two gains as large as it can be:

(G + F − C) × C

The product is a kind of fairness. It rewards neither side for squeezing the other to nothing. If C is close to zero, Virginia's factor is close to zero and so is the product; if C is so large that it uses up everything NextEra gains, NextEra's factor is zero and so, again, is the product. The largest product lies exactly halfway: C = (G + F) / 2.

Try it with numbers. Suppose the deal is worth nothing extra to NextEra's holders, so G is zero, and the fee is $4.83 billion. Then:

The peak is at half the fee: $2.4 billion. The companies have offered $2.5 billion. The package on the table, in other words, is almost exactly what the bargain predicts if NextEra believes the deal adds nothing for its own holders, and NextEra's own adviser comes close to saying so. BofA Securities' discounted cash flow puts a combined share at about $82 to $116, against about $81 to $117 for NextEra alone; the midpoints differ by 25 cents.

NextEra's management says something else. It promises growth in earnings per share of 9% or more a year to 2032 if the deal closes, against 8% or more without it, which by 2032 is about 6.7% more earnings for every NextEra share, worth some $10.6 billion at today's price. Take management at its word, and G is $10.6 billion. Then G plus F is $15.4 billion, the product peaks at C = $7.7 billion, three times what is on the table, and a buyer who believes its own forecast becomes the regulator's best bargaining chip.

This reminds me of the judgment of Solomon. Two women came before the king, each claiming the same living child, and the king called for a sword: "Cut the living child in two and give half to one and half to the other." Then the woman whose son it was "yearned with compassion for her son" and cried out, "Please, my lord, give her the living baby. Do not kill him!" And the king ruled: "Give the living baby to the first woman. By no means should you kill him; she is his mother." (1 Kings 3:25 to 27, Berean Standard Bible.) Solomon's split was never meant to be carried out. It was a test. It worked because what each woman would give up revealed what she truly valued. Nash's split works the same way here. What NextEra finally concedes to Virginia will tell us what its management truly believes the deal is worth.

The doomsday machine

The second model is Thomas Schelling's. On a sabbatical in London in 1958 he wrote chapters of what became The Strategy of Conflict, and his manuscript was typed on Charing Cross Road by Agatha Christie's typist. He published the book in 1960; committees he served on started the hotline between Washington and Moscow; and in 2005 he shared the economics prize. The Library calls his idea: a threat is only as good as its cost to withdraw. His examples were homely and frightening at once: the general who burns the bridges behind his own army; two trucks loaded with dynamite, meeting on a road wide enough for one.

His thinking reached Stanley Kubrick. In 1960 Schelling praised Red Alert, a thriller about an accidental nuclear war, and Kubrick and its author, Peter George, came to spend an afternoon with him. They could not make an accidental war plausible in the missile age without a madman in the Air Force, so they turned the thriller into a nightmare comedy, and Dr. Strangelove has the best line ever written about commitment. When the Soviet ambassador reveals the doomsday machine too late, Strangelove cries that its whole point "is lost if you keep it a secret!" (Even history's most famous commitment was less dramatic than its legend. Hernán Cortés did not burn his ships in 1519; by his own letter, he ran them aground.)

Written out, Schelling's test is simple. NextEra's unspoken threat is that it will walk away if Virginia asks too much. It will rationally walk only if what Virginia asks is worth more than what walking costs it:

walk away only if C > G + F

In Hawaii, F was $95 million, so walking away beat almost any concession, and NextEra walked. In Virginia, F is $4.83 billion. With G at zero, NextEra should accept anything up to $4.83 billion before it walks; with G at $10.6 billion, anything up to $15.4 billion. At these numbers the threat to walk is not believable, and the people deciding can read the contract. The contract is an exhibit to a public filing.

Robert Blue understood Schelling better than anyone in this story. When Dominion's board demanded the 7% fee in March, it tied NextEra's hands, and when the agreement was filed in May, the knot was there for Virginia to see.

The September package fits that reading. When the residential credit was stretched from two years to four, most of the added money, $488 million, was moved from the credits meant for data centers. NextEra's new money for the credits was $88 million. With $100 million each for bill assistance and a workforce fund, the September changes added $288 million. NextEra has bargained as if its threat were still believable. The numbers say it is not.

