The Operating Beat We Asked For Arrived: Upgrading, With Georgia's Pricing Review Six Weeks Out

Results vs. Consensus

Q3 2026 Scorecard

The consensus figure needs two notes. First, the providers we track carried the quarter at $1.69 to $1.71 of adjusted EPS, and we use the most widely cited $1.71. Revenue estimates ranged from $8.36 billion to $8.47 billion; against the widely cited $8.47 billion the beat is 4.4%.

Second, a correction to our July note. We took the company's third-quarter estimate as $1.50, from the call transcript, and argued that it sat 6.3% below the prior year. The company's own July slides show $1.65, and the company measures this quarter's beat against that figure: twenty-nine cents. The roughly $1.65 the Street carried in July was therefore not analysts adjusting for a pattern of conservatism, as we wrote. It was the Street adopting the company's number, as it did in the first two quarters. Our July second-half arithmetic still holds, because it rested on the full-year range, but on the correct estimate the July guide implied a soft fourth quarter of $0.39 to $0.49, not a weak third. Southern has now beaten its own quarterly estimate by twelve, thirteen and twenty-nine cents.

Year-Over-Year Comparisons

Quarter-Over-Quarter Comparisons

The third quarter is Southern's seasonal peak, so the sequential comparison mostly measures summer. It is shown because two lines move against the season and deserve attention: the tax rate and the venture capital marks.

The Thirty-Four Cent Bridge

Southern publishes a cent-by-cent reconciliation of the change in adjusted EPS, and for the third quarter running it is where the quarter is decided. The right-hand column is ours, and it uses the same labels we applied in July so the two quarters can be read side by side.

Quality of Beat: Revenue

Total operating revenue rose $1,020 million, or 13.0%, to $8,843 million, 4.4% above the $8.47 billion consensus. Unlike the first quarter, when most of the increase was commodity pass-through, this time the growth sits where the return is earned. Strip retail fuel revenue and natural gas revenue from both years and the remaining $6,579 million still grew 13.0% against $5,824 million.

The line that matters is non-fuel retail electric revenue, which rose $690 million to $5,132 million, or 15.5%. The company's retail revenue bridge splits the change cleanly: $322 million from rates and pricing, $218 million from weather and $150 million from sales growth. Add the three and you get exactly $690 million. Retail kilowatt-hour sales rose 7.7%, so non-fuel revenue per retail kilowatt-hour rose about 7.2%, from 10.48 cents to 11.24 cents. In the second quarter the same measure fell about 0.3%.

So the pricing test we set in July passed by a wide margin. The reason matters as much as the result. The 10-Q attributes the $322 million "primarily due to higher contributions from commercial and industrial customers with variable demand-driven pricing at Georgia Power and an increase in Rate CNP Compliance revenues at Alabama Power." Georgia alone contributed $296 million of it. In the second quarter the consolidated rates-and-pricing line subtracted $10 million, attributed to lower contributions from the same Georgia customers, and over the first half Georgia's line was a net drag. Georgia's year-to-date rates-and-pricing line is $216 million, which means the third quarter delivered more than all of it. The revenue is real and it is legitimately non-fuel. It is also clearly sensitive to the summer, because demand-driven pricing rises when the system runs hot.

Assessment: Pricing per kilowatt-hour rose for the right structural reason, a growing commercial base paying demand-linked tariffs, and for a cyclical one, a record-hot quarter. We give credit for the first and discount the second, and the December 15 Georgia decision will show how much of the first survives.

Quality of Beat: Margins

Operating margin was 33.9% against 33.2%, seventy-seven basis points better, and operating income rose $406 million, or 15.7%. After two quarters of essentially no operating growth this is a different result, and it came in spite of the cost lines rather than because of them.

Non-fuel operations and maintenance rose $205 million, or 12.5%. The 10-Q names the pieces: $46 million more generation expense across the electric utilities, $42 million more transmission and distribution cost and $39 million more storm damage recovery at Georgia Power, $24 million of charges at Southern Company Gas including the Nicor disallowance, and the absence of a $33 million credit Georgia Power booked a year ago on Vogtle (which the company excludes from adjusted earnings). Alabama Power also drew less from its reliability reserve, which helped expenses in the first half and contributed less this quarter.

Taxes other than income taxes rose $146 million, or 50.7%, and $116 million of that is Georgia Power property tax "largely resulting from the timing of prior-year actualizations." In the second quarter the same mechanism reduced the line by $37 million. On our reading, that is roughly 7.5 cents of the quarter's "interest expense and other" drag, and it is timing, not run-rate.

Depreciation rose only $22 million, for the third consecutive quarter of the same offset: $114 million more from new plant in service and $100 million less "resulting from the extension of Georgia Power's 2022 ARP." Year to date the extension has removed $297 million of depreciation.

Assessment: The margin expansion is genuine and it absorbed real cost growth, including about seven and a half cents of non-recurring property-tax timing. It still leans on the rate-plan depreciation offset, which has now reached $297 million in nine months and ends when the Georgia rates reset.

Quality of Beat: EPS

Below the operating line the quarter was unusually clean. The construction allowance rose $50 million to $140 million and contributed four cents, the same as in the second quarter. Earnings from equity method investments rose $44 million to $77 million, and the 10-Q again attributes the increase to gains on Southern Holdings' energy venture capital funds, $46 million this quarter. That book now stands at $180 million, up from $156 million in June and $124 million at year end. On our arithmetic it is about three cents of the four-cent parent-company row.

The effective tax rate was 18.5% against 19.1%. That headline includes a $54 million tax charge tied to the debt extinguishment items, which the company excludes from adjusted earnings; without it the rate was about 16.4%. The income-tax row in the bridge is three cents, down from eight in the second quarter, because the Georgia valuation-allowance comparison that made the second quarter did not recur at the same scale.

Dilution took nine cents, more than double the four cents of the second quarter. Average shares rose 5.0% year over year to 1,157 million as the last of the forward-sold equity settled: $1.2 billion in the quarter, which completes all $5.3 billion priced under the at-the-market program.

Assessment: The two sources of earnings that worried us most in July both shrank as a share of the increase: the tax row fell from eight cents to three, and the construction allowance held at four cents of a larger thirty-four. The venture marks did not reverse, which is fortunate rather than structural. Dilution, at nine cents, now ties operations and maintenance as the second-largest drag in the bridge, behind the eleven-cent interest-and-other row.

