Two companies, one compact, and a merger nobody expected
NextEra is two businesses that could hardly be less alike, bolted together.
The first is Florida Power & Light — a regulated monopoly serving about six million customers. Nobody else may sell them electricity. In exchange for that protection, a state commission sets the prices, and sets them at a level calculated to give investors a specified return on the money they have sunk into poles, wires and power stations. That arrangement has a name: the regulatory compact. It is the oldest deal in American infrastructure.
The second is NextEra Energy Resources — the largest generator of wind and solar power in the world, selling electricity under long-term contracts, in competitive markets, with economics that depend heavily on federal tax law. It is not protected by anything.
And as of May this year there is a third thing to know, which is large enough to reframe everything else. NextEra is in the middle of the biggest utility merger in American history: it is buying Dominion Energy, in an all-stock deal valuing Dominion at around $67 billion, expected to close in the second half of 2027.
Almost every ratio in a standard screening toolkit was built for a business that sets its own prices and fights for its customers. Point those ratios at a regulated monopoly and they return numbers that are arithmetically correct and analytically worthless. Part II is about nothing else.
We have done this before — funds from operations rather than earnings for REITs, net asset value and net investment income for business development companies, net interest income and tangible book for banks. This is the utilities instalment, and it is the clearest one yet.
| ★ The compact, in one sentence | ★ NextEra raises capital, spends it on assets the regulator approves, those assets enter the "rate base", and the regulator then sets prices to return the depreciation plus an authorised return on that base. <b>Capital spending is not the cost of standing still — capital spending is the product.</b> |
| CEO | John Ketchum — chairman, president and chief executive |
| Adjusted earnings (2025) | $3.71 per share, up 8.2% · ★ reported earnings were $3.30 and FELL — see Part II for why the reported line is unusable here |
| ★ Guidance | ★ 2026 adjusted earnings of $3.92–$4.02, targeting the high end · and 8%+ annual growth through 2032, then targeting 8%+ again through 2035 — a <b>ten-year</b> commitment, and one the company RAISED rather than cut |
| The scale of the machine | ★ Roughly $24 billion of capital spending a year, funded by raising $23–25 billion of long-term debt annually. Long-term debt has gone from $72.4bn to $98.8bn in eighteen months. |
| Market capitalisation | ~$179.8B · dividend yield 2.76% · ⚠️ the all-time high was $97.17 on 30 April 2026 — this year, not years ago |
Four numbers that look alarming and are not
Our own data feed returns four figures for NextEra that, on any ordinary company, would end the conversation. Here is each one, and why it means nothing here.
| The number | Why it is a category error |
|---|---|
| ★ Altman-Z of 1.08 — nominally "distress" | ★ Edward Altman built this score in <b>1968</b>, on a sample of <b>66 American manufacturers</b>. Three of its five inputs are structurally hostile to a utility. It penalises thin working capital — but a utility collects monthly from millions of customers and funds hundred-year assets with long-dated debt, so thin working capital is efficiency. It reads leverage as a management gamble — but ★ <b>Florida's regulator explicitly sets the capital structure, at a 59.6% equity ratio.</b> The regulator chose the leverage. And it penalises low asset turnover — but a utility's entire purpose is to accumulate assets it is allowed to earn on. <b>The score asks whether a firm can survive a bad year of competition. A regulated monopoly does not face competition.</b> |
| ★ Free cash flow yield of MINUS 5.66% | ★ <b>Negative free cash flow is the machine working, not breaking.</b> Every dollar of approved capital spending becomes a decades-long regulated income stream. A utility with positive free cash flow has stopped growing its rate base — which means its earnings growth has stopped too. ⚠️ <b>But see the genuine limit below, because this is where the bull case usually stops and should not.</b> |
| ★ Return on invested capital of 3.98% | ★ A regulated utility does not <i>earn</i> a return; it is <i>awarded</i> one. The consolidated figure blends Florida's regulated returns, the competitive arm's contracted assets with their large minority interests, and billions of dollars of <b>half-built power stations that earn nothing yet by design</b>. The denominator includes a building site. ★★ The right numbers: <b>authorised return on equity of 10.95%, in a band of 9.95% to 11.95% — and Florida Power & Light actually earned about 11.70% last year and 11.40% the year before.</b> |
| Net debt of 6.06× earnings | Elevated for an industrial; ordinary for a utility mid-build. It also treats non-recourse project debt and tax-equity structures as if they were corporate obligations. ★ Use the rating agencies' own tests instead: S&P's funds-from-operations-to-debt threshold, Moody's cash-flow-to-debt threshold, Fitch's leverage trigger. <b>Those are the numbers that decide whether the machine keeps running</b> — and all three agencies affirmed NextEra in May and <i>loosened</i> those thresholds. |
This is the one that catches professionals. NextEra's competitive arm hedges the power it will sell in future years, and many of those hedges do not qualify for hedge accounting. So the change in market value of contracts covering electricity to be delivered years from now lands in this quarter's income statement — no cash moving, no electricity generated.