Who can wait

A third thinker completes the picture. In 1982, at 30, Ariel Rubinstein showed in Econometrica that when two sides trade offers back and forth, the more patient one takes more, and a weak bargainer can be left with almost nothing; four years later he and two colleagues showed that as the delays between offers shrink, his answer becomes Nash's. Rubinstein, who keeps an Atlas of Cafés Where One Can Think, some 750 of them, has said he is proud never to have pretended his theory was useful advice: the 1982 paper describes bargaining and does not teach it. I use it here in that spirit, as a way of seeing.

Who can wait here? Virginia's law deems the application approved if the commission fails to act by January 11, 2027, which looks like a gift to the companies. It is less of one than it looks. The attorney general's office has asked the commission to treat the September package as a new filing, which would move the deadline to about March 13, and the General Assembly convenes on January 13 with a study of a longer review and a broader standard already under way; laws from that session normally take effect on July 1. NextEra's contract runs to November 15, 2027, extendable to August 15, 2028, and its promise assumes the deal closes in the second half of 2027. Virginia can wait longer than NextEra would like.

Packages grow

The record of past mergers points the same way. In eight contested utility deals where both figures are public, the final customer package was a median 2.8 times the first, and in five of them it at least doubled. Of the three exceptions, two were rejected. One of those was NextEra's bid for Hawaiian Electric, where the package never grew.

So I put the odds at about two in three that Virginia's final terms will cost NextEra's holders $3.5 billion to $5 billion before tax, against $2.5 billion now, either as more money or as conditions NextEra fought in Texas, such as limits on the dividends Dominion Energy Virginia may pay its parent. On October 6, when State Senator Scott Surovell asked about safeguards on dividends, NextEra's vice chairman, Armando Pimentel, told Virginia legislators that NextEra had opposed such limits in Texas but that Dominion's situation differs, as The Center Square reported. I expect NextEra to accept.

What it costs a share

Here the numbers turn gentle. The credits are excluded from the adjusted earnings NextEra promises to grow, so their cost shows up in cash, not in the promise. The cash has to come from somewhere. Fitch, one of the rating agencies, has listed significantly higher than expected customer refunds among the things that could cost NextEra its rating if they are not funded in a way that supports its credit, and NextEra says it will keep its ratings. So assume the worst for shareholders: every extra dollar Virginia wins is raised by selling stock at today's price.

• Another $0.5 billion is about 7 million shares and three hundredths of a point of growth a year.
• Another $1.5 billion is about 20 million shares and nine hundredths of a point.
• Another $2.5 billion is about 33 million shares and sixteen hundredths of a point.
• Another $5 billion, twice what is on the table, is about 66 million shares and three tenths of a point.

Three tenths of a point is what the share price has already taken. The equity plan is written in dollars: about $4 billion a year, $20 billion to $28 billion from 2027 to 2032, plus equity units NextEra sold in 2024 and 2026 that turn into shares on fixed formulas. At May's price, NextEra would have about 3,146 million shares by 2032, counting those issued for Dominion; at today's price, about 3,210 million. The difference, 64 million shares, takes 2% off earnings per share in 2032, and 9% a year becomes 8.7%. A lower price means more shares; more shares mean slower growth per share; slower growth is a reason for a lower price. George Soros called this kind of loop reflexivity. Here it is weak, about a tenth of a point of growth for every 6% fall in the price, and it runs in both directions.

Who pays for AI's electricity

Strip away the history and the politics, and this case is about one question: who pays for the power that artificial intelligence needs?

Northern Virginia is where the answer is being written. Loudoun County alone counted some 233 data center buildings this March, windowless and humming, about 56.5 million square feet in all. Data centers bought 28% of the power Dominion's Virginia utility sold last year, and Dominion counts over 53 gigawatts of data-center demand at some stage of contracting. NextEra's board, explaining its vote, cited a forecast that American demand for electricity will grow about 60% by 2045, and a pipeline of large customers wanting more than 130 gigawatts. Ketchum told his employees that demand is rising faster than at any time since the years just after the Second World War. On the day of the announcement, Bloomberg's Liam Denning wrote that this deal would define the AI power boom. Combined, the two companies would join the world's largest generator of wind and solar power to the utility at the center of the data-center boom: a generator and a grid in one, which is exactly what worries its critics in Congress.

This reminds me of a line of Victor Hugo's, from his history of the coup of 1851: one resists the invasion of armies; one does not resist the invasion of ideas. Artificial intelligence is that kind of invasion. No commission will stop it. None is trying. What the commissions can decide is who pays for its electricity, and every part of the Virginia fight is an argument about that bill. The September package took the credits meant for the largest data-center customers and gave them to households. Virginia enacted a statewide consumption tax on data centers this year, the first of its kind. The president has told the technology companies to provide their own power. And fourteen members of Congress worry that one company will be both a giant generator and a giant utility, in effect selling power to itself.