The GAAP-to-Adjusted Wedge

Adjusted EPS exceeded GAAP EPS by seventeen cents against five cents a year ago, and for nine months the wedge widened from twenty cents to thirty-eight. Three items make up the third quarter. Wind repowering at Southern Power cost $118 million pre-tax, $92 million after tax, about eight cents. Debt extinguishment cost about eight cents more: a $43 million loss ($40 million after tax) on repurchasing part of the 2024 convertible notes and redeeming junior subordinated notes, plus a $57 million "change in tax benefits related to the extinguishment of debt in 2025." Nicor Gas's disallowance added $9 million pre-tax, about one cent. A year ago the wedge was narrower because Georgia Power booked a $33 million credit on Vogtle demobilization costs.

The repowering estimate rose again. At June 30 the company projected about $205 million remaining in 2026 and $120 million in 2027. It incurred $118 million in the quarter, which should have left about $87 million for 2026. It now projects about $100 million in 2026 and $135 million in 2027. Total expected remaining charges rose about $28 million in the quarter, after rising about $33 million in the second quarter. The completion date is unchanged at the third quarter of 2027.

Segment Performance

Operating-company revenues do not sum to the consolidated total: the complementary businesses (PowerSecure, Southern Holdings, the fiber and telecommunications operations) and intercompany eliminations sit outside the five named registrants. Consolidated net income is stated after the parent company and other, a loss of $299 million against a loss of $237 million a year ago.

Georgia Power

Revenue rose 18.6%, earnings before income taxes rose 19.0% and net income rose 22.0%. The contrast with July is the point. In the second quarter roughly two-thirds of Georgia's net income growth came below the pre-tax line, from the valuation allowance. This quarter almost all of it came above it. Georgia's retail revenue rose $723 million, or 21.5%: $296 million from rates and pricing, $152 million from weather, $115 million from sales growth and $160 million of fuel recovery that does not affect earnings.

Weather-adjusted commercial sales rose 10.2%, "largely driven by data centers," while weather-adjusted residential rose 1.0% and industrial fell 1.2% on paper, textiles and primary metals. Taxes other than income taxes nearly tripled, from $75 million to $211 million, on the $116 million property-tax timing item. Depreciation fell $10 million to $516 million because the $100 million rate-plan offset exceeded the $77 million added by new plant.

Two regulatory items set the frame for 2029. On August 26 Georgia Power agreed with the commission staff to put at least $949 million a year of downward pressure on rates for non-large-load customers into its next base rate case, levelized over 2029 to 2031, and agreed not to seek recovery from those customers for any revenue shortfall if a data center terminates early. And the commission opened its review of the variable-pricing methodology, with a decision expected December 15.

"Beyond lowering rates in 2026, let me remind you, Georgia Power has committed to deliver nearly $3 billion of customer benefits that will help lower costs for the typical residential customer by approximately $180 annually after the current rate freeze period." — Chris Womack, Chairman, President and CEO

Assessment: Georgia earned this quarter in the pre-tax line, which is what we wanted to see. It also concentrated the quarter's upside in a single tariff that is now in front of the commission, and it has promised existing customers close to $2.85 billion of rate relief over the three years after the freeze, the same years in which the capital it is building finally enters rates.

Alabama Power

Revenue rose 8.2%, pre-tax earnings rose 14.2% and net income rose 12.4%, a third consecutive quarter of clean pre-tax growth this year. Retail revenue rose $193 million, or 9.4%: $59 million from weather, $31 million from rates and pricing, $28 million from sales growth and $75 million of fuel and other cost recovery. The rates-and-pricing increase is Rate CNP Compliance revenue, a cost-recovery mechanism, and the matching compliance expenses rose $16 million to $85 million, so the earnings contribution is smaller than the revenue line suggests. Industrial sales rose 4.7% on primary metals and pipelines, the strongest industrial number in the system.

In July we said to watch whether pre-tax growth persisted once the reliability reserve stopped contributing. It did. Alabama has now used the full $60 million it planned to draw this year, with $124 million left in the reserve, and the quarter's operations and maintenance rose partly because it drew less. Roughly 2 GW of new large-load contracts were signed in Alabama during the quarter.

Alabama is also building an accounting cushion against its own growth. On October 16 it petitioned the commission to record part of what it bills large-load customers under an "Incremental Revenue Component" as a regulatory liability, effective October 1, to cover potential cost exposure from investments made for those customers. The liability would be released into earnings only once risk thresholds are met.

"You can think about it where a cost would be incurred, if you will, from a rate-making perspective, and that will flow through the income statement and create a regulatory liability." — David Poroch, Chief Financial Officer

Assessment: The cleanest utility in the group again, and the reliability-reserve question we raised is answered. The new liability mechanism is prudent and good for the regulatory relationship. It also means some of Alabama's large-load revenue will reach the income statement later than it is billed, and the company has not sized how much.

Southern Power

Revenue rose 14.0% and net income rose from $3 million to $52 million despite $118 million of pre-tax repowering charges, and the segment added five cents to adjusted EPS. The 10-Q attributes the improvement to "higher revenues driven by increased demand and higher market prices of energy and an increase in capacity revenues related to natural gas PPAs." Of the 147 MW Grant Plains repowering project, 82 MW were in service by the time of the filing, and the 400 MW of gas turbine uprates at Franklin and Wansley carry $560 million to $620 million of remaining cost.

The recontracting update was promised for "later this year" in April and again in July. It did not come. Asked directly, management said negotiations are ongoing, that opportunities extend beyond the Southeast, and that the work is "probably going to stretch out over the next few years."

"I mean, I think a lot of times, because of the growth we're seeing in the regulated part of the business, sometimes Southern Power activity it's kind of put to the side." — Chris Womack, Chairman, President and CEO

Assessment: A good quarter on hot-weather market pricing plus new gas capacity revenue, the first real contribution from the segment this year. The recontracting disclosure has now slipped three quarters running, and "the next few years" is a longer horizon than "later this year."

