Capital Power

Phase A — Understand the business

Lens 1 · Company Overview

Capital Power is a North American independent power producer (IPP) headquartered in Edmonton, Alberta — a merchant generator that owns power plants and sells electricity and capacity into wholesale markets, plus a growing book of contracted supply. As of year-end 2025 it operated ~12 GW of generation across 35 facilities. The critical fact about this company in 2026 is that it is not the company it was three years ago: it has executed a deliberate, capital-heavy pivot from an Alberta coal-legacy utility into a US-weighted merchant natural-gas platform explicitly positioned to sell power to AI data centers.

Three structural facts define the business today:

1. Natural gas is now the spine. Management frames high-efficiency gas as "the flexible backbone" for surging grid demand. The US Flexible Generation portfolio is the single largest segment at 6.2 GW of gas capacity, and the US now represents roughly 60% of both capacity and adjusted EBITDA. This is a fleet-level bet that gas peakers/combined-cycle plants are the winning asset class in a grid that must firm intermittent renewables and absorb data-center load.

2. It exited its coal identity ahead of schedule. The Genesee Generating Station west of Edmonton — its flagship — became 100% natural-gas-fueled on June 18, 2024, five years ahead of the Alberta mandate, via a ~$1.6B repowering that lifted Genesee to up to ~1,857 MW (a ~60% capacity increase, ~40% emissions cut).

3. It is monetizing the demand story via contracts, not just merchant exposure. In December 2025 it signed a binding MOU with an investment-grade data-center developer for a 250 MW Energy Supply Agreement, 10+ years, anticipated 2028 start, backed by its Alberta portfolio (with a termination fee if the final deal fails).

How it makes money. Revenue = energy sales + capacity payments + ancillary/flexibility revenue across three market regimes — (a) Alberta, an energy-only market with a $1,000/MWh cap (merchant, volatile); (b) US merchant markets, primarily PJM (energy + capacity auctions); and (c) a contracted tail (PPAs, the pending data-center ESA, tolling on some assets). FY2025 revenues and other income were $3,720M on adjusted EBITDA of $1,580M. Contract structure is mixed — this is not a fully-contracted regulated utility (which is the crux of both the bull and bear case): a material slice of cash flow is merchant and rides wholesale power prices.

Customers / suppliers / competitors. End buyers are grid operators / load-serving entities and, increasingly, hyperscale data-center developers. The key input is natural gas (a direct commodity-cost and margin driver) plus the AESO/PJM capacity-market rules. Competitors: TransAlta (the other Alberta gas/hydro heavyweight), and in the US merchant-gas arena Vistra, NRG, Talen, Constellation, LS Power (the seller of the two PJM plants Capital Power just bought).

Lens 2 · Supply Chain

Named, end-to-end — inputs → Capital Power → offtaker:

Upstream (inputs & counterparties)

• Natural gas — the primary fuel; sourced from Western Canadian (AECO hub) and US (Henry Hub / regional PJM basis) gas markets. Gas price is a direct pass-through risk to spark spreads. Trading/hedging run through the internal Supply & Trading desk (SVP Jason Comandante).
• Turbine OEMs / EPC — the Genesee repowering used combined-cycle gas turbines (Genesee 1&2 marketed as "Canada's most efficient natural gas combined cycle units"); OEM not named in the sourced material — n/a on the specific turbine vendor.
• Capital / financing counterparties — a load-bearing part of the "supply chain" for a capital-intensive IPP. Named: Apollo Global Management (US$3B merchant-gas acquisition partnership), Alberta Investment Management Corporation / AIMCo ($150M private placement in the April 2025 equity raise), and the BBB- syndicate (S&P, plus DBRS and a new Fitch BBB- — three investment-grade ratings).

Midstream (the assets)

• Alberta hub: Genesee (~1,857 MW gas), plus other Alberta gas/cogeneration.
• US Flexible Generation (6.2 GW gas): the June 2025 acquisition of Hummel Station (1,124 MW CCGT, Shamokin Dam, PA) and Rolling Hills (1,023 MW combustion-turbine peaker, Wilkesville, OH) from LS Power for US$2.2B (~C$3.0B); plus Frederickson 1 (265 MW CCGT, Pierce County, WA; 50.15% interest, US$100M, Dec-2023) and legacy US gas (e.g. Goreway in Ontario, Genesee-adjacent US plants).
• Renewables/storage tail (shrinking): wind (Halkirk 2 operational 2025), solar, and 170 MW of Ontario battery storage added 2025 — but management is selling down renewables (see below).

