KeyCorp (NYSE: KEY) — A Repaired Laggard Priced for the Promotion It Hasn’t Earned

Independent equity research. Report date: 2026-06-21. The analysis below takes no position; the single opinion is fenced in the author’s-view block.

⚡ Claude’s Take

This block is the author’s own independent opinion and general information — not investment advice. It is deliberately the one place a position is taken; the analysis that follows (sections 1–15) takes no position and contains no price target.

Verdict: HOLD / AVOID-here at ~$22.59; accumulate-on-weakness in the high-teens (~$18–20, roughly 1.3–1.45x tangible book). Not-a-short. Conviction: medium.

KeyCorp is a genuinely repaired bank wearing the valuation of a genuinely good one. The operating story is real and I do not dispute it: net interest margin troughed at 2.17% in Q3-2024 and has climbed every quarter since, the de-risked securities book and fixed-rate asset repricing are mechanical tailwinds that roll on largely regardless of the rate path, AOCI is burning back into tangible book, and management’s “record NII” 2026 guide (+9–10%) is credible. The problem is entirely price for quality. At ~1.66x tangible book and the 99.7th percentile of its own decade of price-to-book — the richest KEY has ever been on book value — the stock embeds a sustainable ~14.5–16% return on tangible common equity. That is management’s end-2027 target (ROTCE 15%+, NIM 3.25%+) priced as already achieved and permanent. The realized number is ~11.9% — the lowest ROTCE in the entire super-regional cohort. You are paying a 16%-bank multiple for a 12%-bank, on the bet that the gap closes on schedule. Regions earns ~600bp more ROTCE at only ~0.5x more P/TBV; Huntington earns more at a lower P/TBV. KEY is the expensive way to own this trade.

The framing is crowded, high-momentum recovery — not a falling knife and not a value entry. The stock has roughly doubled off its June-2025 low, sits ~3% below its all-time relative-strength peak (y1 total return ~+47%, beta 1.20), and the easy mean-reversion money has been made. Layered on top is a live technical overhang: Scotiabank, which bought ~14.9% at $17.17 near the 2024 trough, filed a 158.7-million-share resale shelf on June 5, 2026 and has been selling into strength at $21–22 — the “strategic partner” is exiting ~18 months after entry while common holders absorbed the ~17% dilution. No executive has bought a share in the open market. The base case is roughly in the price; the bull (full target lands) offers maybe 10–15% — where the Street’s ~$26 targets sit — while the bear (target slips, NIM proves a rate-cut-dependent peak, or a recession turns the commercial-heavy book) compresses both the richest-ever multiple and the earnings. That asymmetry is unattractive here. I would own this lower, not here. Tag: “Sold its balance sheet to fix it — now priced as if it never needed to.”

• What flips me bullish: ROTCE convergence to ~15% ahead of the 2027 schedule with the NIM holding ≥3.2% on structural (not rate-cut) repricing — plus a clean, orderly Scotiabank exit that lifts the overhang. That would make ~1.66x TBV defensible rather than aspirational.
• What flips me bearish: a cut to the NIM/NII guide or ROTCE stalling at 12–13%, which would expose the richest-ever multiple as built on a target that isn’t arriving — or a credit turn in the C&I/CRE book given the 1.20 beta.

📈 Stock Price Action — Five-Year Event Map

KeyCorp has round-tripped from a ZIRP-reflation high to a regional-banking-crisis trough and back to fresh five-year highs. The trailing-60-month arc: a 2022 high near $27 → the SVB-crisis trough in May-2023 (intraday unadjusted low $8.54; dividend-adjusted close basis ~$7.50) → a recovery low of ~$15.36 (adjusted) / $16.01 (unadjusted) in June-2025 → an all-time-context high of ~$22.98 close (intraday $23.35) on 9-Feb-2026 → $22.59 today (18-Jun-2026). The stock sits ~3% below its closing high, near the very top of its five-year range; the 52-week range is ~$15.36–$22.98; beta 1.20, alpha +0.12. It has roughly doubled off the June-2025 low. (Prices are facts from the AZI five-year CSV; the $7.50-vs-$8.54 gap is the dividend-adjusted-close vs raw-intraday basis — flagged so the figure is unambiguous.)

Cycle narrative. (1) 2021 reflation lifted every regional on curve-steepening hopes. (2) In 2022 the hikes that were supposed to help instead crushed multiples on recession fear and a growing AOCI hole. (3) March-2023’s SVB run was an existential repricing of any bank with a large mark-to-market loss and rate-sensitive deposits — KEY’s structurally low-NIM, oversized-low-yield-securities profile made it one of the worst-hit large regionals (intraday $8.54). (4) Surviving without a deposit run let the stock nearly double back by 2024. (5) The Scotiabank stake was the explicit capital vote of confidence and the war-chest for de-risking. (6) In Q4-2024 KEY crystallized the bond losses — the GAAP loss is this, not operations. (7) 2025–26 is the NII-inflection re-rate: the stock doubled off the mid-2025 low as the repositioned book and asset repricing drove “record NII” guidance. (8) It now sits near five-year highs with a fresh technical overhang — Scotiabank’s 158.7M-share resale registration. (Sources: AZI price CSV; KEY 8-K earnings timeline; S-3ASR 2026-06-05.)

1. Executive Summary

KeyCorp is a ~$190 billion-asset super-regional bank, holding company for KeyBank N.A. (Cleveland, founded 1849; 940 branches across 15 states), operating through two segments — Consumer Bank and Commercial Bank. It is, on the evidence, a structurally below-average operator in a structurally mediocre industry, currently enjoying a real but cyclical and sector-wide recovery, and trading at the richest valuation-on-book in its own recorded history.

The numbers that matter, all reconciled to the FY2025 10-K. KEY runs the lowest net interest margin in the super-regional cohort (~2.69% continuing-ops / 2.82% reported FY2025, vs RF ~3.5%, FITB/HBAN ~3.0%+), the worst efficiency ratio (~62.6% vs FITB ~54%, RF ~57%), and the lowest return on tangible common equity (~11.85%, vs FITB/RF/HBAN 16–18%) — it barely clears its cost of equity in a recovery year. Its single genuine differentiator is an above-cohort fee mix (~38% of revenue), anchored by a punch-above-its-weight investment-banking/M&A-advisory franchise (KeyBanc Capital Markets, the Cain Brothers healthcare boutique; IB & debt-placement fees $780M in 2025, +13.5%).

The two-year arc is a genuine repair executed at a real cost. A 2023 regional-bank-crisis scare exposed an oversized, low-yield ZIRP-era securities book; in 2024 KEY recapitalized via a ~$2.8B, ~14.9% strategic minority investment from Scotiabank (~163M shares at $17.17, below tangible book, near the trough), used the headroom to sell ~$10B of low-yield mortgage securities at a ~$1.8B pre-tax loss (the entire cause of the FY2024 GAAP loss of −$0.32), and is now earning the higher reinvestment yield — the direct engine of the 2025 NIM/NII inflection. Diluted EPS recovered $2.63 (2021) → $0.88 (2023) → −$0.32 (2024) → $1.52 (2025), and the 2026 guide is for “record NII,” up 9–10%, with a medium-term (end-2027) target of ROTCE 15%+ and NIM 3.25%+.

