BCE Inc. (BCE) — Saskatchewan AI Fabric MOU charts a path to 1.2 GW; ~5.6% indicated yield

Watch the video

---

The short version: an income stock funding an AI build

BCE Inc. (TSX: BCE), the parent of Bell and Canada’s largest telecom by revenue, traded at CA$31.20 on the September 18, 2026 screener, down about 3% over twelve months, 4% over the last month and more than 51% over five years. The new story is AI infrastructure: on September 14, Bell signed a Saskatchewan agreement that frames a path to 1.2 gigawatts of AI Fabric data-centre capacity, on top of a dividend that still yields about 5.6% after its earlier reset.

That second act costs money. Cap-X jumped about 41.5% in the second quarter and free cash flow fell 9.5%, so the question for a new buyer is whether the AI build can grow into contracted cash flow before it squeezes the dividend. Below we go through the Saskatchewan MOU, the Q2 Cap-X-versus-cash trade-off, and how BCE compares with its closest TSX telecom peers — Rogers, TELUS and Quebecor — on size, returns, valuation, dividends and analyst targets.

What happened: the Saskatchewan AI Fabric MOU

On September 14, 2026, Bell signed a non-binding memorandum of understanding with the Government of Saskatchewan. The agreement brings:

• up to 900 MW of additional AI data-centre capacity in the province,
• on top of 300 MW already under contract, with tenants including CoreWeave and Cerebras,
• a framed path to about 1.2 GW — with public commentary describing a phased, multi-decade ecosystem investment of more than C$50B.

The key word is non-binding. The extra 900 MW still needs customers, power supply and permits, and the C$50B figure is a long-range framing rather than a Bell spending plan. What the MOU does change is the scale of the ambition: Bell is positioning AI Fabric as a second growth engine alongside its fibre and wireless networks. Analysts reacted quickly — TD raised its target to C$40 and Desjardins to C$43, both with Buy ratings.

The Q2 print: Cap-X up, free cash flow down

The MOU builds on a second quarter reported August 6, 2026:

• Revenue C$6.18B, up 1.5%, with internet revenue up about 14% and roughly 55,000 fibre-to-the-home internet additions.
• Adjusted EBITDA C$2.70B, up 1%, and adjusted EPS of C$0.65, up 3.2%.
• Cap-X of C$1.08B, up 41.5% (17.5% of revenue), driven by Bell AI Fabric and U.S. fibre builds at Ziply Fiber.
• Free cash flow of C$1.04B, down 9.5% as spending rose.
• Guidance reaffirmed: revenue up 1–5%, adjusted EBITDA up 0–4%, capital intensity around 20%, free cash flow of C$2.1–2.3B and an annualized dividend of C$1.75.

Bell expects about C$1.3B of 2026 Saskatchewan data-centre Cap-X, mostly in the second half, and has received its first tenant payment as part of about C$400M of set-up fees and prepayments. The weak spots are the balance sheet and cash: net debt is about C$41.3B against roughly C$24.2B of equity (debt to equity about 1.75), and the screener’s trailing EPS of about C$6.72 is inflated by one-time items, which makes the trailing P/E of 4.6x meaningless. Forward earnings and free cash flow are the numbers that matter below.

Business overview: BCE vs Canadian telecom peers

BCE sells wireless, fibre internet and TV services through Bell, runs Bell Media, and since its Ziply Fiber purchase also owns a U.S. fibre business in the Pacific Northwest. The steady part is the subscription base — wireless plans and home internet that customers pay for every month. The cyclical or uncertain parts are media advertising, price competition in wireless and, now, the large AI data-centre build.

On the TSX screener, BCE sits in the Communications sector. Rogers (RCI.A) is the closest peer in size, with wireless, cable and media assets. TELUS (T) is the third national carrier, now in its own dividend-reset story. Quebecor (QBR.A) owns Videotron and Freedom Mobile and is the fast-growing fourth player. Smaller and illiquid names in the sector were left out, and Cogeco, which the screener classes outside Communications, was not included.