Downstream (offtakers)

• AESO / Alberta grid (merchant + the 250 MW data-center ESA from 2028).
• PJM (energy + capacity auctions across PA/OH assets).
• Hyperscale data-center developers — one investment-grade developer named via MOU; not publicly identified. n/a — developer not disclosed.

Chokepoints / single-source dependencies.

• Gas fuel + basis risk is the dominant chokepoint — a fleet-wide input with no substitute on the timescale that matters.
• Regulatory single points of failure: the AESO market-design reform (Restructured Energy Market / REM, nodal LMP, ~2027) and the Alberta $125/MWh-or-25×-gas secondary offer cap on large non-renewable generators directly govern Alberta merchant margins.
• PJM capacity-auction outcomes — a policy-set price that swings US EBITDA.

Names present, chain mapped — lens passes. Gaps (turbine OEM, the data-center counterparty) are labeled, not invented.

Lens 3 · Competitive Advantages (moats)

Be honest: for a merchant IPP the "moat" is mostly location, scale, and the regulatory regime — not a durable franchise. Capital Power's real edges:

1. Alberta incumbency + the Genesee cost position. Genesee is a scaled, newly-repowered, high-efficiency baseload-capable gas complex sitting in a province with a genuine, policy-visible data-center demand wave (AESO bracing for ~11 GW of new demand). Being the lowest-heat-rate, largest-single-site gas generator in your home market is a moat when new-build lead times are years. This is the strongest, most durable part of the thesis.
2. Scale + capital access. Post-2025 it is "one of only five North American IPPs with >10 GW of gas capacity". Scale plus three investment-grade ratings and the Apollo US$3B partnership (which lets it deploy US$750M of its own capital to control US$3B of assets at 25–50% working interest and collect management + performance fees) is a genuine capital-efficiency advantage — it manufactures fee income and optionality on top of equity returns.
3. Flexibility as product. In grids increasingly defined by intermittency (AESO is "a flexibility market disguised as an energy market"), a dispatchable peaker fleet earns scarcity value renewables can't.

Bargaining power. Weak-to-moderate. Against gas suppliers it is a price-taker on the commodity (hedged, not owned). Against the grid/offtaker its power depends entirely on scarcity — high when reserve margins tighten (the AI-demand thesis), very low when the market is oversupplied (Alberta 2024). The 250 MW ESA with an investment-grade counterparty is an attempt to convert merchant volatility into contracted certainty — i.e. management itself is signaling the merchant "moat" isn't reliable enough to fund the growth plan on.

Verdict on moat: Real but conditional. It is a cyclical scarcity moat, not a structural franchise moat. It widens dramatically if AI power demand shows up on schedule and narrows to near-zero in an oversupplied market. That conditionality is the whole investment debate.

Lens 4 · Segments

Capital Power does not publish a clean product×geography EBITDA matrix in the sourced releases, so the segment picture is reconstructed from disclosed portfolio and the "US ~60%" statement — every number labeled.

Precise segment EBITDA/revenue split (US vs Canada vs Ontario): n/a at the sub-segment level. The only firm split disclosed is US ≈ 60% of capacity and adjusted EBITDA; Canada ≈ 40%.

The trend — and the cause — is the whole story:

• US is accelerating, deliberately. From a Canadian-majority company three years ago to US ≈ 60% of EBITDA in 2025, driven by (i) the US$2.2B Hummel/Rolling Hills PJM acquisition (June 2025), (ii) Frederickson (Dec 2023), and (iii) the Apollo pipeline that is designed to keep adding US merchant gas. The 2030 plan targets a 50% cumulative increase in US capacity (+3.5 GW).
• Renewables are decelerating by choice. Capital Power sold 49% of two Canadian wind facilities (Quality Wind BC + Port Dover/Nanticoke ON, 246 MW total) to Axium Infrastructure for ~$340M pre-tax, explicitly "crystallizing a levered equity return above its capital-allocation thresholds". Read plainly: management is rotating capital out of contracted renewables and into merchant gas — a contrarian, cash-return-driven, arguably late-cycle bet that is the single most important strategic signal in the file.

---

Phase B — Measure performance

Lens 5 · Earnings Result (FY2025 / Q4 2025)

FY2025 was a record year — but read the composition carefully.

