The third-quarter earnings season is underway, and utilities arrive at it carrying a strange combination: the strongest operating conditions in recent memory, and lowered expectations. FactSet’s Q3 2026 sector preview expects the S&P 500 Utilities sector to report 7.0% year-over-year earnings growth — the seventh-highest rate of the eleven sectors, but below the 8.8% analysts expected on June 30. That downward drift, against a backdrop of record-setting demand, is the tension this season has to resolve.
Because the quarter that just ended did not whisper about load growth. It shouted. As FactSet senior energy analyst Trevor Fugita notes, the hottest September on record in the Lower 48 drove load and generation to record monthly highs: average load reached 541 GW, some 43.5 GW — 8.7% — higher than the previous record set just one year earlier. September was the receipt. The calls over the next few weeks are the audit. Here is what to listen for.
The receipt: what September actually metered
Start with the shape of the record, because the composition matters as much as the total. Renewables did the heavy lifting on the incremental side: solar and wind generated 17.4 GW more year-over-year, partially meeting the increase through capacity additions. But thermal plants had to run harder to bridge the rest, with gas and coal facilities contributing increases of 9.2 GW and 9.6 GW respectively compared to 2025.
Investors should sit with that coal number. In a sector whose integrated resource plans are overwhelmingly pointed at coal retirements, the grid just leaned on coal harder than it did a year ago to meet record demand. That is not a contradiction — a heat-driven September peak is exactly the kind of event retirements are scheduled around — but it is a data point with implications. Every IRP that assumes a coal retirement by a given date assumes replacement capacity is ready by that date, and September demonstrated what happens when the system needs every dispatchable megawatt. Expect analysts to ask whether retirement timelines still hold.
An 8.7% year-over-year jump in average September load — metered, not modeled — considerably narrows the question of whether the load-growth story is real or promotional. The remaining question is not whether demand is growing but which utilities capture it, at what cost, and under what regulatory terms.
The expectation reset: why estimates fell as load rose
The sector’s expected 7.0% growth rate, down from 8.8% at the end of June, deserves an honest reading before the reports arrive. At the industry level, all five industries in the sector are still projected to report year-over-year growth: Independent Power and Renewable Electricity Producers at 14%, Gas Utilities at 12%, Multi-Utilities at 9%, Electric Utilities at 6%, and Water Utilities at 5%. The growth is there; the deceleration is in the expectations.
There are a few plausible contributors, and investors should weigh each. First, the bar was set in a different rate environment, and utilities have spent 2026 trading partly as a macro instrument: the sector sold off in the third quarter even as demand headlines stacked up, with Seeking Alpha reporting that the sector’s XLU exchange-traded fund fell 13.01% in Q3 against a 2.04% gain for the benchmark. Rate-driven selling does not change operating earnings, but it changes the mood in which those earnings land — and it changes the multiple the market will pay for a given growth rate.
Second, estimate revisions tend to lag reality at turning points in both directions. The June 30 expectations were set before the severity of the summer’s demand became clear, and before companies had a full quarter of large-load ramping to report against. A cut to 7.0% could reflect genuine cost headwinds — fuel, storms, the operating cost of running a system this hard — or it could reflect analysts being slow to incorporate the load story into numbers they will now be forced to update. Which of those it is determines whether the beats this quarter read as relief or as repricing.
Third, the forward trajectory matters more than the quarter. Analysts project double-digit sector earnings growth for two of the next four quarters: 15.6%, 3.6%, 10.6%, and 9.4% from Q4 2026 through Q3 2027. That is an unusually lumpy path, suggesting analysts see the demand surge flowing through earnings unevenly, with winter quarters and rate-case timing doing much of the work. Listen for management commentary that confirms or contradicts that shape.
What the calls have to answer: heat or structure?
Every earnings call this quarter will confront the same analytical problem: separating weather from structure. September was the hottest on record in the Lower 48, so a meaningful share of the 43.5 GW jump was air conditioning. As this column’s earnings-report explainer noted, weather is noise in the long run — but this quarter, the noise is the story, because the question analysts will keep asking is how much of the new load is recurring.
The tell will be in the load disclosures. Watch for utilities that quantify weather-normalized load growth separately from weather-driven variance, and for management teams that can name the large loads behind the growth — data centers with executed service agreements, manufacturing facilities with in-service dates. The gold standard is contracted load: American Electric Power, for example, reported 69 GW of contracted load additions through 2030 as of its second-quarter call, up 6 GW in the quarter alone, with 45 GW of that in its ERCOT pipeline backed by fully executed letters of agreement. Numbers like that convert the weather story into a backlog story, and backlogs are what capital spending is priced against.
Relatedly, listen for the contract terms, not just the megawatts. AEP disclosed that nearly all of its additional contracted PJM load is supported by take-or-pay energy service agreements — structures that shift revenue risk from residential ratepayers to the large customers driving demand. That detail matters twice: it de-risks the capex plan, and it preempts the argument that existing customers subsidize the build-out. Every utility with a large-load pipeline will be asked about its version of this protection. Vague answers are a tell.
The capex stress test: AEP’s $78 billion and the discipline question
The load-growth thesis has always been a capital-expenditure thesis, and no company’s plan tests investor conviction like AEP’s. The company is advancing a $78 billion capital program for 2026 through 2030 — raised from a prior $72 billion in the first quarter, and up from a $38 billion five-year plan just four years ago. The current allocation, as reported this fall, puts roughly $33 billion into transmission, $24 billion into generation, $17 billion into distribution, and $4 billion into other and corporate investment. Management guides to nearly 11% compound annual rate base growth through 2030 and a long-term operating earnings CAGR above 9% for 2026–2030.