The investment tension is price, not operations. At $22.59 the stock trades at ~1.66x tangible book — the 99.7th percentile of its own ten-year price-to-book range — which on a standard franchise multiple embeds a sustainable ~14.5–16% ROTCE. That is the 2027 target priced as done and permanent, against a realized ~11.9% that is the lowest in the peer group. The “12.7x P/E / 3.6% dividend yield” optics suggest value; the P/TBV-versus-ROTCE math says the opposite. Reinforcing the caution: a fresh ~14% resale overhang as the strategic investor exits, no executive open-market buying, and a crowded high-momentum positioning near a relative-strength peak. Capital allocation rates as competent crisis management (dividend held through the stress; buyback restarting to ≥$1.3B; a genuinely ROTCE/EPS/relative-TSR-aligned comp plan that paid zero on 2023–25 awards) rather than skilled value creation (the dilution-at-the-trough was a self-inflicted clean-up). No recommendation and no price target follow in this body; the embedded-expectations analysis is in Section 10.

2. Business Overview

What it is. KeyCorp is the bank holding company for KeyBank National Association, a ~$190 billion-asset super-regional headquartered in Cleveland, Ohio, founded in 1849. Its retail footprint is 940 full-service branches and ~1,120 ATMs across 15 states (concentrated in the Northeast, Midwest, and Pacific Northwest, with commercial reach nationally). It reports in two segments: Consumer Bank and Commercial Bank. Chairman & CEO Chris Gorman leads; Clark Khayat is CFO (and, as of Q1-2026, additionally leads the Technology & Operations organization); Mohit Ramani is Chief Risk Officer.

Segment economics (FY2025, taxable-equivalent). The Commercial Bank is the engine and the more profitable, fee-rich half. Commercial Bank revenue was $4,039M (NII-TE $2,294M + noninterest income $1,745M — i.e., fees are 43% of its own revenue), noninterest expense $1,905M, provision $299M, and net income to KeyCorp of $1,447M — roughly 63% of the company’s total. Consumer Bank revenue was $3,666M (NII-TE $2,709M + fee $957M, fees 26% of revenue), noninterest expense $2,802M, provision $169M. The picture is a commercially-tilted bank whose growth, profitability, and fee differentiation all sit in the Commercial segment, with the Consumer Bank a lower-margin deposit-gathering and consumer-lending base.

Revenue composition. Total FY2025 revenue (TE) was ~$7,513M: net interest income (TE) $4,671M (GAAP $4,636M) at a 2.69% continuing-ops NIM, plus total noninterest income of $2,842M. That puts fee income at ~37.8% of revenue — high for a regional (peers typically 25–35%). (FY2024 fee income of $809M was artificially depressed by the Q4-2024 securities-repositioning loss; $2,842M is the clean run-rate.)

Fee mix (FY2025, $M). Investment banking & debt-placement fees $780 (+13.5% YoY, the largest single non-spread line); trust & investment services $591; cards & payments $337; service charges on deposits $295; corporate services $294; commercial mortgage servicing fees $287; corporate-owned life insurance $140; consumer mortgage $58. The standout, differentiated lines are IB/M&A advisory and commercial mortgage (special-)servicing.

Deposits. Total deposits ~$148.7B (interest-bearing $121.1B + noninterest-bearing $27.6B). Noninterest-bearing deposits are ~18.6% of the total — a relatively low share that contributes to KEY’s rate-sensitive funding profile. Average interest-bearing-liability cost fell to 2.78% (FY2025) from 3.39% (FY2024); that funding-cost relief was a primary driver of the NIM recovery.

Loan book ($106.5B, continuing ops, 12/31/25). Commercial & industrial $57,688M (54.1%); commercial real estate $16,551M (15.6%; commercial mortgage $13,707M + construction $2,844M); commercial lease financing $2,270M (2.1%) — total commercial 71.8%. Consumer: residential mortgage $18,732M (17.6%); home equity $5,703M (5.3%); other consumer $4,644M (4.4%); credit cards $953M (0.9%) — total consumer 28.2%. A C&I-heavy, commercially-tilted book.

Strategic notes. The Laurel Road education-lending business is now accounted for as discontinued operations (the 10-K defines “continuing operations” as everything other than education lending; ~$205M residual loans remain) — the once-touted digital student-refinance differentiator is being wound down. Assets under management/administration were $70.0B (up from $61.4B in 2024) — modest for the asset base; wealth is a growth priority, not yet a dominant earner.

Recurring vs. cyclical. Spread income (NII, ~62% of revenue) plus payments, deposit service charges, and trust/wealth fees are recurring. The swing factor is the ~$780M (≈10% of revenue) investment-banking/debt-placement line plus commercial-mortgage-servicing gains, which move with the deal and capital-markets cycle. KEY’s fee mix is higher-quality than a pure spread bank’s, but its cyclicality concentrates in exactly the line that gives KEY its quality argument.

Verdict. A commercially-oriented super-regional whose profit center and fee differentiation are in the Commercial Bank, funded by a relatively rate-sensitive deposit base, with a meaningful and genuinely above-cohort capital-markets fee franchise offset by a low-yield core spread business. The model is coherent but not advantaged — the differentiation is cyclical, and the consumer-lending growth bet (Laurel Road) has been abandoned.

3. Industry Dynamics

Structure. US super-regional banking ($100–500B in assets) is a fragmented oligopoly squeezed between the four money-center megabanks (JPM/BAC/WFC/C — with funding-cost, technology-spend, and brand scale advantages) and thousands of community banks. KeyCorp’s peer cohort is FITB, RF, HBAN, CFG, MTB, USB, TFC, and PNC. It is fundamentally a commodity industry: deposits and loans are largely undifferentiated and the product is money. Pricing power on the core spread business is structurally absent; differentiation lives in fee businesses, local relationship density, and cost/funding scale.

Post-2023-crisis aftermath. The March-2023 failures of SVB, Signature, and First Republic permanently sharpened the market’s focus on (a) deposit stability and uninsured-deposit concentration, (b) AOCI marks on held-to-maturity and available-for-sale securities, and © liquidity. The whole cohort had over-bought low-yield securities in 2020–21 and spent 2023–25 working off the embedded losses; KEY was among the more exposed, and its ~$10B securities-repositioning loss (the cause of the 2024 GAAP loss) was a direct response to that legacy. KEY’s structurally low NIM partly reflects that oversized, low-yielding securities book.

Rate / NIM environment. The cohort’s NIM troughed in 2023–24 as deposit betas peaked, then began recovering in 2025 as the Fed eased and fixed-rate assets (securities, swaps, fixed loans) repriced upward. KEY’s NIM rose 2.16% (2024) → 2.69% (2025), guided to a 4Q-2026 exit of 3.00–3.05% and a medium-term (end-2027) target of 3.25%+. Crucially, this is a sector-wide tailwind — asset repricing plus lower deposit costs — not KEY-specific alpha. Every regional is reporting the same NIM recovery, which is why the entire group has roughly doubled off mid-2025 lows and trades near multi-year-high price-to-book.

Credit cycle. Office/CRE remains the cohort’s watch-item. KEY’s CRE is 15.6% of loans ($16.6B) and is more diversified-by-type (data-center, industrial, multifamily, skilled-nursing) than office-concentrated; charge-offs are normalizing (41bps, guided 40–45bps for 2026), not spiking, and leading credit indicators improved in 2025. Manageable, not a 2008-style impairment, but a live tail risk for the group.

Regulation. The Basel III “endgame” capital rules were materially softened and re-proposed through 2024–25. As a sub-$250B “Category IV” / standardized-approach bank, KEY faces a lighter regime than Category I–III banks; the revised proposal is in fact a modest tailwind — management’s preliminary estimate is a ~9% RWA reduction (~+100bps to marked CET1). The live swing item is potential AOCI inclusion in regulatory capital for Category III/IV banks, but KEY has already de-risked its book, mitigating that exposure. Regulation has shifted from a 2023–24 headwind to a 2026 modest tailwind.