Q4 2025 specifically: adj. EBITDA $414M, AFFO $244M, but a net LOSS of $(13)M. The full-year EPS of $0.88 against a ~$73–75 share price implies a trailing P/E near 69× and a large gap between GAAP net income ($160M) and AFFO ($1,066M) — a ~6.7× AFFO-to-net-income multiple. That gap is normal for a capital-intensive generator (heavy D&A on a repowered/acquired fleet, plus fair-value and one-time items), but it means the equity is valued almost entirely on AFFO/cash flow, not accounting earnings — a critical framing for Lens 7 and 11.

What drove it: the record was powered by (i) the newly-consolidated US PJM assets (a bought step-change, not organic growth), (ii) asset optimization / strong contract renewals, and (iii) the full-year contribution of the repowered Genesee. Margin quality flag: a chunk of the YoY EBITDA jump is acquisition-additive and capacity-market-driven, not same-store margin expansion — and it landed despite a brutal Alberta pricing year (below). That resilience is a genuine positive; the reliance on M&A to grow is a genuine caution.

Balance-sheet flags: capex ($864M) ran below operating cash flow ($962M) at the sustaining level, but the growth was debt+equity-funded — the US$2.2B acquisition was financed with ~$667M equity (incl. $150M AIMCo private placement) + ~US$1.2B senior notes + additional notes (total ~$2.3B notes issued in 2025). Liquidity of $1.8B available at year-end. Net debt / net-debt-to-EBITDA: n/a (not disclosed in the press releases; would require the audited financials / MD&A). The BBB- rating and "well within financial guardrails" language is the only leverage read available.

Market reaction: the stock is up ~30% over the trailing year to ~$72–75, near a 52-week high of $73.80 (range $54.03–$73.80). The tape has already rewarded the pivot.

Lens 6 · Earnings Calls (sentiment trend)

No transcripts on the shelf (transcripts=0); sentiment is reconstructed from the sourced release language and Investor Day framing across 2025 — labeled ``.

• Recurring, escalating phrases: "flexible generation," "natural gas backbone," "surging demand for reliable power," "data centres," "scale and diversification." The rhetorical center of gravity moved decisively from decarbonization to gas-fired reliability + AI load over 2024→2025.
• What they stopped saying: the aggressive net-zero / carbon-capture framing that dominated 2022-2023. The Genesee CCS cancellation (May 2024, see Lens 9) marks the pivot point — the "low carbon leader" narrative was quietly replaced by "gas is the transition fuel."
• Tone: confident, growth-forward, capital-markets-fluent (Avik Dey is ex-CPPIB/Carlyle — the messaging is investor-grade, M&A-centric). The December 2025 Investor Day ("Powering Growth through Natural Gas") is the clearest statement: 13–15% annual TSR, 8–10% AFFO/share growth, 2–4% dividend growth to 2030. Management is selling a growth-compounder story, not a bond-proxy utility story — a meaningful re-positioning of the shareholder base.

Lens 7 · Comps

Read:

• Capital Power trades at a premium EV/EBITDA (~16×) to Vistra (~11×) and TransAlta (~14×), and roughly in line with NRG (~16×) — i.e. the "AI power" re-rating is already in the multiple. It is not a cheap merchant IPP; it is priced as a growth compounder. Talen's 67× is a data artifact (nuclear/contract-mix distortion) — ignore it.
• Its standout feature is yield: ~4.55%, roughly 4–8× the US merchant peers. That reflects (a) its Canadian dividend-utility heritage and shareholder base, and (b) that the market still partly prices it as an income name even as management re-pitches it as growth. This yield-vs-growth identity tension is the single most interesting valuation feature — bulls get paid to wait; bears note that a 4.5% yield on a merchant-gas grower funded by BBB- debt is not a safe bond proxy.
• Do not treat the 69× trailing P/E as meaningful — GAAP EPS ($0.88) is depressed by D&A on the acquired/repowered fleet; the market values AFFO ($7.08/sh → ~10.3× P/AFFO at ~$73 ), which is a far more sensible ~10× cash multiple. P/AFFO ~10×, not P/E ~69×, is the number that matters.

