BCE Inc. (BCE:CA) Shareholder/Analyst Call Transcript

BCE announced construction of a new 300-megawatt Bell AI Fabric data center in Saskatchewan. The project materially expands BCE’s AI infrastructure footprint and positions the company to capture enterprise AI workloads, implying higher near-term capex but potential long-term revenue from AI services. Management (CEO Mirko Bibic, CFO Curtis Millen and others) provided details on a conference call and flagged standard forward-looking risks. Strategically positive for BCE, but not a market-transformative event.

Analysis

BCE’s strategic step into AI-scale infrastructure changes its capital mix and optionality more than its immediate revenue profile: the primary lever is control over latency, pricing and product bundling for enterprise AI workloads. That control can translate into outsized ARPU gains for high-margin managed services if BCE converts even a mid-single-digit percentage of national enterprise spend on private AI clouds; model a 3–5% ARPU uplift for business services over 24–36 months to justify meaningful multiple expansion.

Second-order beneficiaries are providers of renewables, grid upgrades and high-efficiency cooling — a sustained increase in local demand will force multi-year PPAs and substation upgrades, creating a callable revenue stream for project developers but also a timing and permitting risk that can delay go-live and push incremental opex into late-stage years. On the hardware side, sustained procurement of datacenter GPUs/accelerators will exacerbate vendor concentration (NVIDIA/AMD) and lengthen lead times, increasing financing needs and locking BCE into vendor price cycles.

Material tail risks are execution (permitting, grid reinforcement) and technology obsolescence: a 6–24 month delay materially increases IRR pressure and opens the door for hyperscalers to offer cheaper, elastic alternatives; conversely, early wins in commercial AI contracts would shorten payback to under 5 years. Watch 3-to-12 month catalysts: PPA signing cadence, announced enterprise contracts, and quarterly capex cadence — each can swing consensus EPS and re-rate multiples rapidly.

Key Decisions for Investors

• Long BCE.TO equity (or buy BCE Jan-2028 35/45 call spread): 12–24 month horizon. Rationale: optionality on managed AI services and multiple expansion; target +15–25% upside if ARPU lift materializes. Risk: capex overruns or delayed monetization could deliver a 10–15% downside; hedge with a 6–12 month put if risk-off.
• Long NVDA (or buy NVDA Jan-2027 1200 calls): 6–18 month horizon. Rationale: GPU scarcity and accelerated procurement for AI fabric benefits NVIDIA’s pricing power and ASPs. Risk/reward: asymmetric upside if ASPs hold; downside if macro demand softens or competitor pricing accelerates—limit position size to 2–4% of risk budget.
• Relative trade — Long BCE.TO / Short RCI.B (equal notional): 6–12 month horizon. Rationale: BCE’s vertical integration into AI infrastructure favors share gains in enterprise managed services vs peers reliant on third-party colocation. Risk: regulatory intervention or a competitive capex response from peers narrows spread; cap position size to a small relative-value sleeve.