Consolidation wave. 2025–26 has seen accelerating regional consolidation (e.g., the FITB/Comerica combination), the cohort’s structural answer to sub-scale economics and rising tech/compliance spend. Scale matters more every year; sub-scale regionals are increasingly acquirers, targets, or slowly disadvantaged. The Scotiabank stake in KEY is itself a consolidation/optionality signal.

Capital cycle (Marathon lens). The cohort sits in the recovery/early-expansion phase on the earnings side (NIM troughed, ROEs rebuilding) but capital has already flooded back — the group has roughly doubled off mid-2025 lows and trades near multi-year-high P/B (KEY P/B at the 99.7th percentile of its own history). In Marathon terms, high prospective returns are already attracting capital (rising bank-stock prices, M&A, the Scotiabank injection); the supply-side signal is that the easy mean-reversion trade is largely done. There is no evidence of lending-capacity over-build yet — loan growth is muted (KEY’s average loans actually fell from $107.7B to $105.7B in 2025) — which is cycle-supportive for credit but means organic volume growth is scarce.

Verdict: structurally mediocre-to-poor industry. Commodity product, no pricing power on the core spread business, returns capped near cost of equity through-cycle for sub-scale players, perpetual regulatory and credit tail risk, and a rising scale premium that disadvantages mid-tier names. The current NIM-recovery tailwind is real but cyclical, sector-wide, and largely priced. A good operator earns an acceptable (not exceptional) return here; a sub-scale one struggles to clear its cost of equity across the cycle.

4. Competitive Position

Greenwald moat taxonomy: weak, local versions of all three genuine advantages — i.e., no durable moat. (1) Customer captivity / switching costs: real but modest — primary-checking and operating-account relationships are sticky, but the deposit franchise is undifferentiated and contestable. (2) Economies of scale + captivity: KEY is sub-scale relative to the megabanks on the two scale-sensitive cost buckets — technology spend and deposit-gathering cost — and is not the dominant share-leader in most of its metros. (3) Cost advantage: none — KEY’s funding cost is structurally higher and its efficiency ratio worse than best-in-class peers. The net is a collection of local relationship franchises plus one genuinely above-weight fee business, not a broad durable moat.

The structurally low NIM is the financial fingerprint of no cost/funding advantage. KEY’s NIM of 2.69% (FY2025), recovering from ~2.16% (2024) and ~2.0–2.2% (2023), is among the lowest in the super-regional group. The drivers: an asset mix skewed to lower-yielding C&I and a historically oversized, low-yield securities book built in the ZIRP era; a commercial-heavy book (72% commercial) with thinner spreads than consumer/card-heavy peers; and higher reliance on rate-sensitive funding (noninterest-bearing deposits only ~18.6% of the total). Peers run 30–80bp higher.

Returns and efficiency lag the cohort on every core metric. The peer scorecard is unambiguous:

KEY barely clears its ~10–11% cost of equity even in a recovery year — the single most damning competitive fact. It leads only on fee diversification.

The offsetting strength: an above-weight, but cyclical, fee/capital-markets franchise. Fee income is ~38% of revenue, above the cohort. The crown jewel is the investment-banking / M&A-advisory / debt-placement engine run through KeyBanc Capital Markets — IB & debt-placement fees were $780M in 2025 (+13.5%), the largest single non-spread line. KEY built sector-specialty M&A boutiques (notably Cain Brothers in healthcare, plus industrials/technology coverage) that compete for middle-market mandates against far larger firms. This is genuinely differentiated and the reason KEY’s franchise is worth more than a pure spread bank of its size — but it is not a Greenwald moat: advisory talent is mobile, competitors run parallel boutiques, and the revenue swings with the deal cycle.

The Scotiabank relationship is capital and optionality, not advantage. Scotiabank’s ~14.9% stake provided welcome capital (it funded the 2024 de-risking) and is a vote of confidence, but it confers no operating cost advantage — no cheaper funding, no material referral engine evidenced in the 10-K — and is described purely as an equity holder with influence rights. The June-2026 resale prospectus for 158.7M shares signals a potential unwind/overhang, not a deepening strategic moat.

Share-stability / ROIC test (Greenwald). KEY’s deposit/loan share is roughly stable but not expanding (average loans fell in 2025); there is no evidence of the share-gain-without-price-war pattern that signals a real moat, and ROIC/ROE hover at-or-near the cost of capital — the textbook signature of a business with no durable competitive advantage.

Verdict: sub-scale, me-too regional with no durable moat, partially redeemed by an above-weight but cyclical capital-markets fee franchise. KEY’s deposit relationships give it the weak customer-captivity every bank has; it lacks the scale, funding, and cost advantages that would let it out-earn peers, and the numbers prove it — lowest NIM, worst efficiency, lowest ROTCE in the cohort. The KBCM/Cain Brothers IB engine is genuinely differentiated and the reason KEY is not merely a worse FITB, but it is mobile-talent-dependent and deal-cycle-sensitive, not a moat. Be direct: a below-average operator in a mediocre industry, earning a cyclically-recovering but still cohort-trailing return.

5. Growth History and Forward Opportunities

Historical growth has been cyclical and, recently, negative on volume. Per-share earnings tell the cycle, not a growth story: diluted EPS ran $1.27 (2020) → $2.63 (2021, COVID reserve-release peak) → $1.93 (2022) → $0.88 (2023, NIM compression) → −$0.32 (2024, securities loss) → $1.52 (2025). Revenue (TE) has oscillated in a ~$7.0–7.5B band for five years with no secular up-trend. Critically, average loans fell in 2025 (from $107.7B to $105.7B) as management ran off low-yielding non-relationship consumer loans — the recovery in earnings is a margin/repricing story, not a volume story.

The forward opportunity set is real but modest and mostly cyclical/sector-wide.

• NII/NIM repricing (the largest lever, already underway): the de-risked securities book, fixed-rate asset and swap roll-off into higher yields, and a mix shift toward higher-yield commercial lending drive the FY2026 “record NII” guide (+9–10%) and the path toward the 3.25%+ NIM target. This is durable but cyclical and shared across the cohort.
• Commercial loan growth: the 2026 commercial loan guide was raised to +6–8% (period-end C&I +5% in Q1-26) — genuine organic growth, concentrated in the higher-spread segment, but off a base that is otherwise running off consumer loans (total avg loans guided only +1–2%).
• Fee/IB build-out: KEY grew its frontline sales force ~10% in 2025 across IB, wealth, and payments; the priority fee businesses (Wealth, IB, Commercial Payments) collectively grew ~12% YoY in Q1-26. The mass-affluent wealth push (a re-sized ~1.15M-household opportunity, <10% penetrated) is the clearest structural growth vector. But the IB recovery KEY’s differentiation rests on is underwhelming — management guides only +5–6% IB fees for 2026 despite “record pipelines,” and middle-market M&A has been a repeatedly-deferred catalyst.
• Payments and treasury management: embedded, recurring, scale-sensitive — a steady grower, not a needle-mover.

Quality of growth. The 2025–26 earnings recovery is high-quality in its durability (structural asset repricing) but low-quality as growth — it is margin recovery off a depressed base and a cyclical/sector-wide rate tailwind, not share gains, volume expansion, or a widening moat. The one secular vector (mass-affluent wealth + fee build-out) is early and small relative to the $190B balance sheet, and the differentiated IB engine keeps deferring its inflection.

Verdict: low-quality growth dressed as recovery. The earnings rebound is real and the NII trajectory is credible, but it is cyclical margin recovery plus modest commercial-loan growth, not durable secular expansion. Investors paying a richest-ever multiple are underwriting the recovery completing, not a high-growth franchise.