Lens 8 · Stock-Price Catalysts (last ~5 years)

What actually moves CPX, from the sourced record — mostly ``:

1. Alberta power prices (the dominant macro driver). Alberta's average pool price collapsed 53% in 2024 to $62.78/MWh (from $133.63 in 2023) as ~3,000 MW of new supply hit the grid and the province became a net exporter. Merchant-heavy Alberta IPPs (CPX, TransAlta) live and die on this number — this is the single biggest swing factor and the core bear input.
2. M&A / capital deployment. The US$2.2B Hummel/Rolling Hills acquisition (announced April 2025, closed June 9 2025) and the US$3B Apollo partnership (Dec 2025) were re-rating catalysts — the market rewarded the credible US-scale-up.
3. The data-center / AI-power narrative. The 250 MW ESA MOU (Dec 2025) and Alberta's ~11 GW demand pipeline turned CPX into an "AI power" name — a large part of the +30% trailing-year move.
4. Genesee milestones. Off-coal (June 2024) and repowering commercial ops (Dec 2024) de-risked the flagship.
5. Capital-allocation resets. The Genesee CCS cancellation (May 2024) and the wind sell-down (Nov 2024) were narrative-shifting events — the market broadly rewarded the discipline (walking from a $2.4B uneconomic project; harvesting a levered wind return).

Pattern: the market reacts to (a) Alberta wholesale prices, (b) accretive US M&A, and (c) the AI-demand narrative — in that order of durability. It is not an earnings-surprise-driven name quarter-to-quarter; it is a macro-power-price + capital-deployment story.

---

Phase C — Judge people & books

Lens 9 · Management

CEO — Avik Dey (President & CEO since 2023). This is a capital-allocator CEO, not a lifer operator — and that fact explains the entire strategy.

1. Track record. Ex-Managing Director & Head of Energy/Natural Resources at CPP Investments (CPPIB) — built that group "from an idea to a leading global energy-transition investor," sat on boards of CPPIB-owned oil & gas companies; then co-head of Carlyle International Energy Partners; then CFO of NOVA Chemicals (2021–22). Claims to have "invested over $12B" across energy/energy-transition. Named to Canada's Top 40 Under 40.. This is a private-equity/institutional-capital pedigree — which is exactly why Capital Power now behaves like a PE-style rollup (Apollo partnership, buy-merchant-gas, sell-down-renewables-at-a-return, fee income).
2. Tenure & skin in the game. ~2.5 years in the seat. Insider ownership is thin: direct holding ~0.003% of shares (~C$273K) against ~C$5.39M total annual comp (17.7% salary / 82.3% incentive). Flag: low direct ownership relative to comp — alignment is via incentive plans, not a large founder-style equity stake. For a CEO steering a bet-the-balance-sheet pivot, more personal skin would be reassuring.
3. Capital-allocation history — the most important evidence, and it's genuinely good on discipline:

• Walked from the $2.4B Genesee CCS project (May 2024) when the federal government wouldn't provide revenue certainty — "technically viable but not economically feasible." Avoiding a multi-billion-dollar uneconomic capex is a strong capital-discipline signal.
• Harvested the wind sell-down (49% of 246 MW to Axium, ~$340M, "levered return above thresholds"). Selling contracted assets at a good IRR to fund higher-return merchant gas is textbook opportunistic capital rotation.
• Structured the Apollo partnership to control US$3B of assets with US$750M of equity + fees — capital-efficient, PE-style.
• 12 consecutive years of dividend growth, +6% in 2025 — but note the forward target was cut to 2–4% dividend growth (from a historically higher cadence), signaling capital is being retained for growth over payout.

1. Red flags. (i) Merchant-gas concentration at cycle-arguably-late — rotating into commodity-price exposure and out of contracted cash flow is a high-conviction directional bet that can look brilliant or reckless depending on where power prices go. (ii) Low insider ownership. (iii) Reliance on financial engineering (Apollo fee model, debt+equity-funded M&A on BBB- credit) to hit 8–10% AFFO/share growth — growth that is bought and levered, not organically compounded.
2. Archetype. Professional institutional-capital allocator running an IPP like a PE platform. For this stage (a scale-up trying to ride the power-demand super-cycle), that is arguably the right archetype — it is why the US expansion happened at all. The risk is that a capital-allocator optimizes for deployment velocity and re-rating, and the balance sheet wears the downside if the power-price thesis slips.

Lens 10 · Forensic Red Flags

Grounding caveat: no SEC filings exist (no CIK) and the audited SEDAR+ financials were not on the shelf, so this is a disclosure-limited forensic read from press releases + the funding structure. Where the income-statement/balance-sheet detail needed for a full forensic pass is absent, it is flagged as such rather than guessed.

