NorthWestern Energy delivered a sharp earnings miss for the third quarter. Adjusted EPS of $0.58 fell 20.5% short of the $0.73 consensus and declined 26.6% from a year ago. Revenue was the bright spot on paper: $423.8 million topped the $411.24 million estimate by 3.1% and grew 9.5% year over year. The more important development is that management reaffirmed its 2026 guidance range of $3.68 to $3.83 but now expects to land in the lower half. At the first and second quarters, that range had been described as in line with expectations. The central tension is clear. The standalone earnings story is softening, yet the stock sits near a 52-week high because the Black Hills merger is down to its final approval.
The quality of the top-line beat deserves scrutiny. Utility margin rose 5.5% to $316.5 million. However, only $9.6 million of that increase actually flowed to earnings, because much of the remainder was tracker revenue offset by matching expenses. That $9.6 million included $5.3 million of margin from the Puget Colstrip interests. Expense growth overwhelmed it. Operating and maintenance expense jumped 32.4% to $84.9 million, driven by $10.5 million of higher generation maintenance, of which $11.5 million was tied to the new Colstrip stakes. Depreciation and interest expense also rose. GAAP EPS fell to $0.50 from $0.62, and that was with help from a lower 13.2% effective tax rate and $4.4 million less in merger costs. The trajectory is also worse than the prior quarter suggested. Second-quarter adjusted EPS had grown to $0.50 from $0.40, while year-to-date adjusted EPS now trails 2025 at $2.39 versus $2.41.
Colstrip is the core of the problem. The incremental Avista and Puget interests cost $6.1 million, or about $0.10 per share, in the quarter and $16.4 million, or $0.27, year to date. Pacific Northwest power prices were too weak for the interim PCCAM waiver to recover any of the Avista operating costs. At the second quarter, management had hoped July pricing would help earn back some of that drag. Investors should also note how the adjustments work. The company excludes Avista costs from adjusted EPS, but it does not exclude the Puget costs. According to the call, year-to-date adjusted EPS would be another $0.12 lower if Colstrip were fully stripped out. Lower usage per customer added to the pressure, and the PCCAM cost-sharing decision has slipped to the first half of 2027. Management's fix is a Montana rate review to be filed by the end of 2026, intended to put Avista Colstrip costs into base rates and remove the drag in 2027.
The data center narrative also lost ground. Sabey terminated its development agreement over land issues, cutting the count from three to two. Atlas is now unlikely to sign an electric service agreement by year-end, whereas the second-quarter outlook expected both Quantica and Atlas agreements. The remaining agreements still point to 75 MW of demand beginning in late 2027, potentially rising to roughly 1,250 MW by 2030. Quantica's ramp could start earlier than early 2029, and its agreement is still targeted for year-end. The early-stage request queue grew from 14 to 16, with more South Dakota interest. Even so, the Large New Load tariff docket is moving slowly through heavy discovery, and a Montana legislative proposal with retroactive language adds uncertainty. Bears have legitimate ammunition here: the pipeline got thinner even as the queue got longer.
The market setup is more complicated than the miss implies. Shares closed at $82.00, up 17.1% from the $70.03 open after the July report. That leaves the stock 16.1% above its 200-day moving average and about 1.3% below the $83.04 inter-earnings and 52-week high set on report day. The rally from the late-September low of $65.84 was steep. The price action suggests investors are valuing NWE as a pending component of Bright Horizon rather than on standalone quarterly results. Each NWE share converts into 0.98 Black Hills shares, the Montana commission's order is expected around November 12, and the combined company is framed for 5% to 7% EPS growth versus 4% to 6% standalone. Investor sentiment remained positive but weakened modestly, slipping from 0.14 to 0.10. Price and AVWAP trend readings stayed positive.
That framing cuts both ways. Management itself noted that standalone guidance will no longer apply once the merger closes. Pro forma 2027 guidance is not expected until the February call, and a capex refresh is due in early 2027. Equity issuance to fund South Dakota generation, including the roughly $470 million Hub City project, is planned beginning in 2027. A delayed or conditioned Montana approval would force the market to refocus on a standalone profile that has clearly deteriorated.
The bottom line is that NorthWestern's quarter weakened the standalone case on nearly every front: a 20.5% earnings miss, a lower-half guidance bias, uncovered Colstrip costs, and a shrinking roster of data center partners. The stock, priced near its highs, is effectively a bet on a clean Montana merger approval, and the fundamentals offer little cushion if that final step stumbles.