Southern Company Gas

The third quarter is the gas business's seasonal trough. Revenue rose 5.3% and net income fell from $25 million to $21 million, with no change in the adjusted bridge. Base rate increases at Nicor Gas and Atlanta Gas Light added $50 million of revenue, offset by lower gas cost recovery and $24 million of higher operations and maintenance expense, including another $9 million Nicor disallowance charge the company excludes. The larger news is outside the quarter: Southern Natural Gas, which Southern owns 50% of, committed to the roughly $7.6 billion Southeast Connector pipeline, designed for about 2.6 billion cubic feet a day of capacity.

Assessment: Immaterial to the quarter and a meaningful addition to the federally regulated capital base. The Nicor disallowance has now appeared in three consecutive quarters, $19 million pre-tax year to date, with "further charges may occur" still in the footnote.

Mississippi Power

Revenue rose 3.5%, pre-tax earnings rose 7.4% and net income rose 6.0%, on higher non-fuel wholesale revenue partly offset by lower other revenue. That reverses the 11.9% decline we flagged in July, which was never explained.

Assessment: Back to modest growth and still about 4% of net income. Mississippi did appear on the call, as part of the large-load pipeline management now describes across all three states.

Load and Customer Metrics

Percentage changes are as reported and are computed from unrounded volumes. The final column is stated on a weather-adjusted basis and is not derivable from the two volume columns beside it.

The volume table is the most important single page of the release this quarter, and it says two things at once. The first is that the heat was extraordinary. Raw residential sales rose 10.2% against a weather-adjusted 1.3%, so weather added roughly nine points to residential volume alone. Management called it one of the warmest quarters in its territory in fifty years, with electricity sales higher than any prior quarter on record.

The second is that the weather-normal base broadened again. In July we wrote that the non-data-center demand base was not growing: weather-adjusted residential sales fell 0.7% and industrial was flat. This quarter weather-adjusted residential rose 1.3% on customer growth, industrial rose 1.8% on primary metals, lumber and manufacturing, and commercial rose 7.3% on data centers. Electric customers grew by about 48,000 over the year. Weather-normal retail growth of 3.5% is the highest the company has reported this year, after 2.3% in each of the first two quarters, and year to date it is now 2.7%.

Data center usage accelerated again, to 59% growth from 55% in the second quarter and 42% in the first. Commercial still carries most of the weather-normal growth, but as in the first quarter, and unlike the second, it is not the whole of it.

Key Topics & Management Commentary

Overall Management Tone: Management was the most assertive it has been this year, leading with "we do what we say we will do," a full-year number ten cents above the original range and an explicit signal that February will raise the outlook. The call was less forthcoming on composition than the filing: the largest single driver of the quarter was described only as growing revenue from large customers, and the commission review attached to it went unmentioned. Analyst pushback was concentrated on Georgia politics, and management answered it with confidence and process rather than with numbers.

1. The Largest Line in the Bridge Is a Tariff Georgia Has Put Under Review

The "retail revenue impacts" row added twenty-one cents, the biggest single contribution in the bridge. The 10-Q identifies it as higher rates and pricing revenue of $322 million, "primarily due to higher contributions from commercial and industrial customers with variable demand-driven pricing at Georgia Power." Georgia contributed $296 million. On our arithmetic that is about nineteen cents of adjusted EPS from one tariff in one quarter.

In the same filing, under regulatory matters, sits this paragraph: "In the third quarter 2026, the Georgia PSC commenced a proceeding to review the methodology for assigning a portion of the revenue charged to the variable demand-driven pricing customers to fuel cost recovery revenues. A decision by the Georgia PSC is expected on December 15, 2026." Revenue assigned to fuel recovery passes through to customers and does not earn a return. If the commission changes the methodology, some of what was booked as non-fuel revenue could in future be treated as fuel recovery.

"The primary drivers of our performance for the current quarter compared to last year were warmer weather, increased usage and customer growth, growing revenue from large commercial and industrial customers, including data centers, along with higher earnings from Southern Power." — David Poroch, Chief Financial Officer

That is the whole of the call's description. "Growing revenue from large commercial and industrial customers" is accurate, and it is also how one would describe a pricing outcome that rises with system demand on a hot summer. It is worth recalling the April framing of how Southern recovers the cost of large loads: the CFO said then that the company is "not held captive to a variable pricing methodology" and recovers its costs "through that minimum bill, not through variable pricing." Both things can be true. Minimum bills can protect cost recovery while demand-driven pricing supplies upside. But this quarter's upside came from the variable side.

Assessment: This is the most important open item for the stock and the one the company said least about. We do not know how much of the $296 million the review could reach, whether any change would be prospective only, or how much of the tariff's revenue is weather-driven rather than structural. December 15 will tell us. Our 2027 estimate already removes all of the above-normal weather and a third of Georgia's year-to-date pricing gain, before any regulatory change.

2. One of the Warmest Quarters in Fifty Years

Weather added fifteen cents against the prior-year quarter, and $218 million of non-fuel retail revenue. It also flowed through to the two other lines that moved most: demand-driven pricing in Georgia and market energy prices at Southern Power.

"The third quarter was one of the warmest quarters in our service territory over the last 50 years and we delivered electricity sales that were higher than any prior quarter on record, while providing the best-in-class fleet reliability." — Chris Womack, Chairman, President and CEO

For the nine months the weather row is nine cents, because the first half was five cents unfavorable against 2025. Measured against normal rather than against last year, the company's slides put the quarter at seventeen cents above normal, against three cents a year earlier, and the nine months at seventeen cents above normal, against eight. At normal weather the quarter earned about $1.77 against $1.57, growth of 12.7%. That is a strong number on its own, before counting the indirect effect of the heat through pricing.

Assessment: The reported number is real money, and it is paying for a deliberate pull-forward of costs into the fourth quarter. It is not a run rate. Any reading of the $4.70 full-year figure as a base for 2027 growth should start by removing it.

3. Weather-Normal Growth Broadened Beyond Data Centers

The durable part of the quarter is the eleven-cent retail sales row and the weather-adjusted volume table behind it. All three customer classes grew on a weather-normal basis, as they did in the first quarter, after residential fell and industrial was flat in the second.

"We are seeing the highest level of weather-normal sales growth in generations, and that growth continues to accelerate, up 3.5% in the third quarter and 2.7% higher year-to-date. All customer classes are higher, led by commercial sales, which were up 7.3% compared with the third quarter of 2025." — Chris Womack, Chairman, President and CEO

Residential growth came from customer additions, about 15,000 new electric residential customers in the quarter. Industrial growth came from primary metals, lumber and transportation. Management also cited nearly $25 billion of economic development announcements in the quarter, including Airbus, OpenAI, Pirelli and Walmart, and more than $45 billion for the year including the later Mitsubishi Power and Audi announcements.