• GAAP-vs-non-GAAP gap (the main flag to watch). Net income attrib. $160M vs AFFO $1,066M and adj. EBITDA $1,580M. A ~6.7× AFFO/net-income ratio is explainable (D&A on a repowered + freshly-acquired fleet, fair-value/derivative items on the trading book, one-time acquisition costs), and AFFO is the correct metric for a generator — but it means the entire equity value rests on management-defined AFFO adjustments. The specific add-backs bridging net income → AFFO were not in the sourced releases (n/a); a serious position requires reading the MD&A reconciliation. This is the #1 thing to verify before sizing.
• Acquisition-driven growth optics. The +18% EBITDA / +29% AFFO YoY is materially bought (Hummel/Rolling Hills consolidation). Nothing improper — but "record results" language can flatter what is partly a balance-sheet-funded step-up rather than organic margin gains. Watch the first full-year post-acquisition organic comp (2026) for the underlying run-rate.
• Leverage opacity. Net debt and net-debt/EBITDA were not disclosed in the sourced releases (n/a). The company issued ~US$1.2B + additional notes (~C$2.3B total in 2025) plus $667M equity to fund the US acquisition, and carries BBB- (three ratings incl. new Fitch) — i.e. the lowest rung of investment grade. On a merchant-gas fleet, BBB- with a large debt-funded acquisition is a real fragility: a downgrade to sub-IG would raise the cost of the whole growth model.
• Trading-book / derivatives fair value. An internal Supply & Trading desk means mark-to-market derivative gains/losses flow through results — a classic area where non-GAAP smoothing lives. Detail n/a; flag for MD&A review.
• Receivables/inventory outrunning revenue; SBC flattering non-GAAP: n/a (requires the financial statements). No evidence either way in the sourced material.

Regulatory findings (required sub-section):

• SEC (EDGAR LR + AAER): None — Capital Power has no CIK and is not an SEC filer; no EDGAR enforcement search is possible.
• Non-SEC (FTC / DOJ / provincial / AUC / market-surveillance): Web search "Capital Power" (FTC OR DOJ OR settlement OR fine OR penalty OR "consent decree" OR lawsuit) enforcement returned no material enforcement action, settlement, fine, or consent decree naming the company in 2024–2025. (Note: Alberta generators operate under AESO/AUC/Market Surveillance Administrator oversight — the interim market-mitigation rules of July 2024, e.g. the secondary offer cap, are market-design changes, not enforcement against Capital Power specifically.)
• 10-K Item 3 (Legal Proceedings): not available — no 10-K exists (files on SEDAR+; the AIF was not on the shelf). n/a — SEDAR filer, AIF not ingested.
• Conclusion: No material regulatory or legal findings identified — verified via SEC EDGAR EFTS (LR + AAER, zero), a targeted non-SEC web search (no hits), as of 2026-07-06. The one policy risk (not a legal finding) is Alberta market-design reform, covered in Lens 2/13.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (AFFO-based — the right metric here)

Because GAAP EPS is a poor proxy for this fleet, the projection is built on AFFO/share (the metric management guides and the market prices), then cross-checked to the company's own 2030 target. Every input labeled; output ``. Per the --watchlist rules, no forecast.ts create is logged here.

Anchor actuals: FY2025 AFFO $1,066M, AFFO/share $7.08; management guides FY2026 adj. EBITDA $1,565–$1,765M and AFFO $890–$1,010M, sustaining capex $290–330M. Note the FY2026 AFFO guide midpoint (~$950M) is BELOW FY2025 actual AFFO ($1,066M) — a critical, easily-missed fact: 2025 AFFO/share was flattered by items (likely realized trading/optimization gains and timing) that management does not expect to repeat at the same level in 2026. The base case must not naively grow off $7.08.

FY2026 base: AFFO ~$950M ÷ ~157M shares → AFFO/share ≈ $6.05. This is a step DOWN from 2025's $7.08 on a higher share count — the honest base case for the first year.

Cross-check to the 2030 target: management's 8–10% AFFO/share CAGR to 2030 off a normalized ~$6/sh 2026 base implies ~$8.8–$9.7 AFFO/sh by 2030. That is achievable but assumes both (a) Alberta price normalization and (b) sustained accretive capital deployment — remove either and the plan under-delivers. The plan is credible but conditional on the power-price cycle turning back up.

