Wells Fargo & Company (NYSE: WFC) — The Turnaround Printed; the Price Caught Up

Independent fundamental research · Report date: 2026-07-18 · Report type: UPDATE (follow-up to the 2026-06-04 report) SEC CIK: 0000072971 · Sector: Financials — diversified money-center banking Reference price: ~$87.51 (NYSE close, 2026-07-17, yfinance — unofficial) · Market cap: ~$268B · Tangible book/share: $46.13 (6/30/26, Q2 2026 supplement) Peer cross-read: peer framing cross-read from prior reports on JPMorgan, Bank of America, Citigroup, U.S. Bancorp, PNC, and Truist (money-center industry/regulatory framing and the ROTCE↔P/TBV comp set)

⚡ Kimi’s Take — “The turnaround printed; the price caught up”

The author’s own subjective, independent opinion, offered as general information only; not investment advice. The analysis in sections 1–15 below carries no position.

Position: HOLD / accumulate-on-weakness (medium conviction) — a downgrade in risk/reward, not in the thesis. The June report said BUY/accumulate at ~$81.62 because the market was “paying for only half the recovery”: ~1.81x tangible book capitalized ~16% sustainable ROTCE against a 17–18% target. Six weeks later the recovery has printed — Q2 2026 delivered 17.7% ROTCE (in the target band for the first time), a 60% efficiency ratio (through the 64% bull-case line even adjusted for one-offs), collapsed CRE charge-offs, an 11% dividend hike — and the stock re-rated to ~$87.51 / 1.90x TBV, which back-solves to ~17% sustainable ROTCE: the market now capitalizes the low end of the target. The asymmetric edge that made June compelling is largely spent. What remains is a genuinely improving franchise at a fair price: ~+5–13% of conditional re-rating if 17–18% proves durable, against ~−14% of multiple compression if Q2 was the cycle’s peak — carried either way by ~8–10%/yr tangible-book compounding, a ~2.3% forward dividend, and a still-shrinking share count. Zone: accumulate below ~1.75x TBV (~$81); ~$88–99 (1.9–2.15x) is fair-to-full for 17–18% delivered; above ~2.15x (~$99) the risk/reward inverts. Not a short at any of these levels — the execution evidence is real.

The framing has shifted with the facts. This is no longer “self-help-improvement-at-a-discount with a regulatory catalyst” — the catalyst fired and the discount closed. It is now a quality-compounder-in-proving at a fair price: the burden moved from “can WFC execute?” (it is executing: loans +12%, deposits +10%, efficiency 66%→60%, headcount −7%, all three bull-case kill-switches disarmed) to “how much of Q2 2026 is cycle?” — the honest asterisk being that roughly two-thirds of the headline beat came from $847M of venture-capital marks and a discrete tax item (core EPS ~$1.80–1.85, core ROTCE ~16–16.5%), that the growth engine is deliberately low-spread markets financing funded by a still-eroding noninterest-bearing deposit base, and that Q2 2026 was the best industry quarter in years (JPM printed 23% ex-items; BAC 17% cleaner). The tape agrees this is a fundamental question, not a positioning one: the stock carries a ~zero momentum-factor loading, is dividend/value-factor-carried, and sits in an un-crowded repairing uptrend ~8% off its January high — no momentum crowd to unwind, no falling knife to catch. Governance moves the other way (a failed-grade 65.5% say-on-pay vote, three discretionary executive sales near the highs, zero insider buying, a combined Chair/CEO with no succession signal) and is the soft underbelly the market currently prices at zero.

Conviction and triggers. Medium conviction. The single fact that would flip this bullish: two consecutive quarters of ≥17% ROTCE with efficiency ≤62% excluding equity-gain noise and NII ex-Markets inflecting positive — that trips the bear-case falsification test and makes the 2.08–2.15x justified band (and the BAC/USB relative-discount argument) the base case. The single fact that would flip it bearish: ROTCE fading toward 15% as markets revenue normalizes with NII missing the reiterated ~$50B FY26 guide — confirming Q2 as the flattered cyclical peak and opening the ~14% air pocket to ~1.6x.

Changes since 2026-06-04 (update-mode summary)

• The call changed: BUY/accumulate → HOLD/accumulate-on-weakness. The thesis confirmed (Q2 ROTCE 17.7% in-band; efficiency 60%; CRE reserve-release call vindicated; assets +15% off the cap) — but the price re-rated ~7% and the whole complex with it, so embedded expectations rose from ~16% to ~17% sustainable ROTCE. Confirmation arrived; the discount that paid for it closed.
• Prior falsification scorecard (section 14 of the June report): the bull kill-switch was NOT triggered on any leg (no ROTCE stall below 15% — H1 16.1%; efficiency broke 64% — printed 60%, ~62.8% adjusted; CRE charge-offs collapsed to $16M with another ~$1B of reserves released). The bear falsification (“sustained sub-62% efficiency with ROTCE >16%”) is TRACKING, not triggered — one quarter qualifies, H1 sits at the threshold, and the easing cycle it referenced has inverted into a hawkish Warsh Fed.
• New facts that matter: post-cap balance-sheet growth is real and wholesale-led (CIB loans +26%, Markets financing +~$200B since end-2024) at an explicit NIM cost (2.43%, −25bp YoY); a new disclosed 10–10.5% CET1 operating range (answers June’s open question #3); buybacks deliberately slowing (~$14B annualized vs $17.5B FY25); dividend +11% to $0.50 (pending July board approval at writing); SCB frozen at 2.5% through 2027; 2026 say-on-pay was 65.5% (a correction to June’s 92.4% figure); a February 2026 cluster of ~$10M discretionary (non-10b5-1) executive sales near the highs; zero insider buying.
• Corrections to the June baseline: Q2 EPS is $2.00 GAAP (the “$1.96” circulating is a Zacks ex-item normalization); “~$10B returned in Q2” is actually the H1 figure (~$9.8B; Q2 alone ~$4.4B); the card book is ~$53B, not ~$83B (auto, ~$57B +32%, is the real consumer growth engine); the June 13D/A in WFC’s EDGAR feed is WFC’s own muni-fund-preferred housekeeping, not an activist filing.

📈 Stock Price Action — Five-Year Event Map

The arc (FACT — AZI adjusted prices, pull 2026-07-18). Five years ago WFC traded at ~$40; it bottomed at $33.36 (adj. close, 2023-03-24) in the regional-bank crisis, more than doubled to a $95.37 all-time closing high (2026-01-06; intraday 52-week high $97.76), round-tripped a −23% drawdown to $73.42 (2026-05-15), and now sits at $87.51 — 52-week range $72.78–$97.76 (intraday), −8.2% off the January closing high, +19% off the May low, +162% off the 2023 low.

Cycle narrative. (1) The 2022 drawdown was macro, not idiosyncratic — the fastest hiking cycle in four decades hit all banks, and WFC’s scandal-era discount compounded it. (2) March 2023 marked the cycle low: SVB’s failure indiscriminately repriced every bank with uninsured-deposit and AOCI exposure; WFC’s tangible book was never impaired the way regionals’ were, and the stock recovered within two quarters (INTERPRETATION). (3–4) The 2023–24 recovery ran on the soft landing and then the post-election deregulation bid — sector beta again, with WFC’s consent-order exits (13-plus terminated since 2019, per the June baseline) adding idiosyncratic fuel. (5) The asset-cap removal itself was a non-event on the day (+1.2%) — seven years of anticipation were already in the price; the repricing happened afterward, a ~25% grind over H2-2025 as the growth optionality got underwritten. (6) The 2026 drawdown was the market stress-testing that new multiple: a sell-the-news Q4 print off the all-time high, then a rates scare as expected cuts evaporated — a −23% peak-to-trough test that held well above the pre-cap-removal range (INTERPRETATION). (7) The June–July recovery is fundamentals-led — fee guidance, then a Q2 beat with 17.7% ROTCE — and notably the −2.7% print-day dip (an NII-miss reaction) was bought within two sessions, a different tape character than January’s sell-the-news, which started a five-month decline (INTERPRETATION). Price above rising 21/50/200-day EMAs ($85.27 / $83.08 / $81.66) with a ~zero momentum factor loading: a repairing uptrend, not a momentum crowd (FACT loadings; INTERP characterization).

Price moves are FACT (AZI adjusted OHLC, pull 2026-07-18; yfinance intraday 52-wk range); attributed drivers are INTERPRETATION, cross-referenced to the Q4-25/Q1-26/Q2-26 prints, the 2025-06-03 Fed action, and contemporaneous press (news feed). No price target, no recommendation.

1. Executive Summary

Wells Fargo’s Q2 2026 print (2026-07-14) is the strongest quarter of the Scharf era and the first hard evidence that the post-asset-cap thesis is working: diluted EPS $2.00 (+25% YoY) on revenue of $22.6B (+8.6%), ROTCE 17.7% — inside management’s 17–18% target band for the first time (H1 2026: 16.1% vs 14.6% FY2025) — an efficiency ratio of 60% (66% FY2025; ~62.8% excluding equity gains), net charge-offs of 0.34% (−10bp YoY), TBVPS $46.13 (+7%), and an 11% dividend increase to $0.50 expected in Q3. Thirteen months after the Fed lifted the ~$1.95T asset cap, average total assets are $2.23T, +15% YoY, with loans +12% and deposits +10% — the growth the cap suppressed for seven years is now printing.