Overview table

Company (Ticker)	Price	Market cap	Revenue (LTM)	Net income (LTM)
Bell (BCE)	$31.20	$29.1B	$24.8B	$6.42B
Rogers (RCI.A)	$50.74	$26.7B	$22.6B	$6.17B
TELUS (T)	$12.47	$19.8B	$20.2B	−$919M
Quebecor (QBR.A)	$64.80	$14.3B	$5.79B	$944M
Average (incl. BCE)	—	$22.5B	$18.4B	$3.15B

Source: TSX screener (September 18, 2026, tsx-2026-09-18.csv) · CAD · LTM = last 12 months. Average includes BCE; blank cells skipped.

BCE is the largest of the four at about C$29.1B of market value and C$24.8B of trailing revenue, with Rogers close behind. Its C$6.42B of trailing net income looks strong, but it includes one-time items — treat it carefully when comparing. TELUS shows a trailing net loss after a large writedown, and Quebecor is the smallest by revenue and market value.

Performance: Quebecor wins, TELUS lags, BCE treads water

Performance table

Company (Ticker)	1M	3M	6M	1Y	5Y	10Y
Bell (BCE)	−4.1%	−4.9%	−11.3%	−3.1%	−51.6%	−48.0%
Rogers (RCI.A)	+0.7%	−5.2%	−7.0%	−0.5%	−17.2%	−9.6%
TELUS (T)	−7.2%	−23.1%	−30.6%	−43.2%	−56.3%	−40.2%
Quebecor (QBR.A)	+3.3%	−4.7%	+10.1%	+60.0%	+112%	+225%
Average (incl. BCE)	−1.8%	−9.5%	−9.7%	+3.3%	−3.3%	+31.9%

Source: TSX screener (September 18, 2026). Average includes BCE; blank cells skipped. Price change by window.

Over twelve months BCE is down about 3.1% and Rogers about 0.5%, while Quebecor is up about 60% and TELUS down about 43% — the clear winner and laggard of the group. Over one month BCE fell about 4%, worse than Rogers and Quebecor but better than TELUS. The spread shows this is not a sector-wide move: investors have rewarded growth and punished balance-sheet stress.

Over longer windows BCE is down about 52% in five years (second-worst after TELUS) and 48% in ten (the worst in the table). Quebecor is up about 225% over ten years.

Valuation: cheap on book and sales, not on free cash flow

Valuation table

Company (Ticker)	P/E	Fwd P/E	PEG	P/S	P/B	P/FCF	Price/Cash
Bell (BCE)	4.64	12.49	0.003	1.19	1.41	10.95	60.70
Rogers (RCI.A)	4.46	10.71	0.015	1.20	1.66	9.99	15.90
TELUS (T)	—	16.75	—	0.97	1.48	9.05	14.20
Quebecor (QBR.A)	15.90	14.24	0.70	2.58	5.29	9.52	120
Average (incl. BCE)	8.33	13.55	0.24	1.48	2.46	9.88	52.65

Source: TSX screener (September 18, 2026). Average includes BCE; blank cells skipped. Highlighted = lowest (cheapest) positive value per column. BCE’s and Rogers’ trailing P/E and PEG look unusually low because trailing earnings include large one-time items (PEG shown to 3 decimals, as in the video).

Ignore BCE’s 4.6x trailing P/E: one-time gains distort it, and the same goes for Rogers. On forward earnings BCE trades near 12.5x, a little below the 13.6x group average; Rogers is cheapest at about 10.7x and TELUS the priciest at about 16.8x. BCE is not expensive on price to sales (about 1.2x vs a 1.5x average) and has the lowest price to book in the group at about 1.4x.