Assessment: This answers the concern we raised in July that Southern's demand growth had become a single-class story. It has not. Commercial still leads, but residential and industrial both contributed, and the economic development pipeline behind them is the largest the company has seen.

4. Twenty-One Gigawatts, and What the 10-Q Adds to the Number

The contracted large-load book rose from over 17 GW to 21 GW, with Alabama Power and Georgia Power each adding about 2 GW. The late-stage pipeline fell from 8 GW to 7 GW, so the funnel converted faster than it refilled at that stage, while the prospective pipeline is still described at about 75 GW.

"Large load and data center customers are paying at a minimum their full share, enabling Alabama Power and Georgia Power to pass along meaningful benefits to customers today and into the future." — Chris Womack, Chairman, President and CEO

That wording matters to us. In July the standard was restated as "at least 100% of the incremental cost to serve," which we flagged as weaker than April's "full share." The prepared remarks went back to "full share," and in Q&A the CFO's formulation was minimum bills "that cover at least 100% of the cost being introduced into the system." The slide deck, though, still describes minimum bills negotiated "to cover at least 100% of annual incremental costs to serve" and termination payments "tied to the remaining incremental costs to serve over the life of the contract." The deck also gives a 17-year weighted average contract length across 33 projects, 31 of them already under construction, and about $25 billion of collateral required by 2030 at an average AA- rating.

The 10-Q adds detail the call did not. It describes about 16 GW of contracts already reviewed by the state commissions and about 4 GW more agreed and awaiting review, about 20 GW in all, with service under the remaining contracts expected to begin through 2029. "All of the Large Load Contracts contain minimum bill provisions." But only about 16 GW, "including substantially all signed after 2024," also carry minimum duration, termination payments and collateral.

Assessment: The book keeps growing. The spoken pricing language moved back toward the stronger standard, but the written terms still say incremental, and we treat the written version as the contract. The new fact is that roughly a fifth of the contracted load, mostly older contracts with a five-year average length, sits outside the termination-payment and collateral provisions. The minimum bill level itself remains undisclosed for a third consecutive quarter.

5. Georgia Moved Early-Termination Risk Away From Residential Customers

On August 26 Georgia Power filed a joint statement with the commission staff that raised its commitment to non-large-load customers. It agreed to include at least $949 million a year of downward pressure on their rates in the next base rate case, levelized over 2029, 2030 and 2031. It also agreed that "it will not seek recovery from non-large load customers for any revenue deficiency associated with serving large load data center customers in the event of an early contract termination."

The commission keeps discretion to allocate any such shortfall to other data center customers. But the default changed: if a data center leaves early, the residential and small-business base will not pick up the gap.

Assessment: Good politics and the right commitment to make in this environment. It also puts more weight on the collateral and termination provisions, because collateral, other data center customers or shareholders now stand behind any early exit. That makes the 4 GW without termination payments more relevant than it was.

6. Alabama Will Record Part of Its Large-Load Revenue as a Liability

Alabama Power has included an "Incremental Revenue Component" in certain large-load contracts to cover potential cost exposure from the investments it makes to serve them. On October 16 it asked the commission to let it record matching amounts as a regulatory liability, with a charge to amortization, as the component is billed. Once "risk mitigation thresholds are achieved," the liability would be released back into income.

"So it's intentionally designed to be flexible. And that would probably be more, I would think, in the latter part of the contract lives. So probably taking that well into the next decade and perhaps even touching into the 2040s depending on the contract." — David Poroch, Chief Financial Officer

The petition asks for an October 1 effective date, so if approved it would touch the fourth quarter. The company has not said how large the incremental component is or what the thresholds are.

Assessment: A conservative and commission-friendly design that converts part of Alabama's large-load revenue into a reserve. For earnings it is a deferral, possibly a long one, and it is a modest headwind to how quickly Alabama's 2 GW of new contracts show up in reported EPS.

7. $4.70, and a Fourth Quarter Guided 45% Below Last Year

The fourth-quarter estimate is $0.30 against $0.55 a year ago. Combined with $4.40 for the nine months that gives $4.70 for the year, ten cents above the top of the original range and 9.3% above 2025's $4.30.

"Our adjusted EPS estimate for the fourth quarter is $0.30 per share, which, combined with our strong year-to-date performance, would represent full year adjusted earnings of $4.70 or $0.10 higher than the top of our original 2026 adjusted earnings guidance range, representing 9% year-over-year adjusted earnings growth." — David Poroch, Chief Financial Officer

Asked what sits inside the $0.30, the CFO did not itemize it. He said the company tries "to fix the roof while the sun is shining" and has been doing work "to further bolster our confidence around the future years." Read plainly, the summer windfall is being spent on maintenance and other costs that would otherwise land in 2027. Hurricane Isaias, which reached the Gulf Coast in early October and cut power to more than 650,000 customers, also falls in the fourth quarter.

Assessment: This is a managed number, and we think sensibly managed. Pulling cost forward in a strong year lowers the 2027 cost base and protects the guide that February will reset. It also means the $4.70 is not a clean run rate in either direction: it includes seventeen cents of above-normal weather and an unquantified amount of pre-spent 2027 cost.

8. A February Raise Signaled, and a Capital Plan That Could Approach $100 Billion

Management said it will refresh the outlook through 2031 in February and expects an opportunity "to potentially raise the upper end of our near-term guidance ranges." In Q&A the CFO described the effect as a higher starting point for the 7% to 8% growth path from 2028.

On capital, about $4 billion of projected investment has been added above the $81 billion base plan: the 400 MW of Southern Power gas uprates, the Georgia nuclear uprates done with Google, and Southern's share of the Southeast Connector pipeline. Including potential company-owned wins in the Georgia and Alabama procurements and other opportunities, the CFO said the plan "could potentially approach $100 billion over a 5-year period."

"Our load forecast in Georgia really does kind of suggest that we would be at the near the top end of that 2 to 6 range." — David Poroch, Chief Financial Officer

He also acknowledged cost inflation on new generation: the $2 billion per gigawatt rule of thumb "does span a bit past $2 billion" and excludes transmission. Alabama's procurement will reach clarity "probably later in 2027."