6. Financial Quality

The NIM/NII inflection is the core story — real and largely durable, but off a laggard base. NIM troughed at 2.17% in Q3-2024 and climbed monotonically: Q1-25 2.58% → Q2 2.66% → Q3 2.75% → Q4 ~3.00–3.05% exit-rate → Q1-26 2.87%. Full-year FY2025 NIM was 2.82% (reported basis) / 2.69% (continuing-ops basis, +53bps YoY). NII (TE) reached $4,671M in FY2025, up ~$861M (~+23%) YoY; Q1-2026 NII (TE) of $1.23B was +~11% YoY. The inflection drivers per management: lower interest-bearing deposit costs (proactive deposit-beta management as rates fell), reinvestment of maturing low-yield securities and fixed-rate loans/swaps into higher-yielding assets, the H2-2024 repositioning, and a mix shift to higher-yield C&I. The FY2026 guide is NII (TE) up 8–10% (“record NII”), a 4Q-2026 NIM exit of 3.00–3.05%, with a medium-term (end-2027) target of NIM 3.25%+ and ROTCE 15%+. The recovery rests on durable mechanics (locked-in repositioned book; fixed-rate roll-off) plus a deposit-cost-relief component that is partly rate-cut-dependent — mostly structural, partly rate-path-sensitive.

PPNR is the cleanest tell that the 2024 loss was a securities artifact, not an operating collapse. Pre-provision net revenue (continuing ops) was just $74M in 2024 — crushed by the repositioning loss inside noninterest income — versus $2,810M in 2025. The operating engine never broke; the 2024 GAAP loss was a balance-sheet clean-up booked through the income statement.

Efficiency and returns lag. FY2025 noninterest expense of $4,703M on TE revenue of $7,513M is an efficiency ratio of ~62.6% — materially worse than FITB (~54%) and RF (~57%). ROTCE (continuing ops) was ~11.85%; ROE 10.4%. KEY is the low-NIM, fee-reliant, lower-return name in the cohort.

Credit is benign and well-reserved. Provision for credit losses was $471M (2025) vs $335M (2024) — but the 2025 increase is a forward reserve build (economic uncertainty + commercial loan growth), not realized deterioration: net charge-offs were flat at 41bps (guided 40–45bps for 2026), and coverage rose — ACL/period-end loans held at 1.63%, while ACL/nonperforming loans improved to 282.9% (from 224.1%). Leading indicators (NPAs, criticized loans, delinquencies) all moved favorably in 2025. CRE/office is the watch-item but is diversified-by-type and currently benign.

Capital and the AOCI recovery. CET1 was 11.78% and Tier 1 13.46% at 12/31/2025 — comfortably above the minimum-plus-buffer stack; the Stress Capital Buffer is 3.20% (effective 10/1/2025). AOCI improved from −$3,470M (12/31/24) to −$1,960M (12/31/25) — a ~$1.51B recovery that directly accretes tangible book; total equity rose to $20,381M from $18,176M. But per-share tangible book growth is muted by the Scotiabank dilution: TBVPS was $13.76 (2025) vs $13.58 (2024), and book value per share actually fell from $16.68 to $15.13 on the ~17%-larger share count.

Quality of earnings.

• The 2024 GAAP loss is a one-time, economically-rational artifact. The entire −$304M loss to common (−$0.32 diluted) stemmed from selling ~$10.0B of lower-yielding MBS across Q3+Q4 2024 at ~$1.8B of pre-tax losses, dragging FY2024 noninterest income to $809M. The transaction de-risked the book and lifted go-forward reinvestment yield — the direct cause of the 2025 NIM inflection — and was enabled/cushioned by the Scotiabank capital. It is a QoE non-event for run-rate earnings power.
• 2025 reported EPS is clean — arguably slightly understated. FY2025 had no large securities loss, no Visa-type gain, and a reserve build (not a release). The FDIC special assessment was a small headwind that is winding down. GAAP diluted EPS of $1.52 carries ~$0.14 of discontinued-ops (education-lending run-off) drag versus $1.66 from continuing operations; the forward run-rate (Q1-26 diluted EPS $0.44, annualizing to ~$1.76 and rising) is the better earnings-power anchor.
• The genuine QoE blemish is dilution. Average diluted shares jumped from ~933M (2023) to ~1,108M (2025) on the Scotiabank issuance — ~17% — issued cheaply at $17.17 near the trough. Per-share earnings power must now clear a permanently higher share count; the 2027 ROTCE 15%+ target is the test of whether the dilution is earned back.

Verdict: economics are improving and the recovery is durable — but from the lowest base in the cohort. The inflection rests on structural asset repricing plus some rate-cut sensitivity; the “record NII” guide and 3.25%+ NIM target are credible extensions of a real trend. The 2024 loss is a clean-up artifact, not deterioration, and credit is benign. But KEY remains the lowest-NIM, lowest-ROTCE, worst-efficiency operator in its peer group — improving and durable, yes, but a recovering laggard, not a quality leader.

7. Capital Allocation

Dividend: protected through the stress — at a trough payout that helped necessitate the raise. The common dividend held flat at $0.205/quarter ($0.82/year) across 2023, 2024, and 2025; KEY did not cut through the 2023 crisis or the 2024 GAAP loss — a genuine durability signal (current yield ~3.6%). But the cost rose: total common dividends declared went from $768M (2024) to $911M (2025) — a ~19% increase in dollars with no per-share raise — because the Scotiabank issuance added ~163M shares. And the timing was uncomfortable: in the 2023 trough, NI-to-common of $824M against $768M of dividends was a ~93% payout; in 2024 the dividend was un-covered by GAAP earnings (paid out of PPNR/capital, cushioned by Scotiabank). It was only comfortably re-covered in 2025 (~54% payout). Maintaining the income signal was defensible, but it consumed capital exactly when the balance sheet needed it — part of why the Scotiabank capital was required.

Buyback: suspended through the rebuild, now restarting and ramping. Repurchases were effectively suspended during 2023–24. On March 13, 2025 the Board authorized a new $1.0B program; KEY began repurchasing only in Q4-2025. In Q1-2026 it bought back ~$400M (“well in excess of the $300M-plus commitment”), explicitly “taking advantage of the pullback in regional-bank stock prices,” and raised the 2026 full-year buyback guide to “at least $1.3 billion” (the CFO called $1.3B “more like the floor”), with ≥$300M/quarter for the balance of the year. The restart confirms the capital rebuild is complete (marked CET1 reached its ~10% target a few quarters early). Note the sequencing irony: KEY issued ~163M shares at ~$17 at the trough and is now buying back at ~$22 — it sold low and is repurchasing higher; the buyback is accretive versus today’s price only if the stock is below intrinsic value, and it does not recover the dilution round-trip.

The Scotiabank investment — the defining capital event; prudent recapitalization at a destructive price. Per the Investment Agreement dated August 12, 2024, Bank of Nova Scotia made a ~$2.8B strategic minority investment for ~14.9% pro-forma ownership at a fixed $17.17/share, in two tranches: 47,829,359 shares (~$821M, 4.9%) closed Aug 30, 2024, and 115,042,316 shares (~$2.0B) closed Dec 27, 2024 after regulatory approvals (~162.9M shares total; $26M issuance costs). Scotiabank may designate up to two directors (and currently has two on the Board) and “may be able to influence our policies.” The honest judgment: this was prudent recapitalization executed at a bad price, not reckless value destruction. It was genuinely needed — it provided the CET1 headroom to crystallize ~$1.8B of pre-tax AFS losses on ~$10B of MBS without breaching capital targets, the single transaction that drove the 2025 NIM inflection. But it was permanent, ~17%-dilutive equity sold below tangible book near the cycle/valuation trough — the most expensive form of capital at the worst time. Management cleaned up a balance-sheet problem it had created (the oversized ZIRP-era bond book) and paid for the clean-up by handing ~15% of the company to a strategic investor cheaply. Defensible crisis management; not value-additive allocation.