The catch is free cash flow. At about 11x P/FCF, BCE is the most expensive of the four on that measure (the average is about 9.9x), because Cap-X is soaking up cash. TELUS screens cheapest on price to sales, P/FCF and price to cash. If AI Fabric tenants start paying and capital intensity falls back, BCE’s free cash flow should rise and that multiple should come down; if Cap-X stays near 20% of revenue, it won’t.

Dividend: high yield, thinner cash cover

Dividends table

Company (Ticker)	Dividend / share	Yield	Payout ratio
Bell (BCE)	$1.75	5.61%	26.0%
Rogers (RCI.A)	$2.00	3.94%	17.5%
TELUS (T)	$0.81	6.01%	—
Quebecor (QBR.A)	$1.40	2.78%	36.0%
Average (incl. BCE)	$1.49	4.58%	26.5%

Source: TSX screener (September 18, 2026). Average includes BCE; blank cells skipped. Yield = indicated. BCE dividend = current annualized rate (the screener’s C$2.31 annual DPS predates the reset); TELUS = next-12-month estimate (its 13.4% trailing yield is stale after the cut). Payout = trailing 12 months; TELUS blank (negative earnings).

BCE pays C$1.75 a year at its current rate (the screener’s C$2.31 annual figure predates the reset), for an indicated yield near 5.6% and a payout of about 26% of trailing earnings — a figure flattered by the one-time gains. A cash view is more useful: with about 930 million shares, the dividend costs roughly C$1.6B a year, around three-quarters of the C$2.1–2.3B free cash flow guide. Rogers yields about 3.9%, Quebecor about 2.8%, and TELUS about 6% on its post-cut estimate. BCE is still an income stock, but the cushion depends on the AI build not crowding out free cash flow.

Analyst view: Street targets imply upside

Analyst table

Company (Ticker)	Rating	Price	Target 1Y	Upside	Analysts
Bell (BCE)	Buy	$31.20	$37.70	+20.8%	—
Rogers (RCI.A)	Buy	$50.74	$59.00	+16.3%	—
TELUS (T)	Neutral	$12.47	$14.30	+14.7%	—
Quebecor (QBR.A)	Buy	$64.80	$69.70	+7.6%	—
Average (incl. BCE)	—	—	—	+14.8%	—

Rating & price: TSX screener (September 18, 2026). Target 1Y: MarketBeat consensus as shown in the video’s analyst still (September 2026): BCE ~C$37.70, RCI.A ~C$59.00, T ~C$14.30, QBR.A ~C$69.70. Analyst counts were not recorded. Upside = target ÷ price − 1.

The screener rates BCE a Buy. MarketBeat’s consensus is about Moderate Buy with an average target near C$37.50–37.70, roughly 21% above the C$31.20 price. After the MOU, TD raised its target to C$40 with a Buy, and Desjardins raised its target to C$43, also Buy. Rogers and TELUS show about 15–16% upside to their consensus targets and Quebecor about 8%, so BCE has the most implied upside in the group — the Street is betting that AI Cap-X turns into contracted cash flow.

Why it matters: from income telecom to AI infrastructure

The old narrative was BCE as a classic Canadian telecom income name after a dividend reset: slow growth, heavy debt, a big yield. The new narrative asks investors to fund a Cap-X-heavy AI Fabric build while collecting a mid-5% yield, with named tenants on the first 300 MW and a path to 1.2 GW.

Income investors care whether free cash flow keeps covering the C$1.75 dividend through the build. Growth and AI-narrative investors care whether 1.2 GW becomes real contracted megawatts or stays mostly MOU optionality. The open question is timing: how long BCE can spend ahead of revenue before the balance sheet forces a choice.

Bull case

1. Named tenants and scale. CoreWeave and Cerebras on the first 300 MW, prepayments already flowing, and a path toward 1.2 GW.
2. Core operations still growing. Q2 revenue, adjusted EBITDA and adjusted EPS all rose, with internet revenue up about 14% and wireless postpaid churn near 1.02%.
3. Clearer income math. An indicated yield near 5.6% on a reset dividend, with analyst targets of C$40–43 implying meaningful upside if Cap-X converts.