The quarter still needs to be taken apart, not celebrated. Roughly two-thirds of the ~$0.27 headline beat came from items that will not repeat on schedule: $847M of venture-capital equity gains (vs $119M a year ago) and a $132M discrete tax benefit. Core EPS is ~$1.80–1.85 and core ROTCE ~16–16.5% — strong, but not the headline. The growth is deliberately concentrated in the lowest-spread businesses (markets financing, wholesale deposits), which is why NIM compressed to 2.43% (−25bp YoY) while peers’ margins expanded, and why the stock dipped on the print despite the beat. Noninterest-bearing deposits fell to ~24% of the mix, extending a five-year erosion of the franchise’s cheapest funding. And Q2 2026 was the best industry quarter in years — JPM earned 23% ex-items, BAC 17% on a cleaner quarter — so some of WFC’s improvement is tide, not boat. The honest summary: the turnaround is real and ahead of schedule on costs and credit; the revenue quality of the new growth is asserted, not yet proven.

The prior report’s falsification framework resolves cleanly. All three bull-case kill-switches were disarmed (no ROTCE stall, efficiency through 64%, CRE charge-offs collapsed with reserves released — the office reserve-release call was prescient). The bear-case falsifier is tracking but not yet tripped (one sub-62% quarter; “sustained” unproven). The moat verdict upgrades modestly: the same second-tier deposit/scale moat, now converting — the efficiency gap to JPM narrowed from ~14 to ~8 underlying points, CIB is gaining measurable share (US M&A #4 from #9; LevFin #3), and WIM earns a company-best 32.4% ROAC. Capital allocation remains the relative strength (dividend rebuild, disciplined buybacks paced against +9% RWA growth inside a newly explicit 10–10.5% CET1 range), but the governance ledger deteriorated: a 65.5% say-on-pay vote (a failed grade), a February cluster of ~$10M discretionary executive sales, zero insider buying, and a combined Chair/CEO with no succession signal.

Valuation has caught up with the story. At ~$87.51 the stock trades at 1.90x tangible book, back-solving to ~17% sustainable ROTCE (COE 10.5%, g 3%) — the low end of the target is now capitalized, versus ~16% embedded in June. WFC sits on the peer regression line on H1 returns and modestly below it on the Q2 run-rate (BAC and USB now carry higher multiples for equal returns). The remaining re-rating math is ~+5–13% (17.5–18% delivered) against ~−14% of relapse risk (fade to 14.6%), stacked on ~8–10%/yr of tangible-book compounding and a ~2.3% forward dividend — a narrower, more execution-dependent setup than June’s. The factor read says positioning is not the risk: a ~zero momentum loading, value/dividend-factor support, and an un-crowded uptrend inside a ~$73–$96 range. The debate that decides the stock from here is singular and quarterly observable: is 17.7% the new floor or the cycle’s ceiling? The sections below lay out the evidence both ways, the risk matrix, and the falsification tests we will apply at the next print. (No recommendation, no price target — the only positioned view in this report is the labeled Kimi’s Take block above.)

2. Business Overview

What it is (FACT — Q2 2026 release/supplement, 2026-07-14; FY2025 10-K). Wells Fargo is a diversified, deposit-funded US money-center bank — the #3 US deposit franchise (~$1.50T period-end deposits, ~10% of domestic deposits), #2 by branch count (4,079, down from 4,135 a year ago), and now the ~$2.3T-asset third-largest US bank. Q2 2026: revenue $22.6B (+8.6% YoY), net income $6.4B (+17%), diluted EPS $2.00 (+25%), ROE 15.0%, ROTCE 17.7% (H1 2026: 16.1%), efficiency ratio 60%, NIM 2.43%, CET1 (Standardized) 10.3%, TBVPS $46.13 (+7% YoY), 197,466 employees (−7% YoY; 24th consecutive quarterly headcount decline). G-SIB surcharge unchanged at 1.50% (bucket 1) — far below JPM’s 4.5%. What changed vs baseline: FY2025 figures were NI $21.3B, ROTCE 14.6%, efficiency 66%, NIM 2.64% — every headline return metric has since improved while NIM compressed.

The post-cap balance sheet — the single biggest factual change (FACT). The 2026-06-04 report’s central open question was whether WFC could deploy the freed balance sheet at peer-competitive returns. Thirteen months after cap removal (2025-06-03), the answer is that deployment is happening fast: average total assets $2,227.9B in Q2 2026, +15% YoY, versus $1,916.7B (FY24) and $1,986.3B (FY25) averages that were pinned near the ~$1.95T cap for seven years. Average loans $1,026.5B (+12% YoY); average deposits $1,465.6B (+10%); RWAs +9% to $1,342.2B; CET1 down from 11.1% to 10.3% — capital is being used, while buybacks continued ($3.0B in Q2; ~$7B H1) and the dividend rose 11% to $0.50.

Segment structure (FACT — Q2 2026 supplement; FY2025 baseline from the prior report, 2026-06-04).

¹ CB&L ROAC is flattered by a Q1 2026 allocated-capital methodology change ($45.5B→$33.0B). CIB Q2 included a −$181M provision (CRE reserve release). (Source: WFC 2Q26 Quarterly Supplement, segment pages; 2Q26 presentation pp.12–16.)

All four segments grew net income double digits YoY (CB&L +19%, Commercial +8%, CIB +34%, WIM +28%); CIB’s annualized run-rate (~$8.3B) now clearly exceeds its FY2025 level.

What exactly grew — and at what margin cost (FACT/INTERPRETATION). The growth is concentrated in the wholesale businesses WFC had to ration under the cap: CIB average loans +26% YoY to $359.4B (Banking $125.0B vs $89.0B; Markets $111.2B vs $79.0B; CRE $123.2B vs $117.9B), CIB deposits +16%, and Markets-related balances (trading assets +23%, resale agreements +80%) funded partly by wholesale repo (+86% to $242.9B average). Consumer growth is real but narrower: auto +31% to $55.7B average (originations +41%) — the standout — card loans +5% to ~$53B with new accounts +46% (662K), while home lending shrank 3%. The margin cost is explicit: NIM 2.43%, down 25bp YoY and 4bp QoQ, because the incremental deposits are interest-bearing commercial balances (interest-bearing deposits +14% YoY; average NIB deposits −2% to $354.4B, now ~24% of deposits vs ~26% FY25 and ~38% in 2021) and the incremental assets are lower-spread Markets financing (Markets NII $501M vs $104M). NII ex-Markets rose only 2%. Management’s framing — “what we are seeing in NIM is not happening to us… we’re either gonna get paid for it or we’re not gonna do it” (Scharf, Q2 call) — is a hypothesis: the claimed flywheel (financing → NIB deposits + trading/fee share) is partly visible (Markets revenue +24%, claimed client-level share gains) but the cross-sell payoff is not yet in the fee lines at scale, and the Street pushed back on exactly this point. Management expects modest Q3 NIM compression, then stabilization as Markets asset growth slows to “a natural rate” (+~$200B added since end-2024 — largely re-entry into capacity the cap forced it to cut).

How the money is made — mix shifting toward fees (FACT). Q2 2026 NII $12.3B (+5%) vs noninterest income $10.3B (+13%) — fees now ~46% of revenue, up from ~43% FY2025, closing some of the quality gap with JPM (~48% baseline) that the prior report flagged. Fee drivers: investment advisory/brokerage $3,508M (+13%, market valuations), deposit/lending fees $1,762M (+9%), firmwide investment-banking fees $939M (+35%), card fees $1,222M (+4%) — plus a volatile $728M YoY swing in venture-capital equity gains inside “all other” that flattered the quarter. Correction to baseline: the prior report cited “~$83B card loans”; WFC’s actual card book is ~$53B (period-end $53.3B, +6% YoY) — the baseline figure appears erroneous; the larger and faster-growing consumer book is auto (~$57B period-end, +32%).

Recurring vs. transactional (INTERPRETATION — carried forward, strengthened). The durable streams (deposit spread, WIM asset-based fees, card/payments) all grew; the cyclical streams (Markets, IB) grew fastest — 16% CIB revenue growth in a hot capital-markets quarter is precisely the mix the prior report labeled cyclical. Q2’s 17.7% ROTCE included the VC gain swing, a $132M discrete tax benefit ($0.04), and a CRE reserve release; an underlying ~16–16.5% is a more honest read (INTERPRETATION — not disclosed).

