BCE topped expectations in the third quarter, reporting earnings of $0.53 per share against a $0.49 consensus, an 8.2% beat, on revenue of $4.418 billion versus the $4.34 billion estimate. Earnings still fell 7.0% year over year and revenue grew just 0.6%. Those normalized figures appear to be on a U.S.-dollar basis. In its own Canadian-dollar reporting, BCE posted adjusted EPS of C$0.74, down 6.3%, on revenue of C$6.18 billion, up 2.2%. The gap between the two growth rates is likely currency translation rather than a difference in operations. Either way, this was a beat against lowered expectations rather than a return to earnings growth. Management reaffirmed 2026 guidance, which itself calls for adjusted EPS to fall 5% to 11%.

The quality of the quarter is mixed. GAAP net earnings dropped 84.6% to C$701 million, but that reflects last year's C$5.2 billion gain on the MLSE stake sale and says little about operations. The more telling numbers:

• Adjusted EBITDA rose only 0.9%, and margin slipped 60 basis points to 45.1% on higher costs and a heavier mix of low-margin equipment sales.
• The Canadian core, Bell CTS Canada, saw revenue dip 0.2% and EBITDA fall 0.8%, with service revenue down 1.3%. Excluding Ziply Fiber, consolidated EBITDA was flat.
• Operating cash flow jumped 15.9%, but capital spending rose 49.3% to C$1.33 billion for AI Fabric data centres and Ziply fibre. Capital intensity climbed to 21.5% and free cash flow fell 14.9% to C$854 million.

The central tension is that BCE's growth engines are clearly accelerating while its largest profit pool, Canadian wireless, is still going backward on reported metrics.

The growth engines are moving in the right direction:

• AI-powered solutions revenue grew 41%, up from 29% in Q2.
• Bell Business Markets revenue rose 9.2%, versus roughly 3.4% underlying growth last quarter.
• Bell Media delivered a second strong quarter. Revenue rose 9.7% and EBITDA 8%, and digital revenue grew 20% versus 6% in Q2, helped by the FIFA World Cup.
• Postpaid churn improved 7 basis points to 1.06%, the best third-quarter level in four years.

The pressure points are just as visible:

• Wireless ARPU fell 1.7%, a sharp widening from roughly -0.2% in Q2. The causes were the CRTC's ban on connection fees, lower overage and roaming revenue, and a lower-priced government contract.
• Postpaid net adds slowed to about 24,000 from 41,600 in Q2, and the government contract flattered loadings for a second straight quarter.
• Crave subscriber growth is decelerating, from 25% to 23% to 21% over the year.
• Ziply fibre net adds eased to about 8,600.

The call added several forward-looking developments.

The most important new commitment is on wireless. Management now explicitly expects wireless service revenue and ARPU to return to growth by mid-2027, "perhaps earlier." On the Q2 call, CEO Mirko Bibic declined to predict timing. Supporting that view, new consumer postpaid billed rates are running about 11% higher year over year.

On the balance sheet, La Caisse will buy 49.9% of a new tower company for C$2.45 billion, implying roughly a C$4.9 billion valuation for the business while BCE keeps control. Combined with net debt falling about C$680 million year to date to C$39.6 billion, the deal gives credibility to the unchanged 3.5x leverage target for the end of 2027.

On AI infrastructure, management framed Bell AI Fabric as a business generating more than C$500 million of EBITDA with contracts in hand. A non-binding Saskatchewan MOU could add up to 900 MW of capacity.

Bears have legitimate ammunition:

• Some Saskatchewan capex is slipping into 2027, and Q4 will be capex-heavy.
• The Ziply target of 3 million passings by 2028 has softened to "get as close as we can."
• The Northwestel sale will likely take a different form.
• AI Fabric carries tenant credit risk from neocloud customers and potential community opposition.
• Quebecor continues to take an outsized share of industry net adds.

Investor sentiment improved modestly ahead of the report, moving from slightly negative (-0.04) to mildly positive (+0.11). That is a meaningful shift in tone but hardly euphoric. Price action tells a harsher story. Shares closed at $21.11 on November 4, the day before the report. That was:

• 4.7% below the opening price after the August report;
• nearly 10% below the 200-day moving average of $23.39;
• about 12% below the August 24 high of $23.94.

The stock set a 52-week low of $18.68 on October 9 and has since rebounded about 13%. Most of the Earnings Whispers trend signals remain negative, including price, sequential growth and AVWAP. That fits a stock still recovering from a damaging autumn rather than one already pricing in a turnaround.

The bottom line is that BCE delivered a credible reaffirmation quarter. The earnings and revenue beat, faster enterprise AI and media growth, lower churn and the tower monetization all strengthen the medium-term story. But the earnings line is still shrinking, margins and free cash flow are under pressure from the capex build, and the wireless ARPU recovery remains a management forecast rather than a reported result. Until reported wireless metrics begin to turn, the stock is likely to keep trading on balance-sheet progress and 2027 promises rather than on current earnings momentum.