Valuation frame: at ~$73 and a normalized $6/sh AFFO, P/AFFO ≈ 12×; on 2028 base ($7.1) it's ~10×. Not demanding if the growth shows up; expensive if AFFO stalls at the 2026 guided level.

No forecast.ts create logged (watchlist rule). If promoted to a thesis, the scoreable base call would be: "CPX FY2026 AFFO/share ≥ $6.00" — a binary the guidance ($890–1,010M AFFO) makes trackable.

Lens 12 · Bull vs Bear

Bull case. Capital Power is a scaled, dispatchable-gas platform sitting in front of the two best demand pools in North American power — Alberta's ~11 GW pipeline and PJM's data-center-driven load — with the capital machinery (Apollo + IG balance sheet + fee model) to keep adding assets and an income-grade 4.5% yield paying you to wait. Genesee is a genuine low-cost, newly-repowered scarcity asset. Management has proven capital discipline (walked from CCS, harvested wind at a return) and is converting merchant volatility into contracted cash flow (the 250 MW ESA is the first of many). If AI power demand shows up on anything like the consensus schedule, a merchant-gas fleet is the highest-torque way to own it, and 8–10% AFFO/share growth to 2030 re-rates the stock toward its C$82 street high. Earnings surprise to the upside comes from Alberta price normalization + PJM capacity-price strength hitting a fleet that already grew EBITDA 18% in a bad Alberta year.

Bear case (2–3 permanent-impairment-grade risks).

1. The Alberta merchant thesis is structurally impaired, not just cyclical. Pool prices fell 53% in 2024 to ~$63/MWh as supply flooded in; the province is now a net exporter; and the Restructured Energy Market (nodal LMP, ~2027) + the $125/MWh-or-25×-gas secondary offer cap on large generators are designed to suppress the very scarcity spikes merchant IPPs need. If Alberta is now a structurally-oversupplied, price-capped market, ~40% of EBITDA sits on a permanently lower plateau.
2. The growth is bought and levered on BBB- credit. The record 2025 was M&A-driven; the plan needs continuous accretive deployment via debt+equity at the lowest rung of investment grade. A power-price downturn + higher rates could force a choice between the dividend, the growth capex, and the rating — and a downgrade would break the compounder model.
3. The AFFO base is inflated. FY2026 guided AFFO (~$950M) is BELOW FY2025 actual ($1,066M) — the market re-rated off a 2025 AFFO/share ($7.08) that management itself doesn't expect to repeat in 2026. If investors are anchoring on $7.08, the stock is more expensive than it looks.

Pre-mortem (18 months out, thesis broke): It's early 2028. Alberta pool prices never recovered under the new REM regime and the offer cap; the data-center ESA counterparty delayed to 2030 (or paid the termination fee and walked); PJM capacity cleared soft; and refinancing 2026-vintage notes at higher rates squeezed AFFO. AFFO/share is stuck near $5.5, the 2030 plan is quietly reset, the dividend is frozen, and the "AI power" premium (~16× EV/EBITDA) compressed toward TransAlta's ~14× and below — a 20–30% de-rating even with flat fundamentals.

Are multiples too high? On the "AI power" narrative, yes — ~16× EV/EBITDA already prices the good outcome. On normalized ~10–12× P/AFFO with a 4.5% yield, it's fair-to-full, not cheap. There is little margin of safety in the multiple; the margin of safety has to come from the cycle (buying when Alberta prices are washed out) or from a contract miss de-rating the stock.