Verdict (INTERPRETATION). Business overview — upgraded from the baseline. The 2026-06-04 report described a capped bank optimizing mix inside a frozen balance sheet; that description is now obsolete. WFC is a growing franchise again: double-digit loan and deposit growth, fee mix rising, segment run-rates above FY2025, and the cap-era “~$1.95T” asset ceiling left behind ($2.23T average, +15%). The growth is genuinely broad-based but deliberately concentrated in lower-spread wholesale businesses, so NIM (2.43%) now sits well below peer levels — a self-chosen trade of margin for scale and relationships whose full profitability is asserted, not yet proven. The two-engine model (NII + fees) is intact and the fee engine is gaining share of revenue; the mortgage retreat continues as planned. The business is measurably higher-quality than at the baseline — with the caveat that Q2 2026 was a near-perfect macro quarter and some of the improvement is cycle, not structure.

3. Industry Dynamics

Carried forward from prior work. The money-center industry structure is adopted from our prior JPMorgan report and the WFC 2026-06-04 baseline: US banking is a below-average industry that houses a regulation-protected oligopoly — commoditized products, capital-intensive (regulatory capital caps ROE), cyclical and credit-geared, rate-dependent — in which the top four money-center banks (JPM, BAC, WFC, C) hold ~40%+ of deposits and earn 13–23% ROTCEs while the ~4,000-bank tail earns far less. Regulation remains the central paradox: a tax on incumbents and the highest entry barrier of any major industry. Marathon’s capital-cycle read also carries forward: post-GFC capital rules suppress supply inside the regulated perimeter (favorable for incumbents); the oversupply risk forms outside it (private credit ~$1.7T+, non-bank lenders). This section renders the updated WFC-specific overlay: the rate/macro shift, the capital-markets boom, and the first year of post-asset-cap competition.

Rate backdrop — hawkish surprise, then cooling (FACT). The baseline assumed Fed funds 3.50–3.75% “biased lower.” The actual path of 2026 expectations was more violent: markets opened 2026 pricing two cuts, swung to neutral by July (MarketWatch, 2026-07-13), and by the Q2 call WFC’s CFO said the market priced “a little over one increase” after the spring inflation spike (May CPI 4.2% YoY, a three-year high, driven by the Iran/Strait of Hormuz oil shock — Brent topped $86). June CPI then cooled sharply to 3.5% YoY (−0.4% MoM, the largest monthly drop since April 2020; core 2.6%) on the US–Iran ceasefire and a 5.7% energy decline (financecalendar.com, 2026-07-15). Fed Chair Kevin Warsh gave his debut testimony July 14. WFC-specific interpretation: WFC is asset-sensitive; the evaporation of expected cuts is NII-supportive versus January assumptions (management held FY2026 NII guidance at ±$50B, explicitly noting higher rates replaced assumed cuts), while any disinflationary easing path pressures asset yields but should normalize deposit mix (NIB stabilization). For WFC specifically, the rate question is now second-order to its self-inflicted mix shift: its 2.43% NIM is far below USB’s 2.79% and PNC’s 2.96% — both of which expanded YoY in Q2 while WFC’s fell 25bp — because WFC is deliberately adding lower-spread Markets/wholesale balances. The fee engine (IB, markets, WIM) is therefore the swing factor for WFC’s revenue quality, more than for any large-bank peer.

Capital-markets upcycle — a rising tide lifting exactly WFC’s growth businesses (FACT). Q2 2026 was an industry-wide blowout: JPM IB fees +30% (firmwide $3.3B) and Markets +35% (equities +86%); BAC IB fees +50%, sales & trading +33%; Citi revenue +14% (best in a decade); USB record revenue with BTIG contribution; PNC capital-markets-and-advisory fees +80% YoY. Industry IB fees were up ~26% (KBW) amid a hot issuance pipeline (incl. the SpaceX IPO). WFC participated fully: IB fees +35% ($939M firmwide), Markets +24%, and measurable share gains — YTD Leveraged Finance share 7.2% (#3), ECM share 3.8% (+74bp), announced US M&A rank #4, up from #9 (Dealogic, via WFC 2Q26 deck). INTERPRETATION (Marathon): this is the late-cycle phase where industry returns are cyclically flattered across the board; capacity is being added (every large bank is growing markets balance sheets and hiring bankers/traders), and the same tape that validates WFC’s CIB build-out today will mean-revert. Scharf himself flagged the supply-side risk on the call: on wholesale/AI-data-center credit, “there are more risk assets being created… and there is a lot of capital out there” — WFC claims to be staying in-lane, which is what every management team says at this point in the cycle.

Credit — benign and improving, industry-wide (FACT). WFC NCOs 0.34% annualized (−10bp YoY), coverage 1.40% of loans; BAC NCOs 0.47% (−8bp); PNC 0.25%; card delinquencies improving at BAC (five consecutive quarters) and WFC (30+ day 2.58% vs 2.63%). Consumer and commercial credit are simultaneously strong at every reporter — a cyclical sweet spot (strong labor market, wage growth, narrow spreads), not a structural condition.

Regulatory landscape — WFC’s idiosyncratic overhang substantially done (FACT). No new WFC enforcement actions or consent-order terminations were identified since 2026-06-04. The residual item remains the September 2024 OCC formal agreement on AML/BSA/sanctions; WFC closed 7 consent orders in 2025 (per its January 2026 8-K) and 13+ since 2019; the Fed’s 2018 cease-and-desist shell persists without the economically binding asset cap. Systemically: 2026 CCAR results did not change requirements (SCB stays 2.5%), and the Basel III reproposal remains pending — WFC expects a ~7% RWA reduction when final (a further capital-release tailwind specific to its deployment plans). WFC’s 1.50% G-SIB surcharge vs JPM’s 4.5% remains a relative advantage. INTERPRETATION — carried forward: the regulatory arc (penalty box → cap removal → residual AML item) is substantially complete; this de-handicapped WFC to peer freedom but conferred no premium. The new fact is that WFC is now spending that freedom — CET1 10.3%, at the low end of its 10–10.5% target range — so the next regulatory event that matters is the Basel finalization, not consent orders.

Peer ROTCE set — refreshed (FACT — Q2 2026 annualized, company-reported; FY2025 per prior reports).

INTERPRETATION: WFC gained ~3 points FY25→Q2-26 — but so did everyone (industry-wide hot quarter with Visa gains sprinkled across the group). The ~5-point gap to JPM (17.7% vs 23% ex-items) is essentially unchanged from the baseline’s ~5.4 points; WFC also still trails USB and is now only marginally ahead of BAC (17.7% vs 17.0%, and BAC’s was cleaner). The relative standing from the baseline — mid-pack, behind JPM/USB/PNC, ahead of TFC/C — is intact. WFC is closing into the pack, not breaking away from it.

Verdict (INTERPRETATION). Industry verdict carried forward, with a hotter cyclical overlay: a structurally below-average, commoditized, rate- and credit-geared industry that remains a good one for the protected money-center oligopoly and a bad one for sub-scale banks. What changed since 2026-06-04 is the cycle, not the structure: the rate path turned hawkish-then-cooling (net NII-supportive vs January), capital markets are in a boom that is inflating every bank’s returns and IB pipelines, credit is benign everywhere, and WFC’s regulatory idiosyncrasy is resolved down to a residual AML agreement. Marathon’s framework flags the two supply-side watch items: (i) capital-markets capacity addition across the industry into a hot tape, and (ii) bank + non-bank capital flooding wholesale/AI-linked credit at narrow spreads — the exact conditions in which next-cycle losses are underwritten. WFC’s re-entry growth is, so far, disciplined by the numbers (NCOs down, coverage intact), but the industry backdrop is now a tailwind that flatters — and will fade.

4. Competitive Position

Carried forward from the 2026-06-04 baseline (prior work): WFC has a real but second-tier competitive advantage — a top-three deposit franchise plus retail scale, in Greenwald’s taxonomy a demand-side customer-captivity (primary-checking switching costs) + local economies-of-scale moat, historically unconverted into peer-level returns because of a bloated cost base, a spread-heavy mix, and the regulator-imposed asset cap (the “anti-moat,” 2018–2025). The baseline’s JPM-gap decomposition — roughly three-quarters cost and mix, one-quarter franchise damage — and its verdict (“scaled-but-underearning; gap partly closable; mix disadvantage structural”) are re-tested below against 13 months of post-cap evidence. The update: two of the three gap drivers (cost, cap legacy) are closing measurably faster than the baseline projected; the third (mix) is closing partially and cyclically.

Moat #1 — Low-cost deposit funding (REAL; edge still thinning on mix, but total franchise is growing again). The mechanism is unchanged: primary-checking captivity yields below-beta funding. Q2 2026 evidence: average deposit cost 1.51% (vs 1.52% 2Q25; interest-bearing deposit rate 1.99% vs 2.09%), deposits +10% YoY to $1,465.6B average with growth in every segment, and — the metric that would deteriorate without the moat — NII held up (+5%) despite deliberate mix dilution. Two-sided update: (i) negative — average NIB deposits fell another 2% to $354.4B (~24% of deposits; ~26% FY25, ~38% in 2021), so the free-funding edge keeps eroding and management now expects NIB only “stable” with deposit cost “inching up”; (ii) positive — primary consumer checking grew for the 13th consecutive quarter (management), deposit growth accelerated to +10% without pricing aggression (“we are really careful to not overpay” — CFO; no pricing pressure observed), and the wholesale deposit franchises (CIB +16%, Commercial +6%) are winning operating balances that carry treasury-management and payments attach. Greenwald share-stability test: WFC’s ~10% domestic deposit share held through the cap years and is now growing — consistent with a real moat that was suppressed, not destroyed. Net: moat intact, magnitude still narrowing on mix, but franchise momentum has shifted from holding to gaining.