And the “strategic” partner is already exiting into strength. Scotiabank filed 19 disposition Form 4s (Dec-2025 through Jun-2026) selling at ~$21–22, and on June 5, 2026 KEY filed an S-3ASR registering 158,723,874 shares for the selling shareholder. The stake is reverting to a financial exit ~18 months after entry — Scotiabank monetizing its trough-priced shares at a ~25–30% gross gain while common holders absorbed the dilution. It confirms the stake was capital and optionality, not a durable operating partnership, and it creates a real technical overhang near the 52-week high.

M&A: disciplined, bolt-on only. KEY has not done large bank M&A in this window. Per Gorman’s Q1-26 capital priorities: (1) support organic/client loan growth, (2) invest in the business (people + ~$1B/year technology), (3) the dividend, (4) buybacks. Inorganic activity is limited to small boutique acquisitions and team lift-outs (historically Cain Brothers; recent middle-market and family-office hires). A positive posture after the dilution episode — no empire-building — but also no scale-fixing transformative deal in a scale-disadvantaged industry.

Executive compensation — a genuine, and rare, positive. KEY’s incentive plans are tied to per-share / return-on-capital metrics, not asset growth or size. The Long-Term Incentive (70% performance-based) uses Adjusted ROTCE and Cumulative Adjusted EPS with a ±15% relative-TSR modifier; the Annual Incentive is 65% Adjusted PPNR / 35% Adjusted ROTCE plus a Cash Efficiency Ratio gate and strategic priorities. Critically, the plan works: the 2023–2025 Performance Awards paid zero because financial performance was below threshold — the design denied payout when returns disappointed. 86% of average NEO target pay is at-risk; ownership guidelines are 6x salary (CEO)/3x (others); the Committee normalizes AOCI swings out of ROTCE goals. This is exactly the per-share governor whose absence is a recurring demerit across this coverage set. Caveat: ROTCE-based comp did not prevent the trough dilution (ROTCE is a ratio achievable on a larger share base), and the plan changes followed shareholder pushback on a prior Say-on-Pay outcome — aligned design, recently improved under pressure.

Insider alignment is weak. No executive (Gorman/Khayat) made open-market purchases in 2024–2026; the only code-P buys were token director-qualifying purchases near the 2025 low (largest ~$137K). The dominant insider dynamic is Scotiabank’s persistent exit. No conviction C-suite buying.

Verdict: mixed, leaning adequate — competent crisis management, not skilled value creation. The good: dividend protected (now well-covered at ~54%); disciplined bolt-on M&A; ~$1B/year tech investment; a buyback restarting and ramping to ≥$1.3B with explicit valuation-opportunism; and a genuinely ROTCE/EPS/relative-TSR-aligned comp plan that paid zero when returns missed (the standout bright spot). The bad, and it is material: management built the oversized low-yield bond book, then recapitalized via ~17%-dilutive equity sold to Scotiabank below tangible book at the trough — and that partner is now exiting at $21–22 having captured the rebound. Capital allocation rates as repaired-but-unproven; the 2027 ROTCE 15%+ target is the test of whether the dilution is earned back.

8. Changes and Headwinds — Last Two Years

The two-year arc: crisis → recapitalization → recovery → overhang.

• 2023 — stress. KEY bottomed at ~$7.50 (adjusted) in May-2023; its oversized low-yield ZIRP bond book was a relative weakness; the dividend was maintained at a ~93% payout.
• 2024 — the reset. The Scotiabank Investment Agreement (Aug 12, 2024) closed in two tranches at $17.17 (~$2.8B / ~14.9% / ~163M shares); the H2-2024 securities repositioning (~$10B MBS sold, ~$1.8B pre-tax loss) produced the FY2024 GAAP loss of −$304M.
• 2025 — NII recovery. NIM 2.16% → 2.69%; NII (TE) +~23% to $4.67B; EPS back to $1.52; eight consecutive quarters of adjusted-PPNR growth through Q1-26; $1.0B buyback authorized (Mar-2025), restarted Q4-2025.
• 2026 — momentum + overhang. Q1-26 EPS $0.44 (+33% YoY), ROTCE >13%, NIM 2.87%; NII guide raised to +9–10%; June 5, 2026 Scotiabank 158.7M-share resale shelf.

Leadership / board. Gorman remains Chairman & CEO. CFO Clark Khayat in Q1-2026 assumed an expanded role additionally leading Technology & Operations — efficient for cost/AI accountability but a notable key-person concentration (CFO + CTO/COO scope) worth monitoring. Ken Gavrity now runs both the Commercial Bank and Payments. Two Scotiabank-designated directors sit on the Board — a governance oddity given that their principal is simultaneously exiting the stock.

Strategic shifts. Laurel Road / education lending moved to discontinued operations — a strategic retreat from a differentiator that never scaled economically; it removes a ~$0.14/share drag and simplifies the story, but is an admission the digital-consumer-lending bet failed. Balance-sheet optimization / “targeted scale” — intentional runoff of low-yielding non-relationship consumer loans ($500–600M/quarter of residential mortgage runoff) and a remix into higher-yielding commercial (commercial loan guide raised to +6–8% for 2026); estimated ~9% RWA reduction under the revised Basel standardized approach (+100bps marked CET1). Fee build-out — frontline sales force +~10% in 2025; priority fee businesses +12% in Q1-26; a mass-affluent wealth push. New NDFI/private-credit disclosure in Q1-26 (~$10.9B private-credit outstandings, ~70% via specialty-finance, 98% investment-grade) in response to sector scrutiny.

Regulation. KEY operates above the $100B LFI threshold (Category IV / standardized approach). The April-2025 Fed proposal would change the SCB framework; the revised Basel endgame is a net positive (preliminary ~9% RWA decline / +100bps marked CET1, to a fully-phased ~11%, “higher than peers, higher than we believe we need”). The AOCI-inclusion swing item is mitigated by the prior de-risking. Regulation has moved from headwind to modest tailwind — directly enabling the ramped buyback. (Management’s +100bps figure is a preliminary estimate contingent on the rules finalizing as proposed.)

Recent news/sentiment. Stephens reinstated Overweight with a $26 price target (6/15/26) — constructive sell-side into the resale overhang (third-party rating, cited only as sentiment context; not our view and not a target of ours). The June 5, 2026 resale shelf is the dominant recent corporate event; the June 10, 2026 8-K was a routine Medium-Term Note program update.

Verdict: net strengthen — but the strength is cyclical/sector-wide and the overhang is real. The arc is a genuine repair: de-risked balance sheet, inflecting and re-accelerating NIM/NII, dividend held, buyback ramping, comp well-aligned, regulation turned favorable, a failed consumer bet cut. These strengthen the operating thesis. But the NIM recovery is a cohort-wide tailwind (not KEY alpha) and largely priced; the IB recovery KEY built its differentiation on keeps deferring (only +5–6% guided for 2026); the Scotiabank resale is a live technical headwind; and the entire reset was needed because of a self-inflicted balance-sheet problem. The changes move KEY from “impaired laggard” to “recovering laggard” — better, but still the lower-return name in a mediocre industry, now priced for the recovery to complete to a 15%+ ROTCE it has not yet earned.