| Second quarter 2026 | $m | Per share |
|---|---|---|
| Reported net income | 3,144 | $1.50 |
| less: gains on non-qualifying hedges, after tax | (640) | ★ ($0.30) |
| less: gains on equities in the nuclear decommissioning funds | (134) | ($0.06) |
| plus: merger expenses and other | 37 | $0.01 |
| ★ Adjusted earnings | 2,407 | ★ $1.15 |
★ A year earlier the same line ran the other way, pushing reported earnings below adjusted. In twelve months it swung from suppressing earnings by about seven cents to inflating them by thirty — roughly a third of a $1.15 quarter, from a bookkeeping entry on electricity that has not been generated.
The consequence is stark. Between 2024 and 2025 NextEra's reported earnings per share fell, from $3.37 to $3.30. Its adjusted earnings rose 8.2%, from $3.43 to $3.71. One of those two numbers describes the business and the other describes the weather in the derivatives market.
⚠️ The legitimate objection, which an honest report must state: management chooses what to exclude, and management is paid on the adjusted number — NextEra says so in plain text, that adjusted earnings are used "as an input in determining performance-based compensation." "We adjust out the things that make us look bad" is a fair suspicion. So apply the test: does the adjustment cut both ways? Here it verifiably does — the same item hurt the adjusted figure in 2025 and helped it in 2026. The exclusions are symmetric and rules-based, not opportunistic. That is the standard we would want applied to any company making the same claim.
Everything above defends NextEra against a bad reading. Here is the reading that is not bad, and it is the real risk.
The "negative free cash flow is fine" defence rests on one load-bearing assumption: that the company can keep raising capital at a cost below the return the regulator allows it to earn. The whole thing is a spread business. Borrow at X, be permitted to earn Y, keep the difference. When X rises towards Y, the machine stops paying.
That is why utility shares behave like long-dated bonds, and it is not sentiment. Rising interest rates do three separate things at once: they raise the coupon on debt NextEra must issue every single year in enormous size; they raise the rate at which a very long stream of future cash flows is discounted; and — the subtle one — because authorised returns are reset only in rate cases held every few years, they widen the gap between the market cost of money and the permitted return on money, sometimes for years. There is a structural lag, and shareholders eat it.
★ To be fair to Florida's regulator, it has partly addressed this: the previous rate agreement contained a clause that automatically raised the authorised return from 10.60% to 10.80% in 2022 when Treasury yields rose. Regulators do not write that clause for utilities they dislike.
A capital-markets business with power stations attached
Buffett wrote the thesis for this in seventy words
Buffett has owned regulated utilities for two decades and has explained why better than anyone. From the 2011 Berkshire letter:
"Take care of your customer, and the regulator — your customer's representative — will take care of you. Good behavior by each party begets good behavior in return... We are expected to put up ever-increasing sums to satisfy the future needs of our customers. If we meanwhile operate reliably and efficiently, we know that we will obtain a fair return on these investments." — Warren Buffett, 2011
Florida Power & Light is the cleanest live illustration of that sentence in American utilities. Its bills run roughly 30% below the national average and its non-fuel operating costs about 71% below the industry, with top-decile reliability. And the payoff arrives exactly as Buffett describes it: an authorised return of 10.95%, and an earned return of about 11.70%.
The protection is unusually concrete, too: FPL holds 226 franchise agreements, typically running thirty years and out to 2055, under which municipalities contract not to set up their own utility.