Moat #2 — Branch/retail scale (REAL; the “size without scale” indictment is finally weakening). The baseline’s sharpest criticism was 66% efficiency vs JPM’s ~52% — 4,000+ branches not translating into cost economics. Update (FACT): Q2 2026 efficiency 60% (H1 64%; Q1 67% included severance), on headcount −7% YoY (197K; −79K over six years), branches −56 YoY to 4,079, and positive operating leverage (revenue +9%, expense +2%, with revenue-related comp +19% and non-revenue expense down). JPM’s Q2 overhead ratio was ~47% reported but ~52% excluding its $5.6B of one-off gains (INTERPRETATION — estimate from JPM release), so the underlying efficiency gap narrowed from ~14 points to roughly 8 points. BAC, the closest scale comparator, printed 59% — WFC has essentially caught BAC on efficiency. The moat metric (unit cost of serving the retail franchise) is improving faster than the baseline’s trajectory implied; scale is finally converting. Strengthening.

Moat #3 — CIB (was “partial, cyclical, sub-scale” — now demonstrably gaining share; still cyclical). Q2 2026: revenue +16% ($5,425M), ROAC 19.2% (vs 14.9% a year ago), efficiency 46%, IB fees +36% within the segment, Markets +24% (equities +64%), and Dealogic share gains: LevFin #3 (7.2%), ECM 3.8% (+74bp), US M&A #4 from #9. The cap is the revealed constraint: management says Markets was cut “more significantly than any other place” under the cap, and the +$200B re-deployment since end-2024 is producing both revenue and claimed client-level share gains. But keep the Greenwald discipline: WFC’s IB fee pool ($939M firmwide) is still ~28% of JPM’s ($3.3B), its Markets revenue ($2.2B) ~18% of JPM’s ($12.1B), it lacks JPM’s #1-IB intangible and its payments/securities-services switching-cost annuities, and Q2 2026 was the hottest capital-markets quarter in years — share gained in a boom must survive a downturn to count as a moat. Upgraded from “partial” to “credible challenger, cyclical”: the mix disadvantage is narrowing, but durability is unproven through a cycle.

Moat #4 — WIM (modest moat, strengthening on execution). Q2 2026: revenue +13%, ROAC 32.4% (best in the company), client assets $2.41T segment / $2.69T company-wide (+15% YoY) vs the $2.5T baseline, NII +17% on deposit pricing, near-record advisor recruiting for three straight quarters with record-low attrition (management), and a >$1B multi-year technology modernization now shipping: the Aladdin-based Advisor Gateway desktop (launched May 2026) plus a new AI advisor copilot (July 2026) (BlackRock press release, 2026-05; Banking Dive, 2026-07). The captivity mechanism (advised-asset stickiness) is real, and the advisor-recruiting data suggests the platform is now attracting rather than merely retaining — but 81% efficiency keeps it a lower-margin business than JPM’s AWM, and asset-based fees remain market-level-dependent (a beta exposure, visible in Q2’s +13% advisory fees). Strengthening, still modest.

The anti-moat — retired, with one residual. The asset cap (2018–2025) is gone and the evidence above shows what it was suppressing. The residual September 2024 OCC AML/BSA formal agreement is the last open enforcement item; no new actions emerged since the baseline. The seven-year freeze’s legacy persists in the cost base and in lost CIB share — but both are now moving in WFC’s favor.

The JPM gap, re-scored (INTERPRETATION — the core update). Baseline gap: ~5.4 ROTCE points (14.6% vs ~20%), decomposed ~¾ cost+mix, ~¼ franchise damage. Q2 2026: 17.7% vs 23% ex-items — ~5.3 points, and ~6–7 points on an underlying basis (~16.5% vs 23%). Against the baseline’s three drivers:

1. Cost base — closing fastest. Underlying efficiency gap ~14pts → ~8pts. Headcount −7%, non-revenue expense falling, severance largely behind (Q4-25/Q3-25 charges of $612M/$296M). Management still guides FY26 expense ~$55.7B and says “a lot still to go.” This driver is tracking ahead of the baseline’s “largely fixable” call.
2. Business mix — closing partially, and partly cyclically. Fee mix ~43% → ~46%; CIB ROAC 14.9% → 19.2%; IB share gains; WIM at 32% ROAC. But the quarter that proved it was a boom quarter, and the growth engine (Markets financing) is NIM-dilutive with its payoff in cross-sell asserted rather than banked. Structural underweight vs JPM persists in payments/securities-services and in absolute IB/markets scale.
3. Asset-cap legacy — now visibly reversing. +15% asset growth, +26% CIB loans, CET1 deployed from 11.1% to 10.3%. The baseline’s “forward tailwind” is now a present-tense fact.

Management’s 17–18% ROTCE target was met in Q2 (17.7%) — one quarter, flattered by ~1–1.5 points of one-offs, in the best industry quarter in years; the Street’s question has shifted from whether to when sustainably, and management pointedly declined to commit to timing while stating confidence is “higher, not lower” and calling 18% a “waypoint.” The efficiency ratio remains the scoreboard; NIM-vs-fee-mix is now the second scoreboard.

Greenwald tests, updated (INTERPRETATION). Share stability: deposit share stable-to-rising, primary checking growing 13 straight quarters, CIB/IB share rising — passes, improved from “static” at baseline. Returns test: Q2 ROTCE 17.7% (underlying ~16–16.5%) vs a ~10–11% cost of equity — clears the bar now, but H1 16.1% and the one-off-adjusted figure say “clears narrowly mid-cycle,” not “clears decisively.” Marathon capital-cycle test: WFC’s own asset growth (+15%) is the asset-growth-anomaly question in miniature — so far it passes (returns rose, NCOs fell, coverage intact, growth is re-entry into previously-held share rather than conquest), but it is adding capacity into a hot cycle alongside every peer, which is precisely the setup Marathon warns about. The disciplined-growth language (“we know that such favorable conditions do not go on forever… we are being selective”) is a hypothesis the next downturn will grade.

Verdict (INTERPRETATION). Upgraded from the baseline: WFC remains a real-but-second-tier moat — the same demand-side captivity plus scale mechanism, now demonstrably converting as the anti-moat lifts. What changed: the cost leg of the JPM-gap diagnosis is closing faster than projected (efficiency 66%→60%, gap ~14→~8 underlying points), the cap-legacy leg is reversing in plain sight (+15% assets, CIB share gains, WIM momentum), and the moat metrics that would deteriorate without the advantage (deposit cost 1.51%, primary-checking growth, NII resilience) are all holding or improving. What did not change: the mix disadvantage is structural and only partially and cyclically narrowed — WFC’s IB/markets/wealth pools remain a fraction of JPM’s, its NIM (2.43%) is now peer-trailing by construction, and its 17.7% ROTCE still trails JPM’s ex-items 23%, USB’s 18.7%, and PNC’s 17.9%, while BAC (17.0%, cleaner) has nearly closed on it from behind. The baseline’s framing survives intact: WFC is the same-model-weaker-execution foil to JPM — but the execution deficit is now shrinking on evidence, and the honest residual question is no longer “can it deploy the balance sheet” (it is, profitably so far) but “how much of Q2 2026 is the cycle” — a question only a downturn will answer. Durable competitive advantage: yes, narrow-to-moderate, and for the first time in a decade widening rather than static.

5. Growth History and Forward Opportunities

The headline change since 2026-06-04: the post-asset-cap growth thesis is no longer prospective — it is printing. The prior report framed the cap removal (2025-06-03) as an unlock whose first fully unconstrained year, 2026, was “unproven.” H1 2026 is the first hard evidence, and it is strong (FACT — 2Q26 supplement):

For seven years average assets were pinned near the ~$1.95T cap; they are now compounding at mid-teens. Growth is broad-based by segment (FACT): CIB average loans +26% YoY (markets financing re-entry — ~$200B of balance sheet added since end-2024 per the CFO), WIM +12%, CB&L +6%, Commercial Banking +5%; CIB deposits +16%. H1 2026 net income $11.66B (+12%) and EPS $3.60 (+21%) on revenue +8%.

Where the growth is coming from (FACT/INTERPRETATION). The five forward avenues from the prior report are all active: (i) balance-sheet redeployment — the dominant driver, concentrated in low-spread markets/prime financing (the CFO concedes the pace “was really a reflection of us coming out of the asset cap” and will normalize); (ii) markets — revenue +24% YoY, financing revenue “not quite a double” but close, trading revenue +20%+; (iii) cards — new accounts +50–60% YoY, a deliberately NIM/loss-seasoning-dragging investment in a high-return fee pool; (iv) WIM — client assets $2.4T+ (+15%), near-record advisor recruiting for three quarters; (v) efficiency-led earnings growth (section 7.5). New IB fees +35% YoY to $939M, the highest-quality fee line.