Assessment: If company-owned resources won all of a Georgia procurement near 6 GW, at $2 billion or more per gigawatt that would be $12 billion or more of potential rate base before Alabama. That is the long-dated upside we said was not in the plan. It is now being described in numbers, which is progress, though none of it is in the base plan until the commission certifies it.

9. The Equity Need Went Back Up

All $5.3 billion of forward-priced equity has now settled, the last $1.2 billion in the quarter at an average of $95. With the added $4 billion of capital, the projected remaining equity need through 2030 rose to about $2 billion from $1.1 billion in July.

"Considering the equity addressed to date, along with the incremental $4 billion of projected investment above our base capital plan, we project a remaining equity need of approximately $2 billion through 2030, consistent with our communication of to fund additional investment with approximately 40% equity or equity equivalents." — David Poroch, Chief Financial Officer

In August the parent also issued about $2.73 billion of new convertible notes due December 2027 and September 2029, and used part of the proceeds to repurchase about $369 million of the 4.50% convertibles due 2027. Average shares rose 5.0% year over year and dilution took nine cents from the quarter.

Assessment: The 40% equity rule is doing what it says: more capital means more shares. The increase follows the added investment, though it is smaller than 40% of it would imply, and the 17% FFO-to-debt objective for 2029 is intact. But in July we judged the near-term dilution worse and the long-run need smaller. This quarter the long-run need moved the wrong way too, and a plan approaching $100 billion would add more.

10. Georgia's Election Changes the Commission That Will Hear the 2028 Case

On November 3 Georgia voters elected Democrats to both commission seats on the ballot, re-electing one incumbent and taking a seat Republicans had held, which gives the party a 3-2 majority on the Public Service Commission for the first time in about twenty years. They also elected Keisha Lance Bottoms governor. During the campaign the governor-elect called for a statewide pause on new data center development and said she was unsure whether that could be done by executive order or would need legislation. In Alabama, Republicans held both commission seats on the ballot.

The first question on the call was about this, and management's answer was the affordability record: frozen base rates, the summer fuel-bill reduction and the nearly $3 billion of benefits committed to existing customers after the freeze. Management said it had already briefed the governor-elect on how its large-load contracts work during the campaign.

"I mean if you look at exit interviews from the elections at the top of the list of concerns from those that voted were concerns about issues of affordability. If you look at what we have been doing with rate freezes, with benefits, with how we are providing benefits from growth to customers, existing customers as well as to the communities, we think that's a strong platform to stand on." — Chris Womack, Chairman, President and CEO

The timing matters more than the result. Georgia Power's base rates are frozen until 2029, and the company has "to make a filing in '28." That case is when the construction program finally enters rates, when the $949 million a year of committed rate relief must be delivered and when the rate-plan depreciation offset ends. It will now be heard by a commission whose new majority campaigned on consumer costs. The variable-pricing decision due December 15 is a smaller test of the same relationship.

Assessment: This is the largest change in the operating environment since initiation, and it is a change in risk rather than in results. Signed contracts are protected by their terms, and nothing has been proposed against them. The exposure is to the pace at which the 7 GW late-stage and 75 GW prospective pipeline converts, and to the terms of the 2029 rate reset. We add it to the scorecard as a new risk at emerging.

11. Repowering Estimates Rose Again

The remaining wind repowering charges now total about $235 million, $100 million in 2026 and $135 million in 2027, after $118 million incurred in the quarter. On the June estimate the remainder should have been about $207 million. The program's completion date is unchanged at the third quarter of 2027, and 82 MW of the 147 MW Grant Plains project was in service by the filing date.

Separately, the company excluded about eight cents of debt-related costs from adjusted earnings, including a $57 million change in tax benefits on 2025's convertible repurchases. The release notes that "similar transaction costs may occur in the future."

Assessment: The amounts are small against a $2.2 billion quarterly adjusted net income, but the estimate has now risen in two consecutive quarters. The exclusions table now carries repowering, debt extinguishment, Nicor and Vogtle items in the same year, and the nine-month wedge between GAAP and adjusted EPS has nearly doubled.

12. Nuclear: Uprates Yes, New Units No

On October 6 the Georgia commission approved a 49 MW uprate at Plant Hatch and a tariff that lets new large-load customers buy the zero-emission credits from the Hatch uprate and from 47 MW of previously approved uprates at Vogtle Units 1 and 2. The 10-Q says the credits are fully subscribed for 25 years, and the call identified the partner as Google. The customer pays an upfront enrollment contribution and ongoing credit payments, which reduce the cost to everyone else. The uprates reach commercial operation between 2028 and 2031.

"We're not in line to do this, just making it perfectly clear. That's not for us." — Chris Womack, Chairman, President and CEO

Assessment: The uprate tariff is a neat template: a hyperscaler pays for nuclear capacity that existing customers would otherwise fund. On new large reactors the message is the same as in July, stated even more firmly. Southern will help the industry and the administration, and it will not build the next one.

Guidance & Outlook

Management framed the guidance around credibility and momentum rather than around the quarter's drivers. The full-year number was presented as the outcome of growth across the territory, and the February update as the moment that growth gets written into the plan.

"Consistent with the higher 2026 expectations we shared today, the sustained momentum we see gives us strong confidence for the future and provides an opportunity to potentially raise the upper end of our near-term guidance ranges that ultimately will support an improved outlook into the next decade." — Chris Womack, Chairman, President and CEO

Implied fourth quarter. The full-year figure is a sum, not a range, so the arithmetic is short.

Take out the nine cents of weather against 2025 and the year is about $4.61, 7.2% above 2025 and inside the 7% to 8% long-term objective. Measured against normal weather, the nine months carry seventeen cents of benefit, so a normal-weather 2026 is about $4.53 before adding back the costs pulled into the fourth quarter. That is the right base for 2027, because February's guidance will assume normal weather, and it is the base we use below. The variable-pricing gain and the fourth-quarter pull-forward push in opposite directions from there.

Street at: Before the print the Street carried about $4.59 for 2026 and about $0.42 to $0.44 for the fourth quarter. The company's $0.30 fourth-quarter estimate sits twelve to fourteen cents below that, which is the pull-forward expressed in numbers. Expect the year to move to about $4.70 and the fourth quarter to come down. The larger revision will come in February, when the 2027 and 2028 ranges are reset.