Twelve years later, in the 2023 letter, he returned to the subject in a very different mood:
"The regulatory climate in a few states has raised the specter of zero profitability or even bankruptcy... the fixed-but-satisfactory-return pact has been broken in a few states, and investors are becoming apprehensive that such ruptures may spread... I did not anticipate or even consider the adverse developments in regulatory returns." — Warren Buffett, 2023
That is the whole risk in this security, stated by the man who owns more American utility assets than almost anybody. The compact is not a contract. It is a promise, and it is political. In California and Hawaii it broke, and he admits he did not see it coming.
| Does the compact hold at NextEra? | Evidence |
|---|---|
| ★ Yes — and strongly | ★ The utility is earning <b>above</b> its authorised return. A four-year settlement with a 10.95% authorised return, a regulator-set 59.6% equity ratio, a rate-stabilisation mechanism, a large-load tariff written specifically for data centres, and forward-looking adjustments for new solar and storage. Plus a clause in the previous agreement that automatically raised the authorised return when Treasury yields rose. <b>That is not a grudging regulator; it is a partner.</b> |
| ⚠️ But it is being litigated right now | ⚠️ Appeals against the 2025 rate agreement were <b>consolidated by the Florida Supreme Court in June 2026</b> and the docket is open. This is the most financially material legal matter the company faces, because it goes to the authorised return itself. |
| ⚠️ And the number is a political target | ⚠️ Florida's own public counsel has characterised the 10.95% as the highest authorised return in the contiguous United States. That is exactly the sort of figure that becomes an election issue when bills rise. |
| ★ Florida's version of the wildfire problem is hurricanes | ★ FPL collected roughly <b>$1.2 billion</b> of interim storm surcharges in 2025 for Hurricanes Debby, Helene and Milton — <b>subject to refund pending a prudence review</b>. A genuinely catastrophic season, or a disallowance, is the same shape of risk that damaged utilities in California, if smaller and better mechanised. |
Twenty years of nothing, and then a turn
Here is the fact that makes this whole sector interesting again, and it is not a forecast — it is history.
| US electricity demand, thousand terawatt-hours | 2005 | 2020 | 2025 | 2030 est. | 2045 est. |
|---|---|---|---|---|---|
| Total US demand | 3.9 | ★ 3.9 | 4.3 | 5.1 | 6.8 |
Read the first two columns. American electricity demand in 2020 was lower than in 2005. Fifteen years of economic growth and zero electricity growth — efficiency absorbed all of it. That era has ended.
NextEra's own framing is roughly 10% growth across the twenty years to 2025, against 58% across the twenty years after — what it calls a sixfold increase in the growth rate. And the composition matters more than the headline: on the forecasts NextEra cites, data centres account for about 43% of the growth to 2032, with transport electrification at 20%, commercial and industrial at 20%, and residential at 17%.
That 43% is important in both directions. It is large enough that an artificial-intelligence disappointment would hurt badly. It is also small enough that more than half the growth comes from something else entirely — which is the strongest single answer to anyone who thinks this is purely an AI bet.
Our reading: the demand inflection is real and unusually well evidenced — twenty years of flat load turning upward is not a forecasting artefact, it is already in the delivered data. What is not established is the slope. NextEra is underwriting a slope. The honest bear case is not "demand won't grow" — it is "demand grows at half the assumed rate, and the capital was raised for the full rate."
A cliff in 2030, and a squeeze that accidentally helps
The competitive arm's economics rest on federal tax credits, and those credits were rewritten in July 2025. The detail matters more than the headline, and most coverage got the direction wrong.