Quality of the growth — the skeptic’s discounts (INTERPRETATION). Three caveats keep this from being unambiguously high-quality growth. First, much of the balance-sheet growth is the lowest-spread kind: markets financing and interest-bearing commercial deposits (+16–17% YoY) are NIM-dilutive by construction — management’s defense (“we are either gonna get paid for it or we are not gonna do it”) is a hypothesis asserted client-by-client, not yet visible in a consolidated returns bridge. Second, the funding mix is still degrading: average noninterest-bearing deposits fell to 24.2% of deposits (from 27.1% a year ago), extending the 2021→2025 −31% erosion; deposit growth is being bought with interest-bearing and wholesale money (interest-bearing liabilities +22% YoY), and average deposit cost inflected up 8bp QoQ to 1.51%. Third, Q2’s revenue optics include $847M of venture-capital equity gains (vs $119M a year ago) — ex-gains, revenue grew ~5%, not 9%.

Forward trajectory (INTERPRETATION/ASSUMPTION). Management reiterated FY2026 NII of “~$50B plus or minus” — requiring ~$12.8B/quarter in H2 vs $12.3B in Q2 (+3.9%): achievable on day count, fixed-rate repricing, and volume, but tight. Loan-growth guidance was raised above the original FY26 assumption. The Marathon capital-cycle risk flagged in the prior report — undisciplined volume-chasing post-cap — has not surfaced: commercial underwriting is described as consistent, and Scharf is publicly declining the frothier wholesale risks (data centers, AI supply chain).

Verdict (INTERPRETATION). Growth quality has improved from “buyback-dependent and unproven” to “real and broad-based, with a mix discount.” The 12% loan / 10% deposit growth is the genuine, asset-cap-enabled organic expansion the bull case required, arriving with positive jaws and no visible credit-quality cost. But it is early, the marginal dollar is low-spread markets financing funded by costlier liabilities, and one good quarter of venture marks flattered the revenue line. The open question from the prior report — profitable deployment vs. lower-return volume — is now answerable quarterly and is trending favorable, not yet proven.

6. Financial Quality

Q2 2026 was the strongest print of the Scharf era — and it needs to be taken apart, not celebrated. The GAAP headline: net income $6,407M (+17% YoY), diluted EPS $2.00 (+25%), ROTCE 17.7% — inside the 17–18% target range for the first time (FACT — 2Q26 release/supplement). Stripped of $847M of venture-capital equity gains (≈$0.15/sh after tax and the noncontrolling-interest share, which jumped to $238M) and a $132M (~$0.04) discrete tax benefit, core EPS is ≈$1.80–1.85 and core ROTCE ≈16–16.5% (INTERPRETATION). That is still a strong quarter — revenue ex-gains +5%, expenses +2%, charge-offs down — but roughly two-thirds of the ~$0.27 headline beat vs. consensus came from venture marks and tax, and the stock’s post-print selloff (on NIM commentary) suggests the market did the same arithmetic.

Returns — trend vs. the 14.6% FY2025 baseline (FACT).

ROTCE is tracking toward 17–18%, not yet durably there: H1 at 16.1% is the clean marker, and Q2’s in-range print is venture-assisted (INTERPRETATION). Q1’s 14.5%/67% was the soft patch — elevated personnel cost on seasonal and revenue-related comp, no severance (FACT — Q1-26 10-Q).

Efficiency and jaws (FACT/INTERPRETATION). The prior report called the efficiency ratio “the scoreboard.” Q2’s 60% (62.8% ex-equity-gains) against the 66% FY2025 / 64% Q4-25 baseline is a decisive break below the 64% line the bull case required. The mechanics are genuine: headcount −7% YoY to ~197K (24 consecutive quarters down; −79K from the 2020 peak) and efficiency initiatives funding revenue-related comp, technology (+13% YoY) and advertising (+36%) investment — expenses ex-revenue-related comp declined YoY (FACT, per CEO). H1 jaws are ~+5pp (revenue +8%, NIE +3%). The gap to JPM’s ~52% remains ~8–11 points — closing, not closed.

NII/NIM — the one deteriorating line (FACT). NIM fell to 2.43%, −25bp YoY, −4bp QoQ, with a guided small further Q3 decline before stabilization. The driver mix has changed character since the baseline: this is now deliberate mix shift (markets financing, interest-bearing deposit growth) more than rate beta — Scharf: “what we are seeing in NIM is not happening to us.” Deposit economics confirm the strain: NIB deposits down to 24.2% of average deposits (27.1% a year ago), average deposit cost up 8bp QoQ to 1.51% and guided to “inch up” in H2. NII itself is fine — $12.3B, +5% YoY; H1 $24.4B — and the ~$50B FY26 guide is credible but tight (+3.9%/quarter needed in H2). Note for the scorecard: the market now prices a rate increase (CFO), inverting the prior report’s easing-cycle premise — as a liability-sensitive bank that marginally helps, but the mix shift dominates either way (INTERPRETATION).

Credit — the flagged CRE bet is winning (FACT).

(All five quarters per the 2Q26 supplement credit tables.)

The prior report asked whether releasing office reserves while office charge-offs rose was “prescient or premature.” Four quarters on: prescient (INTERPRETATION). WFC cut CRE reserves ~$1B (2.50%→1.77% of CRE loans) while CRE charge-offs collapsed to $16M and the office book ran off 20% YoY to $21.1B, with office nonaccruals falling in absolute terms ($2.2B from $2.5B). Residual caution: office nonaccruals are still 10.6% of the remaining office book, so the release is partly a shrinking-denominator effect; and new card vintages (+50–60% account growth) season into losses with a lag, against an already 8.5%-reserved card book. Provisions remain a tailwind ($914M vs $1,005M), consumer credit strong (card NCOs $600M vs $622M).

Capital, tangible book, and the buyback inflection (FACT).

TBVPS compounding continues (+7% YoY) but remains buyback-carried — TCE is flat YTD. The material change: management now discloses a 10–10.5% CET1 operating range and sits at 10.3%, i.e., the excess-capital era is over and post-cap RWA growth (+9% YoY) is absorbing earnings. Q2 buybacks slowed to $3.0B (vs ~$4.4B/quarter FY25 pace); H1 total return ≈$9.8B ($7.0B buybacks + ~$2.8B dividends) — the widely-quoted “~$10B in the quarter” is wrong; it is the H1 figure (FACT). Dividend +11% to $0.50 in Q3 (payout ~25%). A Basel reproposal (~7% RWA reduction if finalized) is the potential ~80bp capital release — not yet actionable (CFO). AOCI loss widened to $(8.3)B as rates backed up; HTM unrealized losses are not disclosed in the supplement and the 10-Q is not yet filed — OPEN QUESTION (last verified figure is the FY25 10-K).

Falsification scorecard (tests quoted verbatim from section 14 of the prior report).

• Bull test — NOT TRIGGERED on all three legs. “ROTCE stalling below ~15% for several quarters” — no: Q1’s 14.5% was one quarter; Q2 rebounded to 17.7%, H1 16.1%. “The efficiency ratio failing to break below ~64%” — no: 60% printed, ~62.8% adjusted, H1 64%. “Rising CRE/office charge-offs forcing reserve rebuilds” — no: CRE NCOs $16M, reserves released another ~$1B, office book −20%.
• Bear test — TRACKING, not triggered. “A sustained sub-62% efficiency ratio with ROTCE holding above 16% through the easing cycle” — Q2 alone qualifies (60% / 17.7%), but “sustained” is one quarter, H1 sits at the threshold (64% / 16.1%), the quarter was venture-flattered, and the easing cycle it references has inverted into an expected hike. Two or more clean quarters at sub-62%/>16% would trip it.

Verdict (INTERPRETATION). Financial quality has improved meaningfully since 2026-06-04, with one honest asterisk. The improvement is real: efficiency through 64% even adjusted, ROTCE tracking to target, credit benign with the CRE reserve-release call vindicated, and all three bull kill-switches disarmed. The asterisk is composition — the quarter’s best optics (60% efficiency, 17.7% ROTCE, +13% fee income) leaned on venture marks; NIM is still compressing on a deliberate, unproven low-spread build-out funded by a degrading deposit mix; and capital return has necessarily downshifted now that CET1 sits at target. Economics are improving with scale for the first time in the post-cap era — the H2 test is whether sub-63% efficiency and >16% ROTCE hold without $850M of equity gains, and whether NIM stabilizes after Q3 as promised.

7. Capital Allocation

Update-mode draft, report date 2026-07-18. Builds on the 2026-06-04 baseline; NEW items since 2026-06-04 are marked [NEW], carried-forward items are condensed.

