Pembina Pipelines 4.6B Meta Power Deal|Alberta Grid Full, 37 AI Projects Stranded

The Pipeline That Won Alberta's AI Lottery

Pembina Pipeline announced today it will spend $4.6 billion to build a 932-megawatt power station in Sturgeon County, Alberta — a facility sized to supply a single data centre that previous reports linked to Meta Platforms. That link matters because it is not a pipeline story. The company that moves natural gas across the Prairies has just bet its largest disclosed capital commitment on powering Facebook and Instagram's AI infrastructure. The bottleneck here is not the gas; Alberta has a glut of it, keeping local prices stubbornly low for years. The bottleneck is who controls the power delivery path to hyperscale customers, and as of today, Pembina controls Canada's only contracted gigawatt-scale path. Morgan Stanley Infrastructure Partners joined the project alongside Calgary's Kineticor Asset Management, with a fixed-price turbine agreement with Siemens Energy locking in both construction timing and cost. The CEO described it as "building a business, not a project" — a phrase worth holding onto, because whether that framing survives the next chapter of Alberta's AI story is the question this video resolves.

The 1,200 Megawatt Ceiling and the 37 Projects Left Outside

The same day Pembina's FID crossed the wire, Alberta's grid operator quietly closed the door on everyone else. The Alberta Electric System Operator allocated its entire 1,200-megawatt interim cap for large-load data centre connections to just two projects: Pembina's Greenlight at 970 MW and TransAlta's Keephills at 230 MW. That leaves 37 other data centre proposals in the queue requesting a cumulative 19.4 gigawatts — almost 14 times the power needed to run the city of Edmonton — without a grid connection path. Phase 2 of the Large Load Integration Program, which would govern those 37 projects, is still under development. One University of Alberta engineering professor said developing new transmission infrastructure alone can take 10 years. This is where the two readings of today's announcement split. Pembina's CEO called the project the "first mover" in a new gas-to-power platform, and pointed to additional data centre opportunities ahead. That reading requires the 37 stranded projects to eventually build their own "bring your own power" generation — and in doing so, contract volumes through Pembina's pipeline network, since the Greenlight plant itself requires 150 million cubic feet per day of natural gas transported through Pembina and TC Energy systems. But a competing read sits in the same articles: 37 proposals are frozen, Phase 2 has no timeline, and a province that promised $100 billion in data centre investment by 2030 has allocated just $4.6 billion of committed capital so far. The conflict is between Pembina as the gateway into a durable multi-project platform, and Pembina as the ceiling of a one-shot allocation in a grid-constrained province.

Valuation, Gas Glut, and the Monitoring Variable

Pembina entered 2026 trading near CA$68.45, already up 29.22% year to date before today's announcement. One narrative pegs its fair value at CA$63.61 — meaning the stock was already trading roughly 8% above that estimate before the FID. A separate cash-flow model places intrinsic value at CA$194.98 per share, implying significant undervaluation. The gap between those two estimates is not a rounding error; it reflects whether the gas-to-power pivot is valued as an incremental midstream project or as a new revenue layer that reprices the entire franchise. The counter-evidence in today's pool is the grid ceiling: if the 37 stranded projects cannot resolve their power situation within a horizon that matters to capital markets, Pembina's Greenlight remains a one-client, one-project story generating contracted tolling revenue from 2030 onward — valuable, but not the platform-level re-rating the CEO's language implies. For a holder, the near-term monitoring variable is not the Greenlight construction timeline but AESO's Phase 2 process: a defined framework with timelines for the 37 stranded projects would confirm the "building a business" thesis and validate the higher intrinsic value read. Its absence — an open-ended "still under development" — leaves the 29% YTD run resting on a single contracted tolling deal with an unnamed customer. For a watcher, the entry setup resolves if Phase 2 produces a structured timeline that routes additional data centre build-outs through bring-your-own-power generation, funnelling incremental gas volumes through Pembina's network. The trap confirms if Phase 2 stalls and the 37-project backlog shrinks through attrition or relocation — leaving Greenlight as Pembina's only AI data centre exposure while its share price still prices in a platform story. Watch AESO Phase 2 disclosure, not the Greenlight shovels.