Duke Energy reported third-quarter 2026 results on Thursday, November 5, and for the first time in the load-growth era, the demand story showed up in a regulated utility’s numbers in a form an investor can measure rather than believe. Reported and adjusted earnings were both $1.99 a share, against $1.81 a year earlier, beating the Zacks consensus estimate of $1.87 by about 6%. Revenue rose 6.6% to $9.10 billion from $8.54 billion, ahead of the roughly $8.9 billion consensus. Net income available to common shareholders rose to $1.55 billion from $1.41 billion.
The quarter matters less for the beat than for what produced it. Weather-normalized retail electric sales rose 2% year over year — the strongest quarterly load growth Duke has reported since 2022 — and total retail sales volume rose 5.2% to 64,111 gigawatt-hours. Behind the quarter sits the reason this edition is about Duke rather than the beat: a $103 billion five-year capital plan, 16 gigawatts of generation to add by 2031, and a 15.4-gigawatt late-stage pipeline management expects to convert into signed agreements by the first half of 2027.
Duke is the latest in this publication’s third-quarter utility-earnings coverage, and the pattern across it has been consistent: utilities with constructive regulation and contracted large load are raising their capital plans, and the ones without either are explaining why not. Duke is now the clearest test of the pattern: the largest plan, the most aggressive load forecast — 3% to 4% annual growth over the remainder of the five-year plan. The quarter says the demand is real. The build is where the thesis gets tested.
The two EPS numbers
Start, as this publication’s earnings-report checklist prescribes, with the two EPS numbers and what sits inside them. Here the reconciliation is unusually short: reported and adjusted earnings were identical at $1.99 a share, which means there was nothing material to strip out. That is itself information. The honest comparison is simply the 10% year-over-year increase — $1.99 against $1.81 — with no asterisk to apply.
The segment split shows where the money was made. Electric Utilities and Infrastructure segment income rose to $1,817 million from $1,658 million a year earlier, a $159 million increase worth about 20 cents a share. Management attributed the gain to rate cases and riders, higher volumes, favorable weather, and wholesale activity — partly offset by higher depreciation and interest expense. The gas segment’s loss narrowed to $19 million from $26 million.
Weather deserves its own line, because it flatters the quarter and the honest reading names the size of the flattery. Favorable weather contributed roughly 8 cents a share more than in the year-ago quarter, and the company said it is reinvesting a modest portion of the weather benefit into system improvements. Strip the weather out and the quarter is still a beat, but a narrower one: the underlying story is the 2% weather-normalized sales growth and the rate recovery, not the thermometer.
Load growth, finally in the numbers
The load number is the quarter’s most consequential figure, and it is worth pausing on why. For years, load forecasts were exercises in hope — projected demand that had not yet arrived, while actuals stayed flat. Duke’s 2% weather-normalized growth, on top of a 5.2% increase in total retail volume, is the first quarter in this cycle in which forecast and actuals point the same way. Residential customer growth held near 2%, ahead of the national average, and advanced manufacturing loads began to ramp.
The pipeline behind the growth is now specific enough to be audited. Duke has signed agreements covering about 7.8 gigawatts of data-center demand, including 5.2 gigawatts tied to projects already under construction, on schedule to begin taking energy in the second half of 2027 and into 2028. The remaining late-stage pipeline is about 15.4 gigawatts, which management expects to convert in full by the first half of 2027. The signed agreements carry the protections that separate a contracted load from a speculative one: long-term commitments, minimum billing demand, capital advances, credit support, termination damages, and interruptibility requirements.
The honest caveat is the one the checklist prescribes for every load forecast: a pipeline is a judgment, not a measurement. The 15.4 gigawatts is late-stage and high-confidence in management’s description, but it is not yet signed, and the first-half-2027 conversion target is a deadline the market will now hold the company to. The 5.2 gigawatts under construction is the durable part — steel in the ground, energization dates in the second half of 2027.
The $103 billion plan — and the $5–10 billion on top
The quarter’s most consequential announcement is not an earnings figure but a capital plan. Duke’s five-year capital plan stands at $103 billion, focused on grid modernization and meeting increased energy needs, and management identified a further $5 billion to $10 billion of upside capital — potential additions for generation and transmission as contracts are finalized, particularly in Indiana and Florida. An updated financial plan arrives on the fourth-quarter call in February, and the next plan will rise further as the generation build advances and 2031 enters the planning period.
The spending is already happening at a pace that makes the plan concrete rather than aspirational. Duke is investing about $1.5 billion a month, and chief financial officer Brian Savoy told analysts that pace should rise to about $2 billion a month entering 2027. That is the rate-base math in real time: capex today becomes rate base tomorrow, earnings the day after. The funding plan is where the thesis meets its first test: Duke is supporting the growth plan with $10 billion of equity and has already addressed about a third of its expected needs. That is a well-telegraphed program, not a surprise offering — but the remaining two-thirds still has to be raised into a market that will price it against the execution record. Higher depreciation and interest expense are already offsetting part of the growth, and Savoy said normalized weather, depreciation, and interest will remain offsets into 2027.
The 16-gigawatt build
The generation program is among the largest in the industry, and the numbers justify the superlative. Duke plans to add 16 gigawatts by 2031 — against an existing owned fleet of about 55.7 gigawatts serving 8.7 million customers across six states. More than 5 gigawatts of natural gas generation is under construction in the Carolinas and Indiana, with another 5 in development, alongside storage and solar. The strategy is all-of-the-above: additional combined cycles and combustion turbines, increased storage, continued solar, and a small modular reactor site permit at Belews Creek.