Capital allocation remains WFC’s clearest relative strength versus the money-center group — disciplined divestitures, real cost reduction, a conservative dividend rebuild, and heavy buybacks now running against a growing post-asset-cap balance sheet. The update adds two genuine governance data points: a weak 65.5% say-on-pay vote and a February 2026 cluster of discretionary (non-10b5-1) insider sales.

Capital return — pace slowing as the balance sheet grows (FACT, [NEW]). Q2 2026 buybacks were 37.4M shares / $3.0B, taking 1H26 repurchases to ~$7.0B (~$14B annualized) — a step-down from the ~$17.5B FY2025 and $19.4B FY2024 pace (8-K Ex 99.1, 2026-07-14; call transcript 2026-07-14). Common dividends declared were $0.45/sh in Q2 (~$1.4B derived), so Q2 total capital returned was ~$4.4B against $6.16B net income applicable to common (~71% payout); 1H26 ≈ $9.8B returned vs $11.16B (~88%). NB: press framing of “$10B returned” refers to the first half, not Q2. The slowdown is deliberate capital triage, not a constraint event: Standardized RWA grew +9.5% YoY to $1,342.2B as average assets reached $2.23T (+15% YoY) — the post-cap growth the thesis promised is consuming capital, and management is funding it partly out of the buyback line.

Dividend (FACT, [NEW]). WFC announced (2026-06-24) an expected +11% Q3 dividend to $0.50/sh (from $0.45), subject to Board approval at the July meeting — still pending as of 2026-07-18. Path: $0.40 (Q2’25) → $0.45 → $0.50; annualized $2.00 ≈ ~26–28% of ~$7.20–7.60 annualized EPS (1H26 EPS $3.60) — still a conservative payout with the bulk of return running through buybacks. At $2.00 annualized, the dividend is now essentially back to its pre-2020-cut run-rate ($2.04 in early 2020) after six years; note the prior report’s “$1.70 / $1.92 peak” figures mixed quarterly and annualized bases.

CET1 — new explicit target range (FACT, [NEW]). CET1 (Standardized, binding) is 10.3% at 6/30/26, down from 10.6% at 12/31/25 and 11.1% a year ago — buybacks plus RWA growth grinding it down by design. On the Q2 call, management disclosed a stated CET1 target range of 10%–10.5% and quantified its minimum-plus-buffers at 8.5% (4.5% + 2.5% SCB + 1.5% G-SIB); the CFO is “really comfortable… anywhere in that range” and says buyback capacity remains. This resolves the prior report’s open question #3: the buffer is ~180bp and managed, not accidental. 2026 stress-test results (released 6/24) do not affect requirements — the Fed froze SCBs through 2027 — and WFC’s stressed depletion was below the SCB floor, so its SCB stays 2.5% (WFC newsroom 2026-06-24; BPI 2026-07-01). Scharf flagged that capital plans are set “absent the finalization of the capital rules” — Basel III endgame remains the swing factor.

Buyback arithmetic at ~1.9x book — a refinement, skeptically (FACT/INTERPRETATION, [NEW]). At $88.10 (7/17 close) vs TBVPS $46.13 (+7% YoY), WFC repurchases at ~1.91x tangible book (prior report: 1.81x). A mechanical correction to the prior framing: buybacks above 1x TBV are mildly TBVPS-dilutive (~0.4% per $1B: $139.7B TCE over 3,028.5M shares, retiring 11.35M shares per $1B at $88.10, moves TBVPS $46.13 → ~$45.97). What they are: EPS-accretive (~+0.37% per $1B) and ROTCE-accretive (same earnings on less TCE), at an ~11% earnings yield against near-zero-yield excess capital — and still roughly half the TBVPS drag per dollar of JPM buying at ~2.9x. The share count is down 6% YoY to 3,028.5M and ROTCE printed 17.7% in Q2 (16.1% 1H) — inside the 17–18% target band for the quarter. The accretion story is real but runs through returns-on-equity math, not tangible-book compounding; it reverses sign if ROTCE stalls.

M&A — a divestor, not an acquirer (carried forward). The “12 businesses sold or exited since 2019” discipline stands; [NEW] the rail-car leasing sale (GATX/Brookfield JV) closed in Q1 2026 per the Q2 release (lower lease income/expense YoY). No acquisitions.

Insider transactions (FACT, [NEW]). Six Section-16 filings in the window; all mechanical — zero open-market buys (P), zero open-market sells (S):

Trailing pattern (110 filings scanned, 2025-06-01 → 2026-07-17): zero code-P open-market buys by any insider. The only code-S sales are a February 2026 cluster — Kleber Santos 25,000 @ $87.72 (~$2.2M, 2/20); Bridget Engle 30,000 @ $87.10 (~$2.6M, 2/27); Ellen Patterson (General Counsel) 60,000 @ $87.40 (~$5.2M, 2/27) — and all three forms leave the 10b5-1 checkbox unchecked (aff10b5One=0), i.e., not reported as plan sales. The prior report’s “routine 10b5-1 selling” label was imprecise: ~$10M of discretionary selling near the 52-week high, against zero insider buying at ~1.8–1.9x book. Not damning — executives may sell for many reasons, and two of the three kept large positions — but the honest read is that no insider volunteers capital at this multiple.

Incentive alignment and governance (carried forward + [NEW]). The PSA architecture (3-yr absolute ROTCE 75% / relative ROTCE 25%, relative-TSR modifier; 2026 absolute target raised 100bp) and its caveats stand: no quantified risk-remediation metric in the pay formula, a “principles-based” annual bonus, the ~$60M one-time CEO special award, eliminated pre-set comp targets, and the combined Chair/CEO role. [NEW — correction]: 2026 say-on-pay passed with only 65.53% support (8-K 2026-04-30, Item 5.07) — not the 92.4% the prior report cited (a 2025 figure). A sub-70% vote is a failed grade by proxy-advisor standards and formally obliges a board responsiveness review; it is measurable evidence that the pay flags are not just ours. The independent-chair shareholder proposal drew 33.9% support; a majority-vote-governance proposal nearly passed at 47.9%. No board changes or succession announcements since the proxy.

Verdict (INTERPRETATION). Capital allocation stays positive on substance, with sharpened governance discounts. The substance improved this window: ROTCE hit the 17–18% target band, the dividend rose 11%, the buyback is being paced against post-cap RWA growth inside a newly explicit 10–10.5% CET1 range — that is what disciplined, rules-based capital management looks like, and it answers the prior report’s biggest open question favorably. The discounts also sharpened: a 65.5% say-on-pay rebuke, three discretionary executive sales and zero insider buying, and a buyback whose accretion is returns-math rather than book-value compounding at 1.91x TBV. Net positive — but the governance half of the ledger is deteriorating, not improving, and the 2027 proxy (board response to the vote) is now a dated checkpoint.

8. Changes and Headwinds — Last Two Years

Update-mode draft. The pre-2026-06-04 arc is condensed from the baseline; the 2026-06-04 → 2026-07-18 events are marked [NEW].

The spine of the story is unchanged: the regulatory penalty box is open, and 2026 is the first unconstrained year. The new six weeks show the bank using the freedom — assets +15% YoY — while returns hit the target band and capital return is re-mixed toward the dividend.

Carried forward (2024–2026-06):

• Asset cap lifted (2025-06-03) — the ~$1.95T Fed restriction removed after seven years; other 2018-order provisions remain.
• Consent-order arc (2025) — 13+ orders terminated since 2019, incl. the 2022 CFPB, 2018 OCC compliance, and 2021 OCC home-lending orders.
• Residual overhang: Sept-2024 OCC AML/sanctions formal agreement (EA #2024-72) — [NEW] still open; no termination or status news through 2026-07-18 (OCC enforcement page; web sweep). The one live enforcement item.
• Strategic narrowing — mortgage retreat, WFAM/corporate-trust/rail divestitures (rail sale closed Q1 2026 [NEW]), late card push.
• Governance (2025–26) — Scharf combined Chair/CEO (Oct 2025); ~$60M special award; [NEW] 2026 say-on-pay 65.5%, independent-chair proposal 33.9%.