| What the law now says | Consequence |
|---|---|
| ★★ Wind and solar | ★★ Full credits only if construction began before 5 July 2026 or the facility is in service by end-2027. And then the hard stop, in the company's own words: <b>"There will be no clean energy tax credits for wind or solar facilities placed in service after 2030."</b> That is the cliff, and it is written into statute. |
| ★★ Storage — treated completely differently | ★★ The 30% investment credit runs for storage projects beginning construction through <b>31 December 2033</b>, phasing down in 2034–35. ★ <b>So wind and solar face a 2030 cliff while batteries have a runway to the mid-2030s. That is why storage was 2 of the 3.6 gigawatts NextEra originated last quarter. The mix shift is not fashion — it is tax law.</b> |
| ★ Nuclear | ★ Plants in service after 2024 — <b>explicitly including restarts of reactors previously being decommissioned</b> — qualify if construction begins by end-2033. That is the provision behind the Duane Arnold restart, which has a Google power-purchase agreement attached. <b>The nuclear revival is not sentiment; it is a tax-advantaged asset with a customer.</b> |
| ★★ And the squeeze that helps | ★★ In August 2025 the tax authorities <b>eliminated the 5% spending test</b> as a way to "begin construction". For twenty years a developer could start the clock by writing a cheque. Now it must put <b>steel in the ground</b>. ★ <b>That favours exactly one kind of developer: the very large one with a real supply chain and the balance sheet to break ground on dozens of sites at once. It is a regulatory moat handed to NextEra by a policy designed to hurt it</b> — and almost every write-up reported the legislation as unambiguously bad news. |
★ So how much of the backlog is protected? NextEra's own statement, repeated word for word in its most recent quarterly filing, is that it "believes that its current pipeline of wind and solar facilities to be placed in service through 2030 will qualify for clean energy tax credits."
Read that precisely, because it is more careful than it looks. The claim is not that the backlog is safe-harboured. It is that everything planned inside the statutory window qualifies — and that nothing is planned outside it. ★ The strongest evidence for it is the timing: that sentence was reaffirmed in a filing made on 24 July 2026, which is after the 5 July begin-construction deadline had already passed. The deadline came and went and the company did not revise the claim.
⚠️ What is not available: NextEra has never published a figure for how many gigawatts of the backlog are safe-harboured. Anyone quoting you a precise percentage is guessing. We report the company's qualitative claim, labelled as a company claim.
The largest utility deal in American history, and what it does to the thesis
| The deal | Detail |
|---|---|
| Announced | 18 May 2026 — Dominion Energy becomes a wholly-owned subsidiary of NextEra |
| Terms | All-stock: 0.8138 NextEra shares per Dominion share, plus a pro rata share of $360m in cash. ⚠️ Reported at roughly $67 billion; NextEra holders would own about 74.5% of the combination. |
| ★ Break fees | ★ NextEra would owe Dominion <b>$4.8 billion or $6.5 billion</b> depending on why it terminated; Dominion would owe about $2.2 billion. <b>The size of those numbers tells you how badly each side wants this.</b> |
| Timetable | Applications filed 15 July 2026 · shareholder votes expected early September 2026 · <b>closing expected in the second half of 2027</b>, subject to federal energy and nuclear regulators, three state commissions and both shareholder votes |
| ★ What it creates | ★ About <b>$138 billion of regulated rate base and 110 gigawatts of generation</b> — against Duke Energy's roughly $114 billion and 56 gigawatts. Regulated customers go from about 6 million to 10 million; regulatory jurisdictions from one to four; and the regulated share of cash flow rises from about 70% to 80%. |
| ★★ And the thing that matters most | ★★ On the rating agencies' pro forma figures, <b>Florida falls to about 55% of rate base and Virginia becomes about 36%.</b> Northern Virginia is the densest concentration of data centres on Earth. |
| ⚠️ The integration risk flagged by the agencies | ⚠️ Dominion's Coastal Virginia Offshore Wind project — budget already up from $9.8bn to <b>$11.4bn</b>, about 75% complete, due early 2027. |
Buying Dominion is the most direct possible expression of the electricity-demand thesis — and the most direct possible exposure if that thesis disappoints. It puts a third of the company's regulated assets underneath the world's largest data-centre cluster.
And here is the tension a good analysis has to hold open rather than resolve. Florida's regulatory quality has been NextEra's crown jewel and the entire justification for its premium valuation. After this deal, NextEra is diversifying away from the best regulatory jurisdiction in America into a merely good one. That reduces concentration risk and simultaneously dilutes the very thing that made the shares worth more than everybody else's. Both statements are true.
★ The strongest evidence in the deal's favour is what the credit-rating agencies did with it. On the day it was announced, all three affirmed NextEra's ratings with stable outlooks — and, notably, loosened their downgrade thresholds rather than tightening them. For a company whose entire model depends on capital-market access, that is the single most valuable third-party endorsement available, and it came from people paid to be sceptical.