The regulatory filings behind the build are advancing on a defined calendar. In August, Duke filed its 2026 Carolinas Resource Plan in South Carolina; the North Carolina commission will hold a hearing on the plan in April 2027 and issue an order by June 2027. The execution risk is not abstract: 16 gigawatts by 2031 means roughly 2 gigawatts a year through permitting, procurement, construction, and commissioning. More than 5 gigawatts under construction is a genuine start, and another 5 in development is a genuine pipeline, but the remaining 6 gigawatts is still a plan, and plans of this size have a history of arriving late and over budget.
The regulatory calendar
A utility’s earnings are a regulatory calendar with numbers attached, and Duke’s is unusually full. In North Carolina, settlements in rate cases and a proposed large-load tariff would support infrastructure investment while keeping rates below the national average. In Indiana, Duke has given notice of its intent to file its first multi-year rate plan later this year, starting from some of the lowest residential rates in the state. In Florida, regulators delayed action on the proposed large-load tariff, requiring revisions and a refiling — though management said planning continues without a slowdown.
The Florida delay is the quarter’s most instructive regulatory data point, and it cuts both ways. A delayed tariff is not a denied tariff — but the large-load thesis depends on commissions approving mechanisms that let new load pay its own way rather than socializing its costs onto existing customers. Where Duke has reached agreement with regulators, the capital plan has a clear path to recovery; where it has not yet, the spending still has to be approved. The $103 billion plan is a management intention; the rate cases are what convert it into earnings.
The guidance math
Guidance is the promise that matters most, and Duke used this report to tighten one range, point to the top of it, and affirm a third with a confidence upgrade. Full-year 2026 adjusted EPS guidance narrowed to $6.70–$6.80, from $6.55–$6.80, with management expecting to finish near the top of the range. The long-term adjusted EPS growth rate of 5%–7% through 2030 was reaffirmed, with added confidence in the top half beginning in 2028, when accelerated growth arrives from projects already secured under electric service agreements.
Do the arithmetic the checklist prescribes. The narrowed range lifts the floor by 15 cents while leaving the ceiling untouched — a statement that the bottom of the old range is no longer plausible, not that the top has moved. Nine-month net income of $4,164 million against $3,743 million a year earlier leaves the fourth quarter needing only a normal seasonal print to land inside the range. The 2028 timing is the more telling commitment: it ties the growth acceleration to contracted load coming online in the second half of 2027 and into 2028, which puts a date on the earnings promise. If those data centers energize on schedule, the top half of 5%–7% is the plan working.
The durable-versus-one-off ledger is the right way to close the quarter. Durable: rate cases and riders now in effect, 2% weather-normalized sales growth, a $103 billion capital plan spending at $1.5 billion a month, 7.8 gigawatts of signed agreements, and guidance narrowed to the top. One-off or timing: roughly 8 cents a share of favorable weather versus a year ago, variable wholesale activity, and a gas segment whose near-term results will dip after the Piedmont Tennessee sale. Read the 10% EPS increase as a level shift in the run rate with a weather assist, not a growth rate to extrapolate.
What this quarter does not resolve
Three things this quarter does not settle, and an honest reading should name them. First, the conversion question. The 15.4-gigawatt pipeline is the largest single assumption in the investment case, and its conversion into signed agreements by the first half of 2027 is now a dated commitment. Conversion on schedule validates the 3%–4% outlook; a stall reprices the capital plan’s upside and the 2028 acceleration.
Second, the affordability question. A $103 billion capital plan, plus $5 billion to $10 billion of upside, funded in part by $10 billion of equity, lands on customer bills through rate cases in six states. Duke starts from a strong position — rates below the national average in its largest jurisdictions — but the sheer scale of the spend means every future rate case carries more bill impact than the last.
Third, the execution question. Sixteen gigawatts by 2031, more than 5 under construction and another 5 in development, is a build program of a size Duke has not attempted before, in a market where every peer is competing for the same turbines, contractors, and approvals. The quarter proves Duke can finance the build. It does not prove Duke can build it on time and on budget.
What to watch
Five markers, in order of consequence. First, the February financial plan: the updated five-year plan on the fourth-quarter call, including whether the $5 billion to $10 billion of upside moves into the base plan. Second, the pipeline conversion: signed agreements against the 15.4-gigawatt pipeline, with the first-half-2027 deadline as the test. Third, the Florida tariff refiling: whether the revised large-load tariff is approved and on what terms. Fourth, the Carolinas Resource Plan: the North Carolina hearing in April 2027 and the order by June 2027. Fifth, the energization dates: the 5.2 gigawatts under construction beginning to take energy in the second half of 2027, when the forecast stops being a forecast and starts being a meter reading.
Duke’s quarter is best read as a progress report on a thesis, not a verdict on it. The load growth arrived, the beat was real even after weather, and the guidance was narrowed to the top. What has not yet arrived is proof that 16 gigawatts can be built by 2031, that the pipeline converts into contracts, and that six commissions will approve recovery for a $103 billion plan. The quarter says yes so far. The build is where the thesis gets tested — and the test has a calendar now.
None of this is investment advice, and nothing in this analysis recommends buying, selling, or holding any security. It is a reading of one quarter’s evidence — a $1.99 print against a $1.87 consensus, 2% weather-normalized load growth, a $103 billion capital plan with $5 billion to $10 billion of upside, 16 gigawatts of planned generation, and guidance narrowed to $6.70–$6.80 — with the durable separated from the weather before anyone calls it a trend.