That is where the investment ideas spring from. The price of consent is a toll, and it will not be paid once. Every utility whose growth rests on data centers will pay it, in mergers and in rate cases, as long as households are paying more and the largest technology companies are buying the new power. Nash's arithmetic gives the next buyer a rule of thumb: expect regulators to ask for about half of what walking away would cost you, plus half of whatever gain you have promised your own shareholders. And the incidence is moving. In Virginia this year, costs moved away from households and toward data centers, through the credits, the tax and the rhetoric. Investors in the companies that sell power to data centers should expect the same pressure everywhere the data centers go.

A century ago, Insull's answer to the question of who pays was to make his customers his shareholders. The answer taking shape in Richmond runs the other way. NextEra's shareholders will pay Virginia's customers, a few dollars a month, for permission to grow.

What it means for an investor

Vista makes no buy or sell calls. The analysis still points in clear directions, and three numbers frame them. The price: for a 10% annual return, NextEra's combined earnings per share must grow about 7.9% a year to 2032 if the market still pays 18 times earnings then, or 8.8% at 17 times; today it pays about 17 times the Street's estimate for 2027. The promise: 9% or more a year, from $3.71 in 2025 to $6.78 or more in 2032. The Street: about 8.7% a year for the next five years. My estimate is 8.5% to 8.7%, the promise less the extra shares today's price requires and the extra concessions I expect Virginia to win. On that path the price offers about 10% a year, a little less at 17 times and a little more at 18.

For a NextEra holder, Virginia is the headline. It is not the risk. The whole price of Virginia's consent, even at twice what is on the table, comes to less than two and a half dollars a share, and the worst case, a regulator killing the deal, costs $2.32 a share in fees. What moves a NextEra share is the price at which it sells $20 billion to $28 billion of new stock between 2027 and 2032. If the shares recover, the loop runs in the holders' favor, with fewer new shares and faster growth per share. If they fall, it runs against them. Watch the pace and price of NextEra's equity issuance more closely than the hearing room in Richmond; the company will update its plan with its year-end results.

For a Dominion holder, the premium is gone. The exchange ratio is fixed, so a Dominion share now behaves like 0.8138 of a NextEra share. On May 15 that was worth about $76, a premium Dominion's board put at 23.1%; today it is worth about $62.50, and Dominion trades at $61. Had Dominion merely kept pace with utilities since May, it would trade near $56; add the fee NextEra would pay it if regulators blocked the deal, about five and a half dollars a share, and you are back near $62. A Dominion share is a NextEra share with insurance against a regulatory failure. It is not insurance against NextEra's own price, which Dominion holders now ride all the way down or up. That is also why Dominion's holders have little left to win or to lose in Virginia, and why the party that most needs Virginia's consent is NextEra.

For a Virginia household, the offer on the table is worth about $480 over four years, roughly 6% of a typical bill, and more if the commission gets its way.

The signposts, in order: staff and intervenor testimony on October 19; NextEra's answer to the federal regulator's deficiency letter, due October 23; the evidentiary hearing on November 17; South Carolina's hearing on December 8; the Virginia deadline of January 11, 2027, or about March 13 if the clock restarts; the General Assembly from January 13; South Carolina's final order by January 29; NextEra's year-end report and funding plan; and the contract's outside date, November 15, 2027.

The data that we're missing is...

• What the commission will add. How Virginia's commission has weighed bill credits against conditions, such as limits on dividends or a ring fence, in past transfers, and what it added after the companies' last offer. Its orders record the result, never the bargaining that produced it.
• What the deal is worth to NextEra's holders. The G in the formula. Management promises 9% or more a year; its own adviser's valuation finds almost no difference. Neither is a measurement, and the companies have not put a number on what combining saves or earns.
• Whether the fee moves the buyer. How often a regulatory termination fee has changed what a buyer conceded. Concessions made in private before an order appear in no database.
• How NextEra would pay for more. Whether extra concessions would be funded with debt or with new shares, and at what price. The year-end funding plan will tell part of it.
• How the rating agencies would treat bigger credits. S&P already expects about 14.5% of funds from operations to debt in 2027, below its 17% threshold. Whether more credits funded with debt would tip a rating is a judgment no filing shows.
• Who pays for the grid that data centers need. How much of Dominion's 53 gigawatts is firm, how the cost of the new lines and plants is split between data centers and households, and how the 2026 tariff and tax changes are moving that split. They are too new to show in any filing.