Contrarian view (what the market refuses to see): Consensus treats CPX as a clean "AI power" long. The market is under-weighting that Alberta — 40% of the base — may be entering a multi-year structurally-lower price regime by policy design, and that 2025's record AFFO is a local peak, not a launchpad (per the company's own 2026 guide). The AI-demand upside is real but back-end-loaded to 2028+; the Alberta price and AFFO-normalization risk is now. The asymmetry near a 52-week high is unattractive.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• What structurally breaks the money machine? Merchant power prices. This is a price-taker on both its input (gas) and its output (wholesale power) across a fleet that is now more merchant (gas) and less contracted (sold renewables). The spark spread is the whole business, and in Alberta it just compressed by half.
• Where is revenue concentrated / what if it shifts? Two regimes carry it: Alberta (~40%, energy-only, price-capped, oversupplied) and US PJM (~60%, capacity-auction-dependent). Both are policy-set markets — an AESO offer cap or a soft PJM capacity auction hits EBITDA directly, and neither is in management's control.
• Why is the moat weaker than bulls think? Because it's a cyclical scarcity moat, not a franchise. 2024 Alberta proved the "moat" evaporates the instant reserve margins loosen. New gas build by anyone competes it away; the only durable edge is Genesee's site/cost position.
• Most dangerous competitor bulls underestimate: not another IPP — it's new supply itself (the ~3,000 MW that crushed Alberta prices in 2024, and whoever else chases the same data-center demand). In PJM, the deep-pocketed Vistra/Constellation/NRG can outbid for assets and contracts. And the data-center developers can self-generate or sign with a lower-cost counterparty.
• Worst capital-allocation moves / incentives: rotating into commodity risk near a possible cycle peak; ~0.003% CEO ownership against ~$5.4M comp; growth that leans on financial engineering (Apollo fees, BBB- leverage) to manufacture the 8–10% AFFO/share target.
• What must hold for today's ~$73 price? (i) Alberta prices normalize off 2024 lows; (ii) the data-center ESA (and follow-ons) actually contracts and starts; (iii) Apollo deploys accretively; (iv) BBB- holds through the build-out; (v) investors keep paying a growth multiple on a merchant fleet.
• If growth disappoints by 20–30%: AFFO/share stalls near the 2026 guide (~$6) or below; the ~16× EV/EBITDA compresses toward peers' ~14×; 20–30% downside to ~$50–55 (right at the 52-week low and the low-street target of C$52.5).
• Single permanent-impairment scenario & plausibility: Alberta becomes a structurally price-capped, oversupplied market under REM while the AI-demand wave routes around Capital Power (self-generation / competitor contracts). Plausibility: moderate — not a base case, but far from tail; the 2024 price collapse and the deliberate offer-cap policy make it credible.

Lens 14 · Management Questions (ordered by information value)

1. FY2026 guided AFFO (~$950M) is below FY2025 actual ($1,066M) — decompose exactly what in 2025 does not recur (realized trading/optimization gains, one-time items, timing), and what is the true organic, repeatable AFFO run-rate?
2. What is the net-debt/EBITDA today and the ceiling you will not cross to defend BBB-, and at what power-price scenario does the growth capex compete with the dividend?
3. Give the Alberta pool-price and AESO REM/offer-cap assumptions embedded in the 8–10% AFFO/share-to-2030 plan — what does the plan look like if Alberta averages <$60/MWh through 2028?
4. On the 250 MW data-center ESA: who is the counterparty (credit), what is the termination fee, contracted price/escalator, and what is the realistic probability and timing it converts and starts in 2028?
5. What is the organic vs. acquired split of the 8–10% AFFO/share target — how much depends on Apollo deploying the full US$3B, and what return hurdle governs each deal?
6. What are the realized spark spreads and hedge % on Hummel and Rolling Hills, and how exposed is their EBITDA to PJM capacity-auction clearing prices over the next three auctions?
7. Why is direct CEO/insider ownership so low (~0.003%), and will management commit to a meaningful open-market equity build to align with the bet-the-balance-sheet strategy?
8. You walked from the $2.4B Genesee CCS project — under what carbon-policy scenario does gas-fleet emissions/carbon-cost liability become a material AFFO drag (federal OBPS, future carbon prices)?
9. What is the maintenance vs. growth capex split over 2026–2030, and the expected sustaining AFFO if you stopped all growth M&A tomorrow?
10. How much of the Apollo economics is fee income vs. equity returns, and how durable/aligned are those fees if the merchant-gas assets underperform?
11. What is the refinancing schedule and rate assumption for the 2025-vintage notes, and the AFFO sensitivity to a +150 bps refinancing shock?
12. Renewables sell-down: is the intent to fully exit wind/solar/storage, or keep a strategic tail — and does exiting contracted cash flow to add merchant gas raise the group's cost of equity?
13. In PJM, what stops Vistra/Constellation/NRG or the data-center developers themselves from out-competing you for the next tranche of assets and contracts?
14. What is the downside scenario for the dividend (12-year growth streak now guided to only 2–4%) if power prices stay at 2024 levels for two more years?
15. What would have to be true for you to conclude the merchant-gas rotation was mistimed, and what is your pre-committed response (deleverage, re-contract, pause M&A)?

---