If you own BCE (TSX: BCE) primarily for its dividend, I think there’s some good news: after last year’s painful reset, the current payout looks considerably safer.
BCE cut its annualized common dividend from $3.99 to $1.75 per share in 2025, a reduction of roughly 56%. Management simultaneously established a long-term target of paying out 40% to 55% of free cash flow. The transition initially left the 2025 payout ratio at approximately 64%, but the new dividend gives BCE considerably more breathing room than it had previously.
And management hasn’t blinked since. On August 5, BCE’s board declared another $0.4375 quarterly dividend, payable October 15 to shareholders of record on September 15. That keeps the annualized payout at $1.75.
Yahoo Finance recently showed BCE yielding approximately 5.9%. It also reported about $2.68 billion in trailing levered free cash flow. So, I’m not particularly worried about another dividend cut right now. My concern is the stock itself.
Source: Getty Images
Can BCE actually deleverage?
This is the first thing I would watch. BCE’s balance sheet remains heavily leveraged. Yahoo Finance recently reported approximately $1.38 billion in cash and a total debt-to-equity ratio of roughly 182%.
Cutting the dividend was therefore necessary, in my view. Instead of distributing nearly $4 per share annually, BCE can retain substantially more cash and direct it toward debt reduction. I think management can make progress here.
Telecommunications remains a relatively predictable business. Customers continue paying for wireless and internet service, BCE generates billions in free cash flow, and management now has a much more manageable dividend obligation.
I’m less convinced by the AI buildout
The problem is that BCE is simultaneously attempting something much more ambitious. Bell wants to become a major Canadian artificial intelligence infrastructure provider through Bell AI Fabric.
The centrepiece is a planned 300-megawatt data centre in Saskatchewan. BCE expects the project to require approximately $1.7 billion of incremental capital expenditures, with roughly $1.3 billion occurring during 2026 alone. The first data hall is expected to come online during the first half of 2027. That spending has a major impact on cash flow.
Before incorporating the Saskatchewan project, BCE expected $3.3 billion to $3.5 billion of free cash flow in 2026. After accounting for the additional investment, guidance fell to just $2.1 billion to $2.3 billion, representing a 28% to 34% year-over-year decline. Capital intensity is also expected to jump to approximately 20%.
This is where I become skeptical. The U.S. hyperscalers investing heavily in AI infrastructure generate enormous amounts of free cash flow and generally have much stronger balance sheets. BCE is trying to fund a major new capital-intensive growth initiative while simultaneously repairing leverage and maintaining a $1.75 annual dividend.
Maybe it works. BCE has already secured customers and partners for its AI infrastructure, so this isn’t purely speculative. But I think management is taking on an ambitious execution challenge at a time when financial flexibility remains constrained.
I’d separate the dividend from the stock
That’s the distinction I’d make if I already owned BCE. I think the dividend itself is reasonably safe at $0.4375 quarterly. Management already made the difficult decision to cut it substantially; the new payout is far more manageable, and the board just reaffirmed that level in August. But a safe dividend does not guarantee a good total return.
BCE still needs to reduce leverage while navigating a mature Canadian telecom market and spending heavily on an AI infrastructure strategy whose eventual returns remain uncertain.
I have reasonable confidence in management’s ability to make progress on the first objective. I’m much less confident that BCE can turn its AI ambitions into sufficiently attractive returns on all the capital being deployed.
So, if you’re holding BCE for its 5.9% dividend yield, I don’t see an imminent reason to panic about the payout. I just wouldn’t mistake payout stability for evidence that the stock itself is out of the woods.