Bear case and risks

1. Cap-X pressure. Cap-X up 41.5% in Q2 and guided near 20% of revenue is already cutting free cash flow (down 9.5%).
2. Non-binding MOU. The extra 900 MW needs customers, power and permits, and the C$50B-plus framing is multi-decade and speculative.
3. Weak tape, heavy leverage. BCE is down more than half over five years, and net debt of about C$41.3B leaves little room for error.

What to watch

1. MOU conversion. Whether the Saskatchewan megawatts turn into signed contracts, power agreements and permits.
2. Second-half AI Cap-X — about C$1.3B planned in Saskatchewan this year — versus tenant prepayments and set-up fees.
3. Free cash flow vs guidance. Whether 2026 lands in the C$2.1–2.3B range with capital intensity near 20%.
4. The dividend. Whether the C$1.75 annualized rate holds through the AI build, and whether targets near C$38 move toward the C$40–43 broker calls.

Bottom line

BCE has a credible second act: named AI tenants, a path to 1.2 GW, a growing fibre business and core results that still edge higher. The Saskatchewan MOU makes the ambition bigger without yet adding contracted revenue. The issue is cash, not price. At about 12.5x forward earnings and a 5.6% yield, BCE looks cheap on most measures, but at about 11x free cash flow it is the priciest of its peers on the metric that pays the dividend. The next test is whether the second-half Cap-X brings matching tenant cash.

Video chapters

• 0:00 Hook — AI Fabric path to 1.2 GW
• 0:15 What happened: Saskatchewan MOU (September 14) + Q2 Cap-X vs free cash flow
• 0:47 Overview — telecom peers
• 1:13 Performance heatmap
• 1:26 Valuation
• 1:40 Dividends (~5.6% yield)
• 1:49 Analyst ratings & targets
• 2:12 Why it matters
• 2:37 Bull case
• 2:59 Bear case
• 3:17 What to watch

Data notes

Table figures are TSX screener values (merge tsx-2026-09-18.csv, 2,566 rows; CAD). Peers are liquid same-sector (Communications) carriers: Rogers (closest in size), TELUS (national wireless and fibre) and Quebecor (Videotron/Freedom); smaller and illiquid names were skipped. Average rows include BCE and skip blank cells. Forward P/E comes from the same screener merge (the video’s valuation still did not show it). Dividends use BCE’s current annualized rate (C$1.75; the screener’s C$2.31 annual DPS is stale after the reset) and TELUS’s next-12-month estimate (C$0.81; its 13.4% trailing yield is stale after the cut), as in the video; the payout average is a straight mean of the three non-blank cells (the video rounded it to 26%). Target 1Y is MarketBeat consensus as shown in the video (BCE about C$37.5–37.7); analyst counts were not recorded, so those cells are blank. The dividend-cost estimate uses shares implied by market cap ÷ price. Q2 figures are from BCE’s August 6, 2026 release; the MOU details are from September 14, 2026 reporting. Video chapter times come from the rendered video’s section timings.

Sources

• BCE Q2 2026 results release, August 6, 2026
• Bloomberg and Stockwatch reporting on Bell’s Saskatchewan AI Fabric MOU, September 14, 2026
• MarketBeat — BCE, RCI.A, T and QBR.A consensus targets; TD (C$40) and Desjardins (C$43) target changes (September 2026)
• TSX screener merge September 18, 2026 (2,566 rows)

Media credits

BCE/Bell logo (company trademark, used for identification). The background visuals — a prairie data centre, a server-rack aisle, glowing fibre-optic strands, a telecom tower at dusk and a dividend-versus-Cap-X illustration — were generated with an AI image tool for the video; the data tables were built in-house. Some background images are AI-generated.

By Andrew · stocksummarizer.com