Constellation’s third-quarter earnings jumped, and its full-year forecast moved higher. The useful question is how much of that improvement comes from the business it owns today, versus the nuclear projects scheduled to contribute in later years. Those are two different calendars, and I’d rather not let them share a calculator.
In its November 9 results, Constellation Energy reported adjusted (non-GAAP) operating earnings of $5.06 per diluted share for the quarter ended September 30, up from $3.04 a year earlier. That’s a 66.4% increase. GAAP earnings were $5.19 per diluted share, compared with $2.97.
There’s real growth in those reported figures. But Constellation completed its Calpine acquisition in January, so the comparison includes a larger business. Treating the entire increase as organic growth, or as the payoff from newly announced AI power agreements, would skip some fairly important homework.
A larger profit pool, spread across more shares
Adjusted operating earnings rose to $1.792 billion from $952 million, an 88.2% increase. The per-share gain was smaller because average diluted shares increased to 354 million from 313 million. Both comparisons cover the same three-month period, but they answer different questions: how much profit the company reported, and how much of it was attributable to each diluted share.
Constellation’s earnings presentation and financial attachments identify Calpine’s contribution, strong commercial performance and higher capacity revenue as quarterly drivers, with lower nuclear production tax credit revenue providing an offset. The release also points to unfavorable interest expense.
The quarterly filing offers another lens: unaudited pro forma results assuming Calpine had been acquired on January 1, 2025. On that basis, Q3 revenue was $9.550 billion versus $9.446 billion, while net income attributable to common shareholders was $1.835 billion versus $1.550 billion. That’s approximately 1.1% revenue growth and 18.4% net-income growth, not a measure of adjusted EPS growth.
These figures include acquisition-accounting adjustments and don’t necessarily represent results that would actually have occurred or future results. Constellation also says integration and consolidated financing make standalone Calpine earnings impracticable to determine. It’s a useful comparison, with a clearly marked limit. The portfolio is still changing, too: Brazos Valley was sold in October, while the five-plant PJM sale remained pending in the filing.
The outlook improved, with a slightly smaller share count
Management raised and narrowed its 2026 adjusted operating earnings forecast to $12.50–$12.80 per share. The previous range was $11.50–$12.50. Comparing midpoints, that’s $12.65 versus $12.00, an increase of 65 cents, or about 5.4%.
The company attributes the revision to commercial outperformance in a volatile market, partly offset by higher operations and maintenance costs associated with that outperformance. The current forecast also assumes 356 million average diluted shares, down from 357 million in the previous forecast.
I’d keep that denominator change beside the earnings explanation. It doesn’t erase the improved outlook, but it does mean the two per-share forecasts aren’t calculated with identical share assumptions. These are management estimates, not results already earned.
Why adjusted earnings were lower than GAAP
“Adjusted” doesn’t automatically mean a bigger number. This quarter’s reconciliation is a good reminder.
Among other items, Constellation removed $127 million of unrealized fair-value gains and a $108 million tax benefit, while adding back $73 million of Calpine merger and integration costs and $39 million of acquired commodity-contract amortization. Those amounts are after tax. The contract amortization reflects noncash purchase-accounting effects from Calpine; the tax adjustment relates to deferred taxes and changes in forecasted apportionment.
These adjustments help explain why the company’s adjusted EPS was below its GAAP EPS. They don’t make either measure a substitute for the other. Integration costs still deserve attention, even when management excludes them from its preferred operating measure.
The nuclear agreements belong on the project calendar
Constellation also highlighted its previously announced Amazon and Google nuclear agreement developments. The Amazon arrangement supports a planned 190-megawatt expansion at Calvert Cliffs, expected during 2030–2032. The Google collaboration supports roughly 890 megawatts of additional nuclear capacity, with the first addition expected in 2028.
The planned expansions depend on regulatory approvals, successful uprates, continued operation and relicensing, and other conditions. Those dates are important context for a September-quarter earnings report. Future output can support a longer-term investment case, but it isn’t electricity those expansions already produced in Q3. Google’s separate 15-year, 2,700-megawatt financial energy load-following agreement also shouldn’t be counted as 2,700 megawatts of newly built nuclear generation.
The financing deserves the same care. Constellation signed a conditional Department of Energy commitment supporting a loan guarantee of up to $5 billion. Execution of a loan agreement remains a condition. A conditional financing commitment is useful progress; it isn’t $5 billion of cash already received.
What interests me here is the combination of stronger current earnings and a substantial program of planned nuclear upgrades. The next disclosures can help show how much commercial strength persists, how integration costs develop and when planned upgrades become operating assets. For now, the earnings improvement is measurable. The longer-term project returns still have work ahead of them.