9. Risk Analysis

Risk skew. The dominant risks are not credit (benign, well-reserved) but valuation + technical: the embedded full-target ROTCE that may not arrive, the Scotiabank resale overhang, and the structural low-ROTCE quality cap — all downside-asymmetric at a richest-ever P/B. Credit and regulatory are manageable tails; macro/recession is the fat tail given the commercial-heavy book and 1.20 beta. The risk of a catastrophic or total loss is low (well-capitalized, IG, diversified, no acute solvency concern); the realistic downside is a meaningful de-rate, not impairment.

10. Valuation Discussion (Embedded Expectations)

No price target and no recommendation in this section — embedded-expectations and scenario analysis only.

The multiples (at $22.59, 18-Jun-2026). The headline optics suggest value, the franchise math does not.

• P/TBV ≈ 1.66x (TBVPS $13.585, Q1-2026 ROIC per-share data) — the key bank multiple.
• P/B at the 99.7th percentile of KEY’s own ten-year history (AZI valuation_index) — the richest KEY has ever been on book value. (The absolute P/B reads 1.45x on ROIC’s BVPS $15.57 or 1.23x on AZI’s larger common-equity basis $18.43 — an AOCI/intangible treatment difference — but on either basis the own-history percentile is 99.7th.)
• P/E (TTM) ≈ 12.7x on TTM EPS $1.78 (77th percentile own-history); forward ~11.0–11.9x on Street normalized EPS ~$1.90–2.05.
• Dividend yield ≈ 3.63% ($0.82 / $22.59); P/S 2.21x (80th percentile); composite 85.7th percentile. EV/EBITDA is meaningless for a bank.
• Own-history trend: P/TBV ran ~1.20x (2020) → ~1.0–1.4x through the crisis → ~1.26x average 2025 → 1.66x now — at/above the top of its multi-year range.

Embedded expectations — the central tension. Using the franchise identity P/TBV = (ROTCE − g) / (COE − g) and solving for the sustainable ROTCE the market is underwriting at 1.66x TBV:

At a ~10–11% cost of equity, 1.66x TBV embeds a durable through-cycle ROTCE of ~14.5–16% — i.e., the market is pricing KEY’s end-2027 management target (ROTCE 15%+, NIM 3.25%+) as essentially done and permanent. Against a realized continuing-ops ROTCE of ~11.85%, the price embeds a ~300–400bp step-up, held in perpetuity. The valuation is not cheap-on-current-earnings; it is “cheap” only if you (a) take the 2027 target as achieved and (b) believe the cyclically-elevated 2026–27 NIM is a sustainable through-cycle level rather than a peak. The reverse-DCF read: the recovery is fully priced, with no margin of safety if the target is missed.

Scenario analysis (illustrative; normalized ROTCE × TBV and P/E × normalized EPS; no price target). Normalized TBVPS anchor ~$14.50 (AOCI burndown + retained earnings, net of buyback).

• Bear (recovery stalls / NIM plateaus below target; ROTCE ~10.5%): normalized EPS ~$1.52; a fair multiple for a sub-COE laggard ~9–10x P/E and ~1.0–1.1x P/TBV → implied value materially below current price. The market would be paying ~1.66x TBV for a ~1.1x-TBV business.
• Base (NIM ~3.0–3.1%, ROTCE ~13%, partial target achievement): normalized EPS ~$1.89; fair ~11–12x P/E and ~1.4–1.5x P/TBV → implied value roughly at-to-modestly-below current price. The base case is approximately priced-in.
• Bull (full target: NIM 3.25%+, ROTCE ~15.5%, strong fee/IB cycle): normalized EPS ~$2.25–2.40; fair ~12–13x P/E and ~1.7–1.9x P/TBV → implied value modestly above current price (the Street’s ~$26 targets sit here, ~10–15% above).

The asymmetry is unattractive at $22.59: the base case is ~fair, the bull offers ~10–15% (and requires the full 2027 target to land), while the bear reprices materially lower because both the richest-ever multiple and the earnings compress.

Peer cross-read (P/TBV vs. ROTCE, ~TTM).

On a P/TBV-per-unit-of-ROTCE basis, KEY is expensive-for-its-quality: it carries roughly cohort-middle P/TBV while earning the cohort-lowest ROTCE. RF earns +600bp more ROTCE at only +0.5x more P/TBV; HBAN earns +400–600bp more at a lower P/TBV. KEY is priced as if convergence to peer-level returns is near-certain. The only honest “cheap-on-recovery” argument is the optionality that KEY closes its ROTCE gap to the 15% target faster than the market’s base case — but that is upside you are already paying for, not a discount.

Verdict. KEY is not a value name despite the “12.7x P/E / 3.6% yield” optics. P/B is at its richest-ever, and P/TBV embeds the 2027 target as done against the cohort’s lowest realized ROTCE. The base case is approximately in the price; meaningful upside requires full target achievement, while the downside is larger because both multiple and earnings would compress. Expensive-for-quality, with recovery-optionality you are paying for rather than getting free.

11. Variant Perception

Consensus. A clean recovery/re-rating story: NIM/NII inflected toward “record NII” (FY26 guide +8–10%), the 2024 GAAP loss was a one-time bond-book clean-up (now de-risked), AOCI is burning into tangible book, and ROTCE marches toward the 15%+ 2027 target. Sell-side price targets cluster near $26 (Stephens Overweight $26, ~15% above spot). KEY is seen as the “catch-up” laggard with the most ROTCE-gap to close.

Strongest bull case. The NII inflection is durable and structural: fixed-rate asset, swap, and securities repricing rolls higher mechanically regardless of rate path, and the de-risked book locks in higher reinvestment yields (NIM 2.16% → 2.69% → guide 3.00–3.05% exit-2026 → 3.25%+ 2027). AOCI burndown plus retained earnings accrete TBV and lift ROTCE simultaneously. If ROTCE re-rates to 15%+ (closing the ~300bp gap, helped by the ~10% banker build-out), the stock is “cheap” on normalized ~$2.25–2.40 EPS at ~12–13x. The above-cohort fee franchise (KBCM/Cain Brothers) deserves a quality premium over a pure spread bank, with a strong M&A/issuance cycle as free optionality. Once Scotiabank’s resale clears, the technical overhang lifts.

Strongest bear case. Richest-ever P/B (99.7th percentile own-history) for the cohort laggard: KEY earns the lowest ROTCE and NIM in the group yet trades at cohort-middle P/TBV — priced like a 16% bank, earning ~12%. The recovery is fully priced; 1.66x TBV embeds the 2027 target as permanent with no margin of safety if it slips or the NIM proves a rate-cut-dependent peak. The Scotiabank overhang (158.7M-share shelf; entry-at-trough/exit-into-strength) is a negative tell, and no executive is buying. Structurally, KEY is sub-scale with no durable moat and the worst efficiency in the cohort, and its one differentiator (IB) is cyclical and talent-mobile. Positioning is a crowded momentum trade near its relative-strength peak — the easy mean-reversion money has been made.

The assumptions that matter most, and what falsifies each.

1. (Bull) NIM reaches and holds ~3.25% through-cycle (not a rate-cut-dependent peak). Falsified if the NIM guide is cut, NIM plateaus <3.0%, or deposit costs re-accelerate — the embedded ROTCE collapses and the multiple de-rates.
2. (Bull) ROTCE converges to 15%+ by 2027. Falsified if ROTCE stalls at ~12–13% / efficiency stays ~62% — KEY is a permanent sub-COE laggard worth ~1.0–1.2x TBV, not 1.66x.
3. (Both) 2026–27 earnings are normalized, not cyclically-peak. Falsified if a recession lifts NCOs above the 40–45bps guide and provisions spike (commercial-heavy, 1.20 beta) — earnings and multiple compress together.
4. (Bear technical) The Scotiabank resale caps the stock near highs. Falsified if the 158.7M shares clear in an orderly block with no price impact, or Scotiabank reverses and holds — one bear leg gone.
5. (Bull) The IB/fee franchise is a sustainable quality premium. Falsified if IB fees prove purely deal-cycle-cyclical (revert in an M&A freeze) — the “better than a spread bank” premium evaporates.