The step-down almost nobody has written about
Our house rule is that any dividend-paying company gets its payout examined properly. Here that examination produces one genuinely important finding, and it is not the one anybody expects.
⚠️ NextEra has raised its dividend for roughly 31 consecutive years and has been a dividend aristocrat since 2021. Dividends per share have gone $1.87 (2023) → $2.06 (2024) → $2.27 (2025), with about $2.49 expected this year — growth of roughly 10% a year, which is why income investors have loved it.
Watch the rating triggers, not the payout ratio. They are public: S&P's funds-from-operations-to-debt threshold, Moody's cash-flow-to-debt threshold, Fitch's leverage trigger. A sustained breach forces a choice between capital spending, the dividend and the rating. On adjusted earnings the payout is about 60% and has been stable for two years — an ordinary utility payout, and not the binding constraint.
★★ But there is a precedent inside this corporate family, and it is uncomfortable. NextEra Energy Partners — the yieldco it sponsored, since renamed XPLR Infrastructure — suspended its distribution to zero on 28 January 2025. Same parent, same management, same asset class.
The lesson is not that NextEra will cut. The situations are genuinely different. The lesson is narrower and sharper: when this management team had to choose between a distribution and the growth programme, it chose the growth programme — and it did so abruptly.
The premium is still there, and it is much smaller than it was
| Measure | NextEra | Context |
|---|---|---|
| Share price | $86.19 | 52-week range $69.24 – $96.21 · ⚠️ the all-time closing high was about $97.17 on 30 April 2026 — <b>this year</b>, so the drawdown is roughly 11%, not the multi-year collapse often assumed |
| ★ Forward P/E | ~21.7× | On the midpoint of 2026 adjusted earnings guidance of $3.92–$4.02 |
| ★ Against the peers | Southern ~18.4× · Duke ~17.0× | ★ Sector median around 14.4×. So NextEra carries roughly a 15% premium to the best peers and about 48% to the sector. |
| ★★ Against its own history | ★★ premium compressed | ★★ NextEra historically traded at <b>25–30× against a sector in the mid-teens</b> — a premium of 60% to 100%. <b>That premium has narrowed dramatically.</b> This is the single most important valuation fact in the report. |
| Dividend yield | 2.76% | ⚠️ On a payout whose growth rate halves at the end of this year |
| Reported P/E | 19.3× | ⚠️ Ignore this. See Part II — the reported earnings line is a derivatives price. |
| Balance sheet | $98.8bn long-term debt | Investment grade at all three agencies, all affirmed with stable outlooks and loosened thresholds in May 2026 |
| ★ Analysts | 24 buy · 11 hold · 1 sell | ★ Mean target $102.33, about 19% above the price — and <b>the lowest target on the street, $91, is still above today's price</b> |
What is the premium being paid for? A commitment to grow adjusted earnings 8%+ a year through 2032, and again through 2035 — rising to 9%+ if the Dominion deal completes — against a sector that grows 5–6%. If NextEra delivers 8% while Southern delivers 5.5%, a 15% multiple premium is arithmetically cheap. If NextEra delivers 5.5% — because demand disappoints, or the tax cliff bites harder than claimed, or Virginia's regulators prove less generous than Florida's — the premium is unearned and the multiple compresses toward the sector.
⚠️ One technical caveat on the share price that most readers will not know. With an all-stock merger pending and shareholder votes due in September, NextEra's price currently contains a merger-arbitrage component — arbitrageurs routinely short the acquirer and buy the target. Some of the weakness since April is deal mechanics rather than any judgement about the business.
Ghost candidates, a $150m settlement, and a promise that is political · verified August 2026
Verified against NextEra's most recent quarterly filing (24 July 2026) and its FY2025 annual report — i.e. from the company's own reviewed disclosure rather than plaintiff press releases.