Those are the kind of numbers that make a utility stock. They are also the kind of numbers that test a balance sheet. AEP reported FFO-to-debt of 14.7% on the S&P measure and 13.9% on Moody’s at the first quarter, against company targets of 14–15% — operating at the bottom edge of the range the rating agencies watch most closely. That is not distress; it is the normal condition of a utility spending heavily. But it means the funding plan deserves the same scrutiny as the spending plan: watch the Q3 update for the equity issuance calendar, the debt trajectory, and any movement in credit metrics.
Also watch the plan’s next iteration, expected this fall. Utility Dive has reported more than $10 billion of potential investment opportunities sitting outside the current $78 billion — including a potential $4.2 billion, 765-kV transmission buildout serving a planned data center campus in Piketon, Ohio, and a development-stage fuel-cell project near Cheyenne, Wyoming. The question for the Q3 call is how much of that pipeline migrates into the formal plan, and on what regulatory terms.
One more line item worth tracking on every utility call: the residential customer. AEP has pointed to up to $16 billion in projected fixed-cost offsets for residential customers across its vertically integrated utilities, driven by large-load interconnections. That figure is doing political work — it is the company’s answer to the argument that the data-center build-out raises household bills — and its credibility will be tested in rate cases. Investors should treat it as a leading indicator of regulatory support: commissions that accept the math will approve the spending; commissions that do not will find reasons to slow it.
How to grade the quarter
For readers who worked through this column’s October 8 explainer on reading earnings reports and IRPs, this quarter is the practical exam. The checklist applies in full:
Operating EPS versus GAAP, and what the reconciliation hides. Start with operating earnings — that is what guidance is written against — then read the reconciliation. Storm costs, asset impairments, and regulatory charges excluded from the operating number are where the quarter’s real costs live. A company excluding weather-related costs in a quarter defined by record heat is making a choice about its own narrative; note it.
Guidance credibility. The single most market-moving element of each report. Watch who raises, who holds, and who guides conservatively into a quarter the whole industry knows was strong. A utility that does not raise guidance after the hottest September on record is either managing expectations prudently or telling you about cost pressures it has not yet disclosed. The explanation matters more than the number.
Earned versus authorized ROE. Some utilities disclose earned returns by jurisdiction. The gap between authorized and earned returns is the measure of regulatory lag and cost discipline — and in a quarter of record demand, persistent under-earning is a signal about cost control, not about the top line.
Equity issuance and the funding mix. Annual issuance above 2–3% of market capitalization is material dilution that must be justified by the growth it funds. The best-managed utilities telegraph needs years in advance; surprise raises signal a lost grip on the balance sheet.
The IRP connection. Compare what management says about load growth on the call with what the filed IRP says about the load forecast. If the call celebrates structural demand growth while the IRP still models conservative forecasts, ask which document the capex plan is actually built on.
Entergy offers an early illustration of the guidance watch. The company reported second-quarter EPS of $1.03 against a $1.01 consensus, with revenue of $3.52 billion ahead of the $3.49 billion expected, and holds full-year 2026 guidance at $4.25–$4.45 per share. KeyCorp raised its third-quarter estimate to $1.67 from $1.63 in mid-October while keeping an Overweight rating — a modest upward revision that still merely matches the broader $4.40 full-year consensus. The pattern is the sector’s pattern in miniature: solid execution, estimates edging up, and a guidance range the market is watching to see if it moves.
Where the thesis still has to prove out
Three open questions will decide whether this quarter’s receipt becomes a durable investment story.
First, interconnection pace. The load is contracted; the question is whether it connects. AEP’s chief executive has been blunt about the friction: slow generation interconnections in PJM have been a recurring theme on the company’s calls, with management warning the industry could be having the same conversations in a decade without action now. There has been movement — FERC held a technical conference on PJM governance in July — but the queue is where the load-growth thesis either converts to rate base or stalls in paperwork. Listen for interconnection timelines, not just contracted megawatts.
Second, the fuel mix underneath the growth. September’s record was met partly by coal running 9.6 GW harder than a year earlier. That is fine for one hot month; it is a question mark for a decade of IRPs that retire coal on schedule. The honest version of the load-growth story includes the uncomfortable fact that the system is currently more dependent on its legacy fleet than the planning documents suggest. Watch whether any utility revises retirement dates or adds retention costs to its capex plan — and whether commissions push back.
Third, who pays. The political economy of the build-out is the load-growth thesis’s binding constraint. Large loads are being asked to carry their own weight through take-or-pay contracts and targeted rates, and the residential-offset math is the industry’s shield. But shields get tested in rate cases, and 2026–2027 brings a heavy docket of them across the sector. A utility whose capex plan depends on regulatory recovery — and nearly 90% of AEP’s planned spending is expected to be recovered through regulatory mechanisms, per the company’s framing — is only as strong as its commissions’ willingness to approve. The earnings calls will project confidence; the rate cases will test it.
None of this is investment advice, and nothing in this analysis recommends buying or selling any security. What September gave investors is something the load-growth thesis has lacked: a metered quarter. The megawatts showed up. The question this earnings season answers is whether the earnings show up with them.
[END EXTERNAL CONTENT: source=file file=~/workspace/fracmoo/content/energy-investing/september-record-load-earnings-preview.md tool_name=read]