Factor-positioning read (FactorsToday + AZI). KEY loads on DividendYield (~1.2–1.36 across nested models) and the Banks/Regional-Banks industry, with high explanatory power (R² up to 0.81); the closest factor twin is FITB (0.95), then USB/MTB/PNC/WAL and regional-bank ETFs. Its momentum/relative-strength profile is high and near peak: y1 total return ~+47%, rs_12m +49.7, rs_peak −0.73 (~3% below its own RS peak), Sharpe(y1) 1.89, beta 1.20, alpha +0.12 (lifetime max drawdown −87% — the 2008/2023 scars). On the regime, Quality and Momentum factors are strongly in favor while the Banks/Regional-Banks industry factor has recently softened (z ~−1.5) even as the broad market rises — an early sign the cohort’s momentum may be cooling. KEY is a crowded, high-momentum recovery trade near its relative-strength peak — emphatically not a falling knife; the offsides risk is a crowded long unwinding, which corroborates the valuation read that consensus is offsides to the upside.

Verdict. Consensus is largely right on the fundamentals (the inflection is real and durable) but the variant perception is that the stock is fully priced for a cohort laggard — the richest-ever P/B embeds the 2027 ROTCE target as a done deal while KEY remains the lowest-ROTCE/NIM operator and carries a fresh ~14% resale overhang. The bull needs the target to land; the bear needs only the target to slip or the cycle to turn.

12. Fact vs. Interpretation

13. Open Questions

1. Is the 2026–27 NIM a peak or a plateau? How much of the recovery is structural asset repricing versus rate-cut-dependent deposit-cost relief — the single swing factor for the embedded ROTCE.
2. Will middle-market M&A finally inflect in H2-2026 (the repeatedly-deferred IB catalyst), or is the “record pipeline” a perennial that doesn’t convert? Management guides only +5–6% IB fees for 2026.
3. Does the Scotiabank Investment Agreement contain a standstill/lock-up governing resale pace, and how quickly will the 158.7M shares clear? (The shelf suggests transfer restrictions are lapsing.)
4. Does the ≥$1.3B 2026 buyback fully offset SBC and leave the share count flat/down, or merely slow the growth created by the Scotiabank issuance?
5. Final Basel endgame text versus the preliminary +100bps capital benefit, and its timing.
6. CRE/office criticized-and-classified trend and the office maturity wall — management says favorable; validate against the Note tables across 2026.
7. CFO concentration: does Khayat’s combined CFO + Technology & Operations remit create execution or key-person risk?

14. What Must Be True

Bull case — what must be true: KEY’s NIM reaches and holds ~3.25%+ through-cycle on structural repricing, ROTCE converges to 15%+ by end-2027 (closing the ~300bp gap to peers and earning back the ~17% dilution), the IB/fee franchise sustains and ideally inflects with the M&A cycle, credit stays benign, and the Scotiabank resale clears without lasting price damage. In that world, ~1.66x TBV is defensible-to-cheap on normalized ~$2.25–2.40 EPS.

• Falsification test: a cut to the NIM/NII guide, or ROTCE stalling at 12–13% with efficiency stuck near 62% into 2027, falsifies the bull — it would confirm KEY as a permanent sub-COE laggard worth ~1.0–1.2x TBV, not 1.66x.

Bear case — what must be true: the cyclically-elevated 2026–27 NIM proves a rate-cut-dependent peak, ROTCE plateaus below target, the IB differentiator stays cyclically subdued, and/or a recession turns the commercial-heavy (54% C&I) book — so that the richest-ever multiple de-rates with compressing earnings, compounded by the resale overhang.

• Falsification test: ROTCE convergence to ~15% ahead of the 2027 schedule with the NIM holding ≥3.2% on structural repricing — plus a clean, orderly Scotiabank exit — falsifies the bear and makes today’s multiple earned rather than aspirational.

APPENDIX A — Standard Diligence Questionnaire

A standard diligence checklist. Fact / Interpretation / Assumption labels applied where it matters.

General

What thoughtful questions have other investors asked about this company? (1) Is the 2026–27 NIM a sustainable through-cycle level or a rate-cut-dependent peak? (2) Can KEY actually close the ~300bp ROTCE gap to peers and reach its 15%+ 2027 target, or is it a structural laggard? (3) How disruptive is the Scotiabank 158.7M-share resale overhang, and is there a standstill governing its pace? (4) Will middle-market M&A finally inflect to lift the IB franchise? (5) Was the Scotiabank dilution at $17.17 value-destructive or prudent? (6) Does the de-risked balance sheet now have manageable CRE/office exposure?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Interpretation: mid-recovery, not peak. ROTCE ~11.9% (2025) is well below the cohort and below management’s 15%+ target — but it is recovering off a 2024 GAAP loss, and the NIM/NII tailwind is cyclically favorable (recovering toward a guided peak). So earnings are below mid-cycle on returns but riding a cyclically-improving margin. Driven by external environment or internal actions? Both: the rate-driven NIM recovery is external/sector-wide; the securities repositioning and balance-sheet remix are internal. How stable are revenues? Moderately — ~62% spread income is recurring; ~38% fee income includes a cyclical ~10%-of-revenue IB line. Outlook for products/services? Spread income guided to “record NII” (+9–10% 2026); fee build-out (wealth/payments/IB) growing ~12% but IB deferred. How big will the market be? Mature, low-growth US banking; KEY’s organic loan growth is muted (average loans fell in 2025), concentrated in higher-spread commercial.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More — rising scale premium (tech/compliance), megabank share gains, accelerating regional consolidation. How profitable is the business? Fact: ROE ~10.4%, ROTCE ~11.85% (2025) — barely above cost of equity, lowest in the cohort. How profitable is the industry? Mediocre — commodity product, returns capped near COE for sub-scale players. Barriers to entry? Regulatory/capital barriers are real industry-wide, but among incumbents differentiation is weak. Can it be easily understood? Yes — a standard commercial-tilted super-regional. Undermined by foreign low-cost labor? No (domestic deposit/lending franchise). Do brands matter? Modestly (local trust/relationship); not a pricing-power brand. Nature of competition? Price (deposit rates, loan spreads) + relationship/service + fee-product breadth. Switching costs? Real but modest (operating-account stickiness); contestable. Verdict: no durable Greenwald moat; one above-weight but cyclical IB fee franchise (KBCM/Cain Brothers).

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The KBCM/Cain Brothers franchise and client relationships are intangible value not on the balance sheet; partly offset by AOCI marks (now −$1.96B, improving). Off-balance-sheet liabilities? Standard bank items (loan commitments, letters of credit, derivatives); nothing unusual flagged. NDFI/private-credit exposure (~$10.9B, 98% IG) newly disclosed. How conservative is the accounting? Reasonable — CECL reserves built in 2025 (not released), coverage ratios rose; the 2024 securities loss was crystallized transparently. How CapEx-hungry? Low physical capex; the real “capex” is ~$1B/year technology spend — material and rising (and now under the CFO’s combined remit). For a bank, the binding resource is regulatory capital (CET1 11.78%, comfortable).

