BCE (TSX: BCE) just made one of the biggest corporate investment promises in Canadian history, and it has very little to do with phone plans.
The telecom giant plans to expand Bell AI Fabric, its sovereign AI infrastructure ecosystem in Saskatchewan, by up to 900 megawatts, creating a path to a 1.2-gigawatt (GW) artificial-intelligence hub. At full buildout, Bell says the project could represent up to $52.5 billion in capital investment.
That is a staggering amount for a company investors have spent years treating mainly as a dividend stock. The question is whether this marks a genuine growth transformation for BCE … or simply a very expensive new chapter.
Source: Getty Images
Not your father’s BCE
AI models need enormous computing power, which requires data centres, fibre connections, networking equipment, cooling systems, and reliable electricity. Canada has several advantages in that race, including having relatively abundant power and cooler weather, along with the political will to keep sensitive computing infrastructure inside the country.
Bell wants to sit squarely in the middle of that ecosystem.
The Bell AI Fabric strategy is designed to provide Canadian businesses and governments with computing capacity, fibre connectivity, cybersecurity, and other AI-related services. The Saskatchewan expansion would add up to 900 MW beyond the 300 MW project already underway, potentially creating the country’s largest sovereign AI infrastructure hub.
That makes BCE look rather different from the company investors knew a few years ago.
BCE spending heavily
There’s one problem. Building data centres costs a lot of money before they start producing much cash.
BCE’s second-quarter capital expenditures jumped 41.5% year over year, to $1.08 billion, partly because of spending on Bell AI Fabric and its U.S. fibre network. Free cash flow fell 9.5%, to $1.04 billion.
Management expects 2026 capital intensity of roughly 20%, up from 15.1% in 2025, with about $1.3 billion of incremental spending tied to the initial Saskatchewan data-centre build.
In short, it’s expensive. So why buy BCE stock today? Because there are early signs the strategy is producing growth.
In the first quarter, Bell Business Markets’ revenue rose 9.7%, while revenue from its AI-powered solutions more than doubled. In the second quarter, combined revenue from Ateko and Bell Cyber increased 29% year over year.
Those are still relatively small pieces of BCE. Yet they offer something its traditional telecom business has struggled to deliver consistently: meaningful growth.
What about BCE’s dividend?
For a long time, BCE reliably increased its payout, making the stock a favorite with income investors. But the company reset those expectations when it reduced the dividend in 2025. (The annualized dividend now sits at $1.75 per share.) The cut was painful, but it also freed up cash that the company can use for debt reduction and investment. The lower payout also removes some of the pressure that previously made BCE’s giant yield increasingly uncomfortable.
The trade-off is clear: BCE is no longer asking investors to buy the stock mainly for maximum income today. It’s asking them to tolerate heavier spending now in exchange for a healthier, faster-growing business tomorrow.
Is BCE stock finally a buy?
I think BCE has become much more interesting in the past day, but I wouldn’t buy it because of the $52.5 billion headline alone.
The Saskatchewan agreement is non-binding, the buildout will happen in phases, and the final scale depends on customer demand, available electricity, permitting, and execution. Data centres can become expensive mistakes if capacity grows faster than paying demand.
Still, BCE now has something it badly needed: a believable growth story.
The core telecom network continues to generate billions in cash flow, while fibre, cybersecurity, enterprise AI, and data-centre infrastructure give the company new ways to grow. Investors learning about buying stocks in Canada should view BCE as a turnaround with an AI option attached, not as a guaranteed winner.
Bottom line
If Bell can turn even part of its enormous infrastructure ambition into recurring enterprise revenue, today’s heavy spending could eventually look far more productive than another round of defending market share with discounted phone plans.
That makes BCE a stock I’d consider buying into gradually, with the expectation that the most interesting part of the story may still be several years away.