Primary research

Likely, we can get some of it from primary research, which is what Vista is all about. The people who sat on the other side of these bargains, at the commission, at the negotiating table, at the rating agencies and in the energy offices of the data centers, know already what the filings will take a year or more to show.

For a client engagement, Vista combines this report with that primary research.

The checkpoint, set down now

We will score this inquiry on April 30, 2027, and publish the result on the scorecard whether it flatters us or not. Each test sides with the reading here:

• Virginia gets more, and NextEra pays. By the commission's final order, Virginia's terms exceed the September package: at least $3 billion of shareholder-funded benefits in all, or a limit on Dominion Energy Virginia's dividends or a ring fence beyond the companies' own offer, and NextEra accepts them. Falsified if the commission approves the September package without additions, or NextEra ends the merger.
• The cost lands in shares, not in the ratings. Through its first report after the order, NextEra keeps its promise of 9% or more and its ratings, and no agency turns its outlook negative over the Virginia terms. Falsified if an agency downgrades NextEra or turns its outlook negative citing the Virginia terms, or if NextEra lowers its promise because of them.

Two passes and the reading here gains ground; one is too early to say; none, and the bargain did not work the way Nash and Schelling would have it.

Edison's bridge stood dark for six years because no one had planned to pay for its lamps. In the end a utility paid, two years after a newspaper cartoon made fun of the darkness. The data centers of Northern Virginia will not wait six years for their power, and someone will pay for it. The commission's order will say who, at least for this deal. Solomon never meant his sword to fall, and NextEra, I think, never meant to walk.

Questions this inquiry answers

Why is NextEra Energy buying Dominion Energy?

NextEra's board pointed to the size of the combined company, the largest utility in the United States, and to the growth of demand for electricity, above all from data centers: it cited a forecast that American demand will grow about 60% by 2045, and Dominion serves the largest concentration of data centers in the world, in Northern Virginia. NextEra promises growth in earnings per share of 9% or more a year to 2032 if the deal closes, against 8% or more without it. The deal, announced on May 18, 2026, is worth about $67 billion; each Dominion share becomes 0.8138 of a NextEra share plus a small cash payment.

Why does Virginia have to approve the NextEra and Dominion merger?

Under Virginia's Utility Transfers Act of 1940, control of a Virginia utility cannot pass to a new owner without the approval of the State Corporation Commission, which must be satisfied that "adequate service to the public at just and reasonable rates will not be impaired or jeopardized." The commission may attach conditions to its approval. If it fails to act in time, the application is deemed approved; the current deadline is January 11, 2027, or about March 13 if the clock restarts.

What are the bill credits in the NextEra and Dominion deal?

Credits on Virginia customers' bills, about a cent for every kilowatt-hour, or about $10 a month for a typical household, paid for by NextEra's shareholders and shown as a separate line on the bill. The companies offered $2.25 billion of them and added $88 million in September, along with $100 million for bill assistance and $100 million for a workforce fund. They are not a tax: the buyer offers them, and they bind only if the commission writes them into its order.

How much will Virginia's approval cost NextEra shareholders?

Everything offered so far comes to about $2.5 billion of shareholder money, about $1.22 a NextEra share before tax. This inquiry expects Virginia to win more, $3.5 billion to $5 billion in all, with odds of about two in three, but even twice what is on the table costs less than two and a half dollars a share. The larger cost has come from the fall in NextEra's share price, which means more new shares must be sold to raise the $20 billion to $28 billion of equity NextEra plans for 2027 to 2032.

What happens to Dominion shares if the merger closes, or if it fails?

If it closes, each Dominion share becomes 0.8138 of a NextEra share plus about 41 cents in cash, so a Dominion share now moves with NextEra's. If regulators block the deal, NextEra pays Dominion a regulatory termination fee of $4.83 billion, about $5.49 a Dominion share. On October 5, 2026 the deal valued a Dominion share at about $62.49, against a price of $61.02.

What is the Nash bargaining solution, and how does it apply to this merger?

John Nash's 1950 solution splits the gain from an agreement according to what each side would get if talks failed. Here NextEra's holders gain the deal's value plus the $4.83 billion fee they avoid paying, and Virginia's customers gain whatever concessions they win. Making the product of the two gains as large as possible gives Virginia half of the deal's value to NextEra's holders plus half the fee: about $2.4 billion if the deal adds nothing for NextEra's holders, and about $7.7 billion if it adds what management promises.