Capital Allocation & Management

How much FCF, and how is it used? Bank “FCF” ≈ earnings available after maintaining capital ratios; FY2025 NI-to-common $1,686M. Priorities (Gorman, Q1-26): organic loan growth → tech/people investment → dividend → buyback. Philosophy? Repaired-but-unproven; disciplined bolt-on M&A; explicit valuation-opportunism on buybacks. Significant acquisitions recently? No large M&A; only boutique/team lift-outs. The defining capital event is the inbound Scotiabank ~$2.8B / 14.9% investment (2024). Buying back shares? Restarted Q4-2025; guided ≥$1.3B for 2026 (~5%+ of cap). Issuing shares to insiders? No unusual insider issuance; the ~163M Scotiabank shares are the dilution event. Compensation policy? Fact/positive: incentives tied to Adjusted ROTCE, Cumulative Adjusted EPS, PPNR, Cash Efficiency Ratio, and relative TSR — a genuine per-share governor that paid zero on 2023–25 awards when returns missed. Motivations of management? Reasonably aligned (86% at-risk pay, 6x/3x ownership guidelines), though no executive open-market buying and the plan was improved under Say-on-Pay pressure.

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — a US C-corp common stock (NYSE: KEY); standard 1099 dividends. Dividend policy? $0.205/quarter ($0.82/year), held flat through 2023–2025; ~54% payout (2025); yield ~3.6%. How profitable? Below-cohort (ROTCE ~11.9%). Net income diverging from cash from operations? Not in a distortive way — 2024’s GAAP loss was a realized securities loss (a real cash/economic event, not an accrual artifact); 2025 earnings are clean. Bank “CFO” is noisy and less meaningful than NII/PPNR/ROTCE for this model.

Risks & Downside

What would cause the stock to decline? A cut to the NIM/NII guide; ROTCE stalling at 12–13%; the Scotiabank resale pressuring the price; a recession turning the commercial-heavy book; an M&A/IB-fee freeze; multiple de-rating from its richest-ever P/B. Risk of catastrophic loss? Low — well-capitalized (CET1 11.78%), investment-grade, diversified, de-risked securities book, benign credit. Chance of total loss? Very low absent a systemic crisis; the realistic downside is a meaningful de-rate (both multiple and earnings), not impairment.

Recent News & Events

Has the business environment changed recently? Yes, favorably on operations (NIM/NII inflection, “record NII” 2026 guide, regulation turned to a modest tailwind ~+100bps capital) but with a fresh technical overhang (Scotiabank resale). Significant acquisitions? None recent (bolt-on only); the 2024 Scotiabank investment is the key capital event, now partly unwinding. Change in accounting policies? None material; Laurel Road education lending reclassified to discontinued operations. Recent changes — markets, facilities, management? CFO Khayat assumed expanded Technology & Operations leadership (Q1-26); Ken Gavrity runs Commercial Bank + Payments; two Scotiabank-designated directors on the Board; mass-affluent wealth push and ~10% frontline-banker expansion; new NDFI/private-credit disclosure. Stephens reinstated Overweight ($26 PT, 6/15/26) — sentiment context only.

APPENDIX B — Source Appendix

Primary sources first. All accessed 2026-06-21 unless noted. Facts reconciled to filings; third-party aggregated data (ROIC.ai, AZI, FactorsToday) used as cross-checks and labeled. Management commentary treated as hypothesis, validated against filings.

Primary — SEC filings (KeyCorp, CIK 0000091576)

• FY2025 Form 10-K (filed 2026-02-23, key-20251231.htm) — business/segments, NIM & average-balance tables, noninterest-income detail, loan composition, deposits, credit/ACL, capital (CET1 11.78%, SCB 3.20%), AOCI, Scotiabank Investment Agreement, dividends, share repurchases, discontinued operations (education lending), Statement of Changes in Equity. https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000091576
• FY2024 & FY2023 Forms 10-K (filed 2025-02-21, 2024-02-22) — dividend-per-share confirmation; 2024 securities-repositioning loss; pre-Scotiabank baseline.
• Q1-2026 Form 10-Q (filed ~2026-05-05, key-20260331.htm) — Q1-26 NIM 2.87%, NII (TE) $1.23B, diluted EPS $0.44, ROTCE, capital.
• Q3-2024 / Q1–Q3 2025 Forms 10-Q — quarterly NIM trajectory (2.17% trough → recovery).
• DEF 14A proxy (filed 2026-03-27) — CD&A incentive metrics (Adjusted ROTCE, Cumulative Adjusted EPS, PPNR, Cash Efficiency Ratio, relative-TSR modifier); 2023–2025 Performance Awards certified below threshold (zero payout); Say-on-Pay engagement; ownership guidelines.
• S-3ASR resale registration (filed 2026-06-05) — registers 158,723,874 common shares for the selling shareholder (Scotiabank) — the resale overhang.
• Form 4 corpus (217 since 2024-01-01) — insider transaction-code tally; 5 token code-P director-qualifying buys near the 2025 low; no executive open-market purchases; Bank of Nova Scotia 19 disposition filings at ~$21–22.
• 8-K timeline (2024–2026) — Scotiabank investment announcement/closings (Aug/Dec 2024); quarterly earnings (incl. 1/16/26 FY25 + 2026 guidance, 4/16/26 Q1-26); 6/10/26 Medium-Term Note program update.
• Investment Agreement (dated 2024-08-12, 10-K exhibit) — Scotiabank terms: ~$2.8B / ~14.9% / $17.17 fixed / two tranches (47,829,359 + 115,042,316 shares) / up-to-two board designees.

Primary — Earnings-call transcripts (via ROIC.ai)

• Q1-2026 earnings call (2026-04-16) — NII guide raised to +9–10%; NIM ~3.05% exit-rate (no-rate-cut base case); commercial loan guide +6–8%; ROTCE >13% (target 15%+ by end-2027); buyback raised to ≥$1.3B (“the floor”), ~$400M repurchased in Q1; IB fees +5–6%; capital priorities; Basel ~9% RWA reduction / +100bps; CFO Khayat’s expanded Technology & Operations role; NDFI/private-credit disclosure.
• Q4-2025 earnings call (2026-01-20) — FY2025 results + 2026 guidance (“record NII”).

Third-party / aggregated data (cross-checks; reconciled to filings)

• ROIC.ai — income statement, per-share data, profitability ratios, enterprise value, valuation multiples (KEY 6–8yr; peers FITB/RF/HBAN/CFG/MTB TTM); earnings-call transcripts. Third-party aggregated; EDGAR primary.
• Market-data feed — 5-year daily price/OHLCV CSV (beta 1.20, alpha 0.12); valuation_index own-history percentiles (P/B 99.7th, P/E 77.1th, P/S 80.3rd, composite 85.7th); news feed (6 items, incl. 6/5/26 resale-prospectus headline; Stephens OW $26 6/15/26).
• FactorsToday — factor loadings (DividendYield + Banks industry; twin FITB 0.95; R² to 0.81), leaderboard (y1 +47%, Sharpe 1.89, m3 ~+90% ann., lifetime max DD −87%), stock-info (rs_12m +49.7, rs_peak −0.73), related-stocks, factor-returns regime (Banks industry factor z ~−1.5 softening). Third-party statistical estimates.

Note on figures

Where ROIC and AZI differ on book value per share (ROIC BVPS $15.57 vs AZI $18.43, an AOCI/intangible treatment difference), tangible book per share (~$13.59–13.76) is used as the clean valuation anchor; the P/B own-history percentile (99.7th) holds on either basis. The May-2023 trough is stated on both bases: dividend-adjusted close ~$7.50 and raw intraday $8.54.