Guidance style: Conservative, and now demonstrably so for three quarters. Southern cleared its own quarterly estimate by twelve cents in the first quarter, thirteen in the second and twenty-nine in the third. The fourth-quarter estimate of $0.30 is the first one that looks deliberately low for a reason management has explained, the decision to spend the summer's surplus on next year's costs. We would not expect the usual twelve-cent beat in the fourth quarter, because the company controls how much it spends.

The longer-dated arithmetic. In July we noted that the original 2026 range, 4.7% to 7.0% growth off $4.30, never reached the top half of the long-term objective. At $4.70 the year grows 9.3%, or about 7.2% without the weather gain against 2025. If February raises the upper end of the 2027 and 2028 ranges and then runs 7% to 8% from 2028, the outlook will finally match the language management has used all year.

Analyst Q&A Highlights

Whether the Plan Holds Up in Front of a Changed Georgia Commission

The first question of the call went straight to the election: a party change at the Georgia commission, a rate settlement that ends in 2028 and a major rate case to follow, set against a company preparing to raise its growth outlook. Management answered with the affordability record rather than with any view of the new commissioners.

Q: "I guess, Chris, how do you kind of build confidence here with the plan, especially as you roll forward with sort of the untested PSC dynamic that's formed and a major rate case that's forthcoming, especially as you're thinking about raising the CAGR, et cetera?"
— Shahriar Pourreza, Wells Fargo

A: "If you look at what we have been doing with rate freezes, with benefits, with how we are providing benefits from growth to customers, existing customers as well as to the communities, we think that's a strong platform to stand on."
— Chris Womack, Chairman, President and CEO

Assessment: A reasonable answer and the only one available two days after an election. The bet is that customer benefits funded by large loads are popular regardless of party. That is plausible, and it is untested, which is the questioner's point.

Whether Another Rate Stay-Out Is on the Table

A follow-on question asked whether Southern could offer further stay-outs in Georgia or Alabama and when it would have to start making filings. The answer was the clearest statement on the call about the 2028 timeline.

Q: "Just what's the ability to offer those further stayouts in either Georgia or Alabama? And at what point in the time line do we have to kind of start thinking about that or making subsequent filings, understanding that you don't have many filings in front of the commission right now?"
— Nicholas Campanella, Barclays

A: "Yes. I mean, we have to make a filing in '28. And so I think it's very premature to think about anything different."
— Chris Womack, Chairman, President and CEO

Assessment: No extension is being contemplated. The Georgia filing in 2028 is fixed, and it will go to the commission elected this week. Later in the call another questioner asked whether "until at least 2029" meant a year had been added; the reply did not say it had, and we read the remarks as describing the existing plans.

Extension or Increase: What February's Raise Would Mean

The same questioner asked whether the February update would extend the growth range or raise it. The answer was more specific than the prepared remarks.

Q: "As we wrap in the $100 billion and you reevaluate the CAGR, is this more about an extension or an increase to those ranges, just to be clear?"
— Nicholas Campanella, Barclays

A: "We feel good at the moment to be able to raise that upper end of the guidance in those years. And we think that, that's going to put us in a, if you will, a higher starting point for the 7% to 8% path starting in 2028 and well into the next decade."
— David Poroch, Chief Financial Officer

Assessment: The answer is both: higher upper ends for 2027 and 2028, then 7% to 8% compounding from a higher base. That is a meaningful commitment from a management team that has under-promised all year, and it makes February the next major catalyst.

When Alabama's Deferred Revenue Reaches Earnings

A question on the Alabama regulatory liability asked which thresholds release it and how it affects earnings through 2030. Management offered an analogy instead of a schedule.

Q: "Under the proposed earnings deferral that you have there, which of the risk mitigation thresholds are going to kind of determine when those amounts actually enter earnings? And how could that sort of affect earnings between the years through 2030?"
— Julien Dumoulin-Smith, Jefferies

A: "Yes. That's going to be worked out with the commission as those circumstances give rise. The best way I might want to think about that is almost like storm damage accounting, right?"
— David Poroch, Chief Financial Officer

Assessment: The thresholds are not yet defined, and the release could run "well into the next decade." For modeling, assume some Alabama large-load revenue reaches the income statement years after it is billed, and treat it as conservative accounting that also supports the 2029 rate discussion.

Whether a Data Center Pause Could Reach Contracted Load

Three questioners returned to the same risk from different angles: campaign rhetoric about executive orders to pause new data centers in Georgia, and whether that could touch contracts already signed. The answer drew a firm line between existing contracts and future ones.

Q: "I guess just as you think about what's embedded in your base plan, do you see any potential risk to any of those investments or projects moving forward? Or do you feel those are adequately protected by the tariffs that you have in place?"
— Multiple analysts incl. Carly Davenport, Goldman Sachs; Jeremy Tonet, JPMorgan; Sophie Karp, KeyBanc Capital Markets

A: "My sense is I feel pretty confident about what we have in place. There's not been any talk about retroactive or looking backwards. And so they are contracted loads, and so we've got protections there."
— Chris Womack, Chairman, President and CEO

Assessment: The distinction is the right one. Signed contracts are protected by their terms; the pipeline is not. On the call management also said it had already briefed the incoming governor during the campaign. A pause would most likely slow the conversion of the 7 GW late-stage and 75 GW prospective pipeline rather than the 21 GW already signed.

What Sits Inside the $0.30 Fourth Quarter

One question asked directly whether the company was pulling maintenance into the fourth quarter and whether there were other discrete items in the bridge to $0.30. The answer confirmed the intent without a number.

Q: "Just as we think about the fourth quarter guidance you've now provided, are you planning on pulling forward any O&M into the quarter to take advantage of the strong performance this year? Are there any other discrete items that we should be mindful of as you bridge to the $0.30?"
— Carly Davenport, Goldman Sachs

A: "Well, I think a good way to think about that, and I'll use a term from a former CFO, Art Beattie that we go out of our way to try to fix the roof while the sun is shining. And so we've been taking that upon ourselves kind of all year long and really have made some good progress to take care of some things this year to further bolster our confidence around the future years."
— David Poroch, Chief Financial Officer

Assessment: That is a yes. The roughly 25-cent gap between the fourth-quarter estimate and last year's fourth quarter is mostly discretionary spending choices, plus whatever Hurricane Isaias costs. It supports 2027, and it means the fourth quarter will tell us little about run-rate earnings.