[NEW] Event list, 2026-06-04 → 2026-07-18:

• 2026-06-11 — SpaceX IPO raised $75B at $135/sh (~$1.77T, largest ever; leads GS/MS/JPM/BofA — not WFC). Effect: industry-wide IB-fee tailwind that WFC participates in only at the margin — its sub-scale IB (prior report risk #9) caps the benefit; WFC’s Q2 IB fees rose regardless. (venturewiremedia.com 2026-06-20)
• 2026-06-12 — SCHEDULE 13D/A in WFC’s EDGAR feed — a non-event. It is WFC itself (with its Municipal Capital Strategies subsidiary) amending its own 13D on Nuveen Quality Municipal Income Fund MuniFund preferred shares to report a 750-share redemption (~$75M). No activist or outside 13D filer holds a stake in WFC. Effect: none; flagged plainly so it is not misread. (EDGAR 0001193125-26-268969)
• Mid-June — Iran/Strait of Hormuz crisis and de-escalation. Oil spiked then collapsed ~21% to ~$77 after the ~June 14–16 ceasefire/reopening. Effect: volatility tailwind to markets revenue (WFC Q2 Markets +24%); energy-driven inflation reversal sets up the June CPI print. (futuresearch.ai June-2026 CPI analysis)
• 2026-06-24 — 2026 stress test completed; SCB frozen at 2.5%; +11% dividend intent announced. Fed froze SCBs through 2027 (framework revision pending), so results don’t move requirements; WFC’s depletion was below the SCB floor. Effect: removes a capital-distribution uncertainty for two years; the binding constraint is now management’s own 10–10.5% CET1 range, not the Fed. (WFC newsroom; BPI 2026-07-01)
• 2026-07-14 — Q2 2026 print. EPS $2.00 (+25%), ROTCE 17.7% (target band, vs 15.2% a year ago), efficiency ratio 60% (vs 64%), revenue +9%, NCOs 0.34% (−10bp), buybacks $3.0B, CET1 10.3%, TBVPS $46.13 (+7%). Effect: the strongest “changes strengthen the thesis” datapoint since the cap lifted — first quarter inside the 17–18% ROTCE band, with the bear-case falsification metrics (efficiency, ROTCE, credit) all breaking favorably. One quarter is not a trend; the 1H ROTCE is 16.1%. (8-K Ex 99.1/99.2, 2026-07-14)
• 2026-07-14 — June CPI cools to 3.5% YoY / −0.4% MoM (energy −5.7%); Fed Chair Kevin Warsh’s first testimony the same day — hawkish framing (“no tolerance” for elevated inflation; policy rules prescribe higher rates), yet hike odds before 2027 repriced down to ~48%. Effect: a hawkish-hold Fed with falling inflation is NIM-supportive for a liability-sensitive bank — a mild macro tailwind vs the prior report’s “rate cycle turning” headwind; Warsh’s quieter-Fed style raises rate-path uncertainty. (metatradingclub.com 2026-07-14; octagonai.co 2026-07-15; piptheory.com 2026-07)
• 2026-07-14 — JPM blowout Q2 ($21.2B net income incl. $4.6B Visa gain; adj. $16.9B; dividend to $1.65). Effect: raises the money-center bar and the comparison WFC is judged against; WFC’s print held up well on a like-for-like basis (its quarter had no mega one-time gain, only a $132M tax item). (banksandbankers.com 2026-07-15; Quiver 2026-07-14)
• Governance/pay follow-ups [NEW]: nothing new on Scharf comp, succession, or board composition since the 2026 proxy; the Feb-2026 discretionary insider-sale cluster (~$10M, non-10b5-1 per the forms) and the 65.5% say-on-pay vote are the relevant new governance facts (see section 7.6). Effect: the governance discount widens modestly; watch the board’s formal response to the sub-70% vote.

Headwinds watch (updated): (1) OCC AML agreement still open — the last enforcement item; (2) Basel III endgame “finalization of the new proposed capital rules” (Scharf) could raise effective capital requirements and re-tighten the buyback; (3) buyback pace already slowing (~$14B annualized vs $17.5B FY25) as RWA grows +9.5% — the capital-return rate is now growth-dependent; (4) a hawkish Warsh Fed cuts both ways — NIM support now, credit cost later; (5) insider signal (discretionary sales, zero buying) and the pay-vote rebuke point to a governance discount the market currently ignores at 1.91x TBV.

Verdict (INTERPRETATION). The June–July window strengthens the thesis on fundamentals while widening the governance discount. On the operating side, the arc is now delivering, not just promising: a 17.7% ROTCE quarter inside the target band, a 60% efficiency print, clean credit, a two-year SCB freeze, an 11% dividend hike, and capital return explicitly re-anchored to a stated CET1 range — this is the execution the prior report said was unproven, arriving on schedule. The offsets are second-order but real: the last AML agreement lingers, Basel endgame looms over 2027 capital plans, the buyback is decelerating by design, and the governance ledger (65.5% say-on-pay, discretionary insider selling, no insider buying, combined Chair/CEO, no succession signal) moves the wrong way. Net: the changes strengthen the thesis; the burden of proof has shifted from “can WFC execute?” to “does one strong quarter repeat, and does the board absorb the governance message?”

9. Risk Analysis (Risk Matrix)

The risk set is the June baseline’s, re-weighted on thirteen months of evidence: the execution risks that dominated June have partially de-risked (efficiency, ROTCE, CRE), while the cycle-composition, funding-mix, and governance risks have risen in salience. The thesis is now less “will the turnaround happen” and more “how much of the current run-rate survives a normal cycle.”

Narrative on the binding risks (INTERPRETATION). Risks #1–#3 are the thesis-deciders and they travel together: if Q2’s returns were cyclical-composition rather than structure, the multiple paid for ~17% has ~14% of air beneath it toward ~1.6x. The quarterly scoreboard is explicit — ROTCE ex-items, efficiency ex-equity-gains, NII vs the ~$50B guide, and NIB deposit stabilization. Risk #7 is the idiosyncratic one the market prices at zero: a 65.5% say-on-pay vote formally obliges a board responsiveness review, and the 2027 proxy is now a dated checkpoint; for a bank with WFC’s control-failure history, a governance discount is not academic. Risk #8 reframes June’s “thin CET1 buffer” worry into a managed trade-off — capital return is now deliberately subordinated to growth, which is correct capital allocation but caps the buyback-driven compounding the June thesis leaned on. Risks #5, #6, and #10 are the standard late-cycle credit underwritings — currently invisible in the numbers (NCOs 0.34%, coverage 1.40%) and precisely therefore the ones to date-stamp. Risk #11 is the low-probability, high-severity tail: the AML agreement is the last open enforcement item, and any new action would re-rate the governance narrative overnight.

10. Valuation Discussion (Embedded Expectations)

No price target. No recommendation. Valuation is framed as P/TBV-versus-ROTCE comps, embedded expectations, and scenarios — the same method as the June baseline, so the two read directly against each other.

Reference levels (FACT — yfinance 2026-07-18, close 2026-07-17; Q2-26 supplement 2026-07-14): price $87.51; market cap ~$268B; P/TBV 1.90x on Q2-26 TBVPS of $46.13 (recomputed — price ÷ supplement TBVPS); trailing P/E ~12.8x on recomputed TTM EPS of ~$6.84 (FY25 $6.26 − H1-25 ~$3.00 + H1-26 ~$3.58); forward P/E ~11.5–12x (ASSUMPTION — H1-26 actuals of $3.58 plus modest H2 improvement; consensus not pulled from a primary aggregator); dividend yield ~2.3% forward (Q3-26 dividend raised 11% to $0.50/qtr), ~2.1% trailing; 52-week range $72.78–$97.76 (intraday, yfinance).

One feed caveat that matters (FACT → INTERPRETATION). The AZI valuation index (2026-07-17) shows a composite percentile of 86.1 vs. WFC’s own ~10-year history (P/E 81.7, P/B 98.5, P/S 78.1). Two components are contaminated for bank analysis. (a) The P/B leg uses GAAP book per share ($55.45 — itself stale vs. the actual Q2 BVPS of $54.48), not tangible book; for a bank carrying ~$25B of goodwill/intangibles, GAAP P/B is not the multiple this framework runs on, and a percentile rank of it does not describe where the tangibles-based multiple sits. On tangible book, 1.90x sits below WFC’s own January-2026 peak (~2.2x) and inside the 2025–26 range — extended vs. the 2021–24 regime (~1.0–1.5x) but not at a historic extreme. (b) The feed’s TTM EPS ($6.73) lags the Q2 print; on recomputed ~$6.84 the trailing P/E is 12.8x, not 13.0x — the P/E percentile is mildly stale but directionally intact. Net: WFC trades at the high end of its own 10-year history on every basis (FACT), but the 98.5 P/B percentile overstates where the tangibles-based multiple actually sits (INTERPRETATION).

Peer comp table (FACT — prices yfinance 2026-07-18; TBVPS/ROTCE from each company’s latest release as noted):

*USB/PNC/TFC report Q2 ~7/16–7/21; ROTCE carried at FY25 (TFC FY25 12.7% per Morningstar) — flagged as the table’s one stale column. BAC P/TBV on Q1 TBVPS; ~2.09x if Q2 TBVPS ≈ $29.3. Peer TTM P/E rebuilt at live prices from filing/press-based TTM EPS (approximate, ASSUMPTION-labeled). Context (not comps): GS $1,065, MS $216 — capital-markets-geared, different model.