★★ THE HEADLINE LEGAL FACT, AND IT IS NEW: the political-activities securities class action has been SETTLED. Shareholders sued NextEra, Florida Power & Light and certain executives in the Southern District of Florida in June 2023, alleging false or misleading statements about the company's "alleged campaign finance and other political activities." The case was dismissed with prejudice in September 2024, then revived on 26 November 2025 when the Eleventh Circuit reversed and a rehearing was denied. ★ In June 2026 the parties agreed a settlement under which NextEra would pay $150 million, entirely covered by insurance, with preliminary court approval still pending. Related shareholder derivative actions were also settled in June 2026, under which insurers pay roughly $16 million to NextEra and the company adopts certain governance changes.
★★ THE DISTINCTION THAT MATTERS, AND WE STATE IT DELIBERATELY. ⚠️ Neither NextEra nor Florida Power & Light has been charged with, or found liable for, any of the underlying conduct. In Florida's 2020 elections, no-party "ghost" candidates entered several state Senate races; former state senator Frank Artiles was convicted on 30 September 2024 on three election-related felony counts over $44,000 paid to one such candidate. ⚠️ Reporting has linked FPL personnel to a political consultancy connected to that scheme, and the appeal-court opinion recites the plaintiffs' allegations. But the Eleventh Circuit's ruling was procedural — it held only that loss causation had been adequately pleaded, and decided nothing about whether the allegations are true. The settlement carries no admission and is insurance-funded. We report the allegations, the third party's conviction and the settlement, and we report that the company has not been found to have done what was alleged.
★ Why it still matters commercially even without a finding: Florida Power & Light's entire competitive advantage is a political asset — a regulator willing to authorise 10.95% and a legislature willing to leave the franchise structure alone. Reputational damage in Florida is not a soft cost for this company; it is a direct threat to the core economic mechanism. And the merger puts NextEra in front of three new state commissions at precisely the moment this record is public.
Other matters: a securities class action over the XPLR yieldco is pending in the Southern District of California with a motion to dismiss undecided, and a related derivative action is stayed. An antitrust claim brought by Avangrid in Massachusetts had its federal and state antitrust theories dismissed in September 2025, with a state-law remainder pending. ★ And a genuinely clean note: for a company operating around 45 gigawatts of generation across 44 states including eight nuclear units, no environmental proceeding meeting the $1 million disclosure threshold appears in either filing.
If you put NextEra Energy through a stock screener this morning, it would tell you the company is in trouble. Its Altman bankruptcy score is 1.08, which nominally puts it in the distress zone. Its free cash flow is minus eleven billion dollars a year. Its return on invested capital is under four percent. Any one of those would normally end my interest in a business. All three are arithmetically correct and every one of them is meaningless here, and explaining why is the most useful thing this letter can do for you.
A regulated utility is not a company that happens to be capital-intensive. It is a legal arrangement wearing the costume of a company. In exchange for a monopoly over a territory and an obligation to serve everyone in it, a state commission sets the prices — and sets them so that investors earn a specified return on the money they have sunk into poles, wires and power stations. So the machine runs like this: raise capital, spend it on assets the regulator approves, those assets enter what is called the rate base, and the regulator then sets prices to return the depreciation plus an authorised return on that base. Repeat forever. Capital spending is not the cost of standing still. Capital spending is the product.
Once you see that, the three frightening numbers dissolve. The bankruptcy score was built in 1968 on a sample of sixty-six American manufacturers, and it punishes thin working capital, high leverage and low asset turnover — all three of which are, for a utility, features rather than faults. Florida's regulator literally sets the capital structure, at a 59.6% equity ratio. The regulator chose the leverage the score is complaining about. Negative free cash flow is the machine working: a utility with positive free cash flow has stopped growing its rate base, which means its earnings growth has stopped too. And return on capital is the wrong question entirely, because a regulated utility does not earn a return — it is awarded one. The right question is what the regulator authorised and whether the company is achieving it. Florida authorised 10.95%. Florida Power & Light earned about 11.70%.
I want to give you the fourth trap too, because it catches professionals. NextEra's reported earnings are close to unusable. Its competitive arm hedges electricity it will sell years from now, and the changing market value of those contracts lands in this quarter's profit line even though no cash moves and no power is generated. Last quarter that added thirty cents to a $1.15 quarter. A year earlier it subtracted. Between 2024 and 2025 NextEra's reported earnings per share fell while its adjusted earnings rose eight percent. One of those numbers describes a business and the other describes the weather in the derivatives market. I will add the fair objection: management picks the exclusions and is paid on the adjusted figure. So I applied the test — does the adjustment cut both ways? Here it verifiably does, and I would want that same test applied to anyone making the same claim.