How Much of the Collateral Is Actually Posted

A question framed the collateral requirement at about $25 billion by 2030 and asked how much is in hand and in what form. Management declined the first half and answered the second.

Q: "As we -- and kind of put a little bit of a finer point, I guess, as we think of the $25 billion of collateral that's required by 2030, I guess just how much of that is posted now in what form?"
— Nicholas Amicucci, Evercore ISI

A: "We'll stay away from how much is in hand at the moment just for -- but it's several billion dollars."
— David Poroch, Chief Financial Officer

Assessment: In July the figure was about $21 billion across 17 GW. The question's $25 billion against 21 GW keeps the ratio near $1.2 billion per gigawatt, and the company's slides put the average rating at AA-. "Several billion" posted today is consistent with contracts that ramp through 2029, and it is the first indication of how much protection exists now rather than at full ramp.

Whether Large-Load Cash Flow Lowers the 40% Equity Rule

The second half of the opening question asked whether upfront payments and strong cash flow from large-load customers could reduce the equity needed for each dollar of new capital. The answer was no, for now.

Q: "But I'm just wondering, just given the cash flow nature of sort of the large load pipeline, especially if they're making upfront payments, could that sort of help mitigate some of your historical equity needs? So is 40% the right number?"
— Shahriar Pourreza, Wells Fargo

A: "And so our modeling would suggest that with the contracts that we're signing and with the capital that we're going to be deploying, we continue to believe that about 40% remains about the right term for us."
— David Poroch, Chief Financial Officer

Assessment: Clear, and slightly disappointing for the bull case on dilution. Every billion dollars of upside capital still means roughly $400 million of new equity, so the share count grows with the opportunity until the 17% FFO-to-debt target is reached in 2029.

What They're NOT Saying

1. The Georgia real-time pricing review: the commission opened a proceeding during the quarter on how part of the variable demand-driven pricing revenue is assigned to fuel recovery, with a decision due December 15. That tariff supplied most of the quarter's largest bridge line. Neither the release nor the call mentioned the proceeding.
2. How much of the pricing gain is weather: Georgia's rates-and-pricing line swung from a first-half drag to $296 million in the third quarter. The company did not say how much of that is a hot summer and how much is a larger base of demand-priced customers.
3. What is in the $0.30: the size of the fourth-quarter maintenance pull-forward and any other items in the bridge to $0.30 were not quantified.
4. Hurricane Isaias: more than 650,000 customers lost power in early October, with Alabama hit hardest. Neither the restoration cost nor its accounting treatment was given.
5. Collateral in hand: "several billion dollars" of a requirement framed at about $25 billion by 2030. No figure, no schedule.
6. The 4 GW without termination protection: the 10-Q says about 16 of roughly 20 GW of large-load contracts carry minimum duration, termination payments and collateral. The rest have only minimum bills, which matters more now that Georgia will not pass early-termination shortfalls to existing customers.
7. The size of Alabama's deferral: the incremental revenue component behind the new regulatory liability, and the thresholds that release it, were not quantified.
8. The equity arithmetic: the need rose from $1.1 billion to about $2 billion. About $0.7 billion of the $4 billion, the Southern Power gas uprates announced in April, was already inside July's $1.1 billion, so the capital added since July is closer to $3.3 billion. At the stated 40% that would imply about $2.4 billion. The difference was not explained.
9. The minimum bill, for a third consecutive quarter: still no dollar level, rate or share of contracted revenue that is floored.
10. The Southern Power recontracting update: promised for "later this year" in April and July, now described as work that will "stretch out over the next few years."
11. The "rate reduction": the summer reduction cited on the call was a 12.9% decrease in Georgia fuel billings approved in May. Fuel revenue does not affect earnings, and base rates did not move.
12. Georgia's property-tax timing: $116 million of higher property tax from prior-year actualizations, about seven and a half cents, was not mentioned, nor was the $37 million benefit from the same mechanism in the second quarter.
13. Earned return on equity by subsidiary: still not disclosed, which matters more as the 2028 Georgia filing approaches.

Market Reaction

• Pre-print setup: Closed at $93.12 on 4 November, up 6.8% year to date against 12.3% for the S&P 500, up 1.9% over the trailing twelve months and up 9.0% over the trailing thirty days, inside a 52-week closing range of $82.35 to $99.11.
• Election session (4 November, the day before the print): Traded as low as $90.40, down 3.6% intraday, on 18.7 million shares, 3.1 times the 6.0 million thirty-day average, and closed at $93.12, down 0.7%. The S&P 500 rose 1.6% and the utilities sector fund 0.4% the same day.
• Reaction session (5 November, before-open print): Opened at $94.24, up 1.2%; traded between $91.93 and $95.72; closed at $93.34, up 0.2% or $0.22. The close sat about 37% of the way up the day's range.
• Volume: 15.9 million shares against the 6.0 million thirty-day average, 2.7 times normal, after 3.1 times the previous day.
• Market and sector: The S&P 500 fell 4.4% in a broad sell-off in AI-linked stocks, driven partly by post-election fears that governments could force pauses in data center construction; utilities were the strongest sector on the day and Treasury yields fell. The utilities sector fund fell 0.7%. Duke rose 0.4%; American Electric Power fell 1.9%, Dominion 0.2% and NextEra 0.2%.

Relative strength on a risk-off tape. A 0.2% gain looks like indifference to a twenty-nine cent beat against the company's own estimate. On a day the index fell 4.4% it was 4.6 percentage points of outperformance, and about one point better than the utilities sector. Southern did what a defensive large-cap utility is supposed to do on a day like this, and the print gave it a reason to do it a little better than its peers.

The same open, a different fade. For the third quarter in a row the stock opened up between 1.1% and 1.2% on the release. In April it then rose through the call to close up 3.4%; in July it faded to close down 1.8%. This time it reached up 2.8% intraday before giving back most of the gain into the close, which on a day like this says as much about the tape as about the call. The call added the February raise and the $100 billion framing. It did not add anything about the pricing review.

The election was traded first. The heavier session was the day before the print. On the morning after the Georgia results the stock fell as much as 3.6% on three times normal volume before recovering most of it. Across the two sessions the stock fell 0.4%, in line with the utilities sector's 0.3% and well ahead of the index's 2.8% decline. Read together, the market marked down the policy risk on Wednesday and marked part of it back up on the earnings on Thursday, ending the two days in line with the sector. We weigh the two the other way: for the next twelve months we think the earnings evidence is the more durable of the two.

Street Perspective

Debate: Is a Heat-and-Pricing Beat a Quality Beat?

Bull view: Operating income rose 15.7% after two flat quarters, the construction allowance and tax rows shrank to seven cents combined, weather-normal sales grew 3.5% with every class contributing, and the company is guiding the year to $4.70 with a raise to follow. A beat built on revenue is exactly what the skeptics asked for.

Bear view: Weather and the variable-pricing line together are thirty-six cents, more than the entire increase. One will not repeat and the other is under commission review. At normal weather the year is about $4.53, and the fourth-quarter estimate of $0.30 shows management already spending the windfall.

Our take: The bull has the better of the trend and the bear has the better of the single quarter. The composition improved in exactly the place we said it had to, revenue per kilowatt-hour and operating income, and weather-normal volumes broadened. Underlying growth near 7% is a good result for a utility of this size. How much of the pricing gain is durable is a December question, and we have pre-stated what an adverse answer would do to our rating.

Debate: Does the Georgia Election Change the Investment Case?

Bull view: Signed contracts are protected, base rates are frozen until 2029, and Georgia has already committed close to $3 billion of benefits to existing customers funded by large loads. Affordability was the top voter concern, and a company that is cutting bills while growing has the strongest possible argument in front of any commission.

Bear view: The commission that will decide the 2028 rate filing, the moment when years of construction finally enter rates, is now different and untested. An incoming governor has talked about pausing new data centers. The variable-pricing review is already under way. The risk lands precisely where the earnings are meant to come from after 2028.

Our take: The risk is real, and it is a 2029-and-beyond risk to the rate reset and to the conversion of the pipeline, not a risk to 2026 or 2027 earnings. We prefer to carry it as an explicit trigger than as an uncalibrated cut to the multiple. Our multiple is unchanged, and the first evidence of commission or executive action against signed contracts would end the upgrade.

Debate: Is the Road to $100 Billion Accretive After 40% Equity?

Bull view: A Georgia procurement near the top of its 2 to 6 GW range, the Southeast Connector pipeline, nuclear uprates paid for by a hyperscaler, and Alabama to follow: the plan could grow by a fifth, almost all of it regulated, and February will reset growth from a higher base.

Bear view: The equity need already rose 82% on about $3.3 billion of capital added since July. At 40% equity, a $15 billion increase implies about $6 billion more stock. Generation costs are running "a bit past" $2 billion per gigawatt, and the shares outstanding rose 5% in a year.

Our take: Regulated capital at authorized returns is accretive even after 40% equity, but slowly, and the dilution is front-loaded while the returns arrive at commercial operation. We would rather own the capital plan than fear it. We do not credit any of the procurement upside in our estimates until the commission certifies it.

Model Update

Valuation framework. We keep the 20 to 21 times multiple we have used since initiation and apply it to a 2027 estimate of about $4.95, for a fair-value range of $99 to $104 and a midpoint of about $101.50, up from about $100. On the same earnings years the stock is cheaper than it was in July: 18.9 times our 2027 estimate at $93.34 against 19.3 times in July at $94.34, and 19.8 times our 2026 estimate against 20.4 times. Against the $93.34 reaction close, the midpoint implies about 8.7% of price return plus a 3.3% dividend yield, a twelve-month total return of about 12%. At the bottom of the range it is about 9%, at the top about 15%.

Our fair value moved by less than the headline numbers might suggest, because we build 2027 from a normal-weather 2026 rather than from $4.70. What changed is the evidence and the price. In July the expected return was about 9% and the earnings composition argued against paying for the franchise. Today the expected return is about 12%, the composition test we set has been met, and the stock is a dollar lower than it was then.

Thesis Scorecard Post-Earnings

The pillars below are the ones established at initiation and graded again in July, with one new risk added this quarter and marked as such. Status tags follow the standing convention: bull pillars run on track, at risk or broken; risk points run contained, emerging or materializing.

Overall: The thesis strengthened on its central question and weakened at its edges. The question since initiation has been whether the load would convert into earned operating income before the rate reset, and this quarter is the first evidence that it can: operating income up 15.7%, revenue per kilowatt-hour up about 7%, weather-normal volumes growing across every class, and the accrual rows shrinking to a third of the increase. Against that, the pricing that did the most work is under review, dilution accelerated, and the commission that will set rates from 2029 changed. Our earnings-quality bear point moves from materializing to emerging, the dilution point moves from contained to emerging, and a new political risk opens at emerging.

Action: Upgrading to Outperform from Hold. In July we wrote that we would upgrade "on a third-quarter print in which non-fuel retail revenue grows again with pricing flat or better and the beat comes from the operating lines of the bridge." Non-fuel retail revenue grew 15.5%, revenue per kilowatt-hour rose about 7.2%, and the operating rows produced forty-seven cents of a thirty-four cent increase. That test is met, and we said in July that meeting the letter of a test while defeating its purpose does not count. Here the purpose, operating income growing on revenue, is met too. The other two July upgrade tests were not met: no procurement outcome has been disclosed, and the stock is at $93.34, not the mid-$80s. None of our July downgrade tests fired: no commission outcome has reopened the freeze, the second half is growing rather than shrinking, and operating income is no longer flat. Fair value rises to about $101.50 on 20 to 21 times a 2027 estimate of about $4.95, for an expected twelve-month total return of about 12%. Conviction is 6, lower than for our Hold, because the upgrade rests partly on a tariff under review. We state the reversal conditions now. We would return to Hold if the December 15 Georgia order reassigns more than a third of the third quarter's $296 million variable-pricing gain to fuel recovery, or applies any change to revenue already booked; that is outside our estimate, which already removes all of the above-normal weather and a third of Georgia's year-to-date pricing gain but assumes no regulatory change. We would return to Hold on any executive or commission action that delays service under signed large-load contracts. These two do not net against a February raise: an adverse December order goes to the quality of the revenue behind this upgrade, and it would reverse the call even if February raises the outlook as signaled. We would move toward Underperform, as in July, on a Georgia commission outcome that reopens the rate freeze, or on a return to flat operating income while the non-GAAP wedge widens.