So what is this business actually worth owning for? Buffett wrote the thesis in 2011 in seventy words: "Take care of your customer, and the regulator — your customer's representative — will take care of you." Florida Power & Light is the cleanest living example of that sentence in America: bills about thirty percent below the national average, operating costs far below the industry, top-decile reliability — and, as the payoff, permission to earn nearly eleven percent on an ever-growing pile of capital. It even holds two hundred and twenty-six franchise agreements running out to 2055 under which towns contract not to set up their own utility.
But Buffett also wrote the warning, and it is the more important passage. In 2023, after watching regulation turn hostile in California and Hawaii, he wrote that "the fixed-but-satisfactory-return pact has been broken in a few states" and admitted, with unusual bluntness, "I did not anticipate or even consider the adverse developments in regulatory returns." That is the entire risk in this security. The compact is not a contract. It is a promise, and it is political. Florida's 10.95% is reportedly the highest authorised return in the contiguous United States, which is exactly the sort of number that becomes an election issue when bills rise — and the 2025 rate agreement is under appeal at the Florida Supreme Court right now.
Two more things happened that most coverage has not caught up with. The first is that in May, NextEra agreed to buy Dominion Energy for about sixty-seven billion dollars in stock — the largest utility merger in American history, closing in late 2027. It would create roughly a hundred and thirty-eight billion dollars of regulated rate base against Duke's hundred and fourteen. And it does something subtle: it moves about thirty-six percent of the rate base to Virginia, home to the densest concentration of data centres on Earth. That is the demand thesis made literal — and it also dilutes the Florida regulatory quality that justified paying a premium for this company in the first place. Both things are true and I am not going to pretend otherwise. What I will say in the deal's favour is that all three credit-rating agencies affirmed NextEra on the day and loosened their downgrade thresholds. People paid to be sceptical were not.
The second is the one I would most want an income investor to hear, and it is four months away. NextEra has raised its dividend for about thirty-one years and has grown it at roughly ten percent a year, which is why people own it. In its own guidance it says it expects that ten percent rate "through 2026" — and then six percent a year from the end of 2026 through 2028. That is disclosed, it is confirmed in the investor materials, and it is barely written about. Anyone buying NextEra today as a ten-percent dividend grower is buying something the company has already announced it will stop being at Christmas. The reason is defensible — earnings are guided to grow eight percent while the dividend grows six, so the payout ratio falls and more cash stays inside to fund a three-hundred-billion-dollar building programme. That is prudent. It is also, unmistakably, growth capital being prioritised over income. And there is a precedent inside the family: the yieldco this same management sponsored cut its distribution to zero in January 2025. The lesson is not that NextEra will cut. It is that when this team had to choose between a distribution and the growth programme, it chose growth, and it did so abruptly.
My verdict is "The Screens Say Distress — The Regulator Says 11.7%," and I score it 7.0. An eight for a moat that is genuinely legislated, an eight for a growth commitment no other utility comes close to matching — eight percent a year, guided out to 2035, and raised rather than cut — and a six for financial strength, because a business that must raise twenty-three billion dollars a year to execute its own plan is only ever as strong as the bond market's mood. On price, the premium is still there but it has shrunk a great deal: NextEra used to trade at twenty-five to thirty times against a sector in the mid-teens, and today it is about twenty-two against Southern's eighteen. If it delivers eight percent while the sector delivers five and a half, a fifteen percent premium is cheap. If it delivers five and a half, the premium goes.
So: a fine business, honestly priced, with a real growth commitment and a political risk that no amount of analysis can eliminate. I would own it, and I would rather own it nearer seventy-five dollars — about nineteen times forward earnings, where the premium to the best peers has essentially disappeared and you are being paid to carry the regulatory risk rather than assuming it away. Note that every analyst covering this company, even the most bearish, has a target above today's price, which is the opposite of what I found at Paychex last week and is worth something. But buy it for what it is: a legislated monopoly compounding at eight percent, whose dividend is about to grow at six. Not a bond. Not a ten-percent income grower. And never, ever, a company you can judge from a screen.