The cross-section is still rational — and WFC is still on the line (INTERPRETATION). Ordered by return: TFC 12.7% → 1.78x, C 13% → 1.28x, WFC 16.1% (H1-26) → 1.90x, PNC ~16.5% → 2.20x, BAC 17% → 2.15x, USB ~18% → 2.33x, JPM 23% → 3.01x. On H1-26 ROTCE, WFC sits on the regression line — fairly priced for what it has printed on average this year. Two shifts vs. June: (1) the whole complex has re-rated upward (USB 1.91x→2.33x, BAC 1.89x→2.15x, JPM 2.89x→3.01x), so WFC’s absolute P/TBV rise to 1.90x is partly beta to a sector move, not idiosyncratic re-rating; (2) on the Q2 run-rate ROTCE of 17.7%, WFC screens modestly below the line — BAC gets 2.15x for 17%, USB 2.33x for ~18%. If 17.7% is durable, the cross-section argues WFC carries a small relative discount again (INTERPRETATION, and a genuinely open question — see below).

Embedded expectations (INTERPRETATION; P/TBV = (ROTCE − g)/(COE − g), COE 10.5% assumed). Back-solving 1.90x: ~18.1% at g = 2%, ~17.2% at g = 3%, ~16.3% at g = 4% — versus 17.4 / 16.6 / 15.8% at the June baseline. The market now capitalizes a sustainable ROTCE of roughly 16.3–18.1%, midpoint ~17% — about half a point to seven-tenths more than six weeks ago, and now squarely at the low end of management’s 17–18% target at mid-g. June’s framing — “priced for only half the recovery” — is stale: the Q2 print (17.7% ROTCE, 60% efficiency, TBVPS +6.8% YoY) plus a ~7% price rise have closed most of the embedded-expectations gap. The tape now believes the target at mid-g; what it does not capitalize is ~18%+ sustained (at g = 4%, spot implies 16.3%, and even the full 18% justifies only 2.15x). Nor does it price a relapse: a fade back to the FY25 14.6% justifies ~1.63x, ~14% of multiple compression from spot (INTERPRETATION).

Re-rating math, updated (INTERPRETATION; COE 10.5%, g 4%). Justified P/TBV by sustainable ROTCE: 14.6% → 1.63x; 16% → 1.85x (≈ spot — at conservative g the price already pays for H1’s 16.1%); 17% → 2.00x (~+5%); 17.5% → 2.08x (~+9%); 18% → 2.15x (~+13%) — all before TBVPS compounding. Compare June: +10% / +15% / +19%. The conditional re-rating upside has narrowed by roughly a third to a half — spent by the re-rating itself. What remains is: (i) ~5–13% of multiple expansion if ROTCE proves out at 17–18% through an easing cycle; (ii) TBVPS compounding at ~7–11%/yr (Q2 TBVPS +6.8% YoY, plus ~$3B/qtr of buybacks at sub-2x-book, which is accretive); (iii) a ~2.3% forward dividend. The JPM mirror-image is unchanged in structure but more extreme: JPM’s 3.01x back-solves to ~23.6–27.6% sustainable ROTCE (g 4%→2%) — at or above even its blowout 23% ex-items Q2 print at every g, i.e., peak capitalized as permanent (INTERPRETATION).

Scenarios (value drivers / implied 3-year annualized TBV total return — NOT price targets) (INTERPRETATION):

• Bear. Q2 was the flattered peak: markets revenue normalizes, the reiterated ~$50B NII guide proves there is no NII growth engine yet, the Fed eases and NIM compresses, CRE/office normalizes into reserve builds, efficiency stalls in the low-60s; ROTCE fades to ~14.5–15.5%; the multiple de-rates to ~1.6–1.7x; TBVPS compounds ~6–7%/yr. Driver: low-single-digit annualized total return — compounding plus dividend roughly offset by ~12–16% multiple compression. Fragile assumption: that 17.7% was cyclical markets-plus-tax-items, not run-rate.
• Base. H1’s 16.1% extends to ~16–17% for FY26–28: cost discipline holds efficiency near 60–62%, fee momentum (IB/trading mid-teens, wealth) offsets flat NII, credit stays mid-cycle, buybacks ~$12–16B/yr at <2x TBV shrink the count ~4–5%/yr; TBVPS compounds ~8–10%/yr; the multiple holds ~1.85–2.05x. Driver: high-single-to-low-double-digit annualized total return, dominated by TBVPS compounding + buyback accretion + dividend, with the multiple roughly neutral.
• Bull. The asset-cap unlock compounds: freed balance sheet funds cards/wealth/markets/commercial growth at target returns, efficiency breaks decisively below 60%, ROTCE sustains 17.5–18%+ through the cycle; the multiple re-rates to ~2.1–2.2x (still a discount to BAC/USB’s current marks); TBVPS compounds ~11–13%/yr. Driver: mid-teens annualized total return = compounding + ~9–13% re-rating + accretion + dividend. Fragile assumption: post-cap growth earns at least the average return, not marginal-return volume — the discipline failure the bear case predicts.

What the market gets right vs. wrong (INTERPRETATION).

• Right: the re-rating itself. Q2 delivered 17.7% ROTCE, 60% efficiency, and a dividend raise; capitalizing ~17% sustainable is no longer faith, it is roughly the H2-25→H1-26 trajectory extended. The discount to JPM remains fully earned.
• Possibly wrong, both directions now: at ~1.90x the market underwrites ~16.3–17.2% sustainable (g 3–4%) — below the Q2 run-rate but above anything WFC has sustained for a full year since the scandal era. The asymmetry that made the June setup compelling (pay ~16%, get 17–18% free) has compressed to a much thinner edge: the price now requires ~17% at mid-g. Upside surprise must come from >18% durability or faster TBVPS compounding; downside surprise is a fade toward 15%.
• The cross-sectional wrinkle worth watching: BAC and USB now carry higher P/TBV multiples for equal or barely-higher returns (2.15x at 17%, 2.33x at ~18%). Either WFC’s governance/mix discount is still being applied (INTERPRETATION), or the regionals are the ones ahead of themselves.

Verdict (INTERPRETATION). At 1.90x TBV the market has moved from skeptic to believer: embedded sustainable ROTCE is now ~16.3–18.1% (midpoint ~17%), versus ~15.8–16.6% in June — the low end of the 17–18% target is now capitalized, not merely optioned. WFC remains on the peer regression line on H1 ROTCE and modestly below it on the Q2 run-rate. The remaining re-rating math is ~+5–13% of multiple against ~−14% of relapse risk, stacked on ~8–10%/yr of underlying TBV compounding — a narrower, more execution-dependent setup than June’s, no longer asymmetric on multiple alone. The AZI 86th-percentile composite overstates the stretch (its P/B leg is a GAAP-book artifact); the honest statement is that WFC trades at the high end of its own history on tangible book and earnings, because its returns are at the high end of their own history.

11. Variant Perception

What consensus believes — updated. The June consensus (“credible turnaround, deserves to trade up as it executes”) has concluded: after the Q2 print — EPS $2.00 vs. ~$1.71 expected, ROTCE 17.7%, efficiency 60%, TBVPS +6.8%, dividend +11% — the sell-side debate has shifted from whether the 17–18% target is real to how much beyond it the model goes. The stock re-rated to ~1.90x TBV, back-solving to ~17% sustainable at mid-g. Consensus now underwrites the target. (INTERPRETATION — built from the Q2 reaction, the re-rated comp set, and post-print commentary; note the one-day −2.7% sell-the-news dip on 7/14 was about the NII miss, not the returns story.)

The strongest bull case. The Q2 print is the start of the run-rate, not a flattered quarter: efficiency at 60% (−400bp YoY) shows the cost machine is still compounding; fee engines (IB/trading mid-teens per Scharf at Bernstein, wealth assets up) are scaling because the asset cap is gone; buybacks at sub-2x TBV plus a rising dividend compound TBVPS at double-digit rates; and the cross-section already pays BAC/USB more for the same returns — WFC re-rates toward 2.1–2.2x as 17–18% prints stack up. In one line: the target is hit, the multiple still lags peers for equal returns, and compounding does the rest.

The strongest bear case. Q2 was the best quarter of the cycle wearing a turnaround costume: markets/trading revenue is a windfall environment (JPM’s equities +86% says the tide, not the boat), NII ex-Marks is flat with a reiterated ~$50B guide that missed in Q2, $132M of discrete tax items flattered the EPS beat, and the easing cycle compresses NIM from here. ROTCE settles back to 15–16%, the multiple paid for 17% at mid-g, and ~1.6x is where 15% prices. In one line: the market just finished paying for the target precisely when the tailwinds that produced it are most cyclical.

What the tape and factor positioning are pricing (FACT loadings → INTERPRETATION). FactorsToday (pull 2026-07-18): WFC’s return stream is ~75% factor-explained, dominated by DividendYield (+1.29), the Financials sector (+1.06), Banks industry (+0.53), and the “Wall Street Giants” custom basket (+0.34) in the All-Factors model, with positive rate sensitivity (+0.26) — and a momentum loading of ~zero (−0.003 Base / −0.12 All-Factors). The 2023–2026 advance is value- and dividend-factor-carried, not momentum-carried: this is not a crowded momentum trade vulnerable to a momentum crash, and there is no momentum-riding-low-quality setup (Quality loading ~flat). The regime is currently mildly supportive: Value (+14.3%, z +1.70) and DividendYield (+16.8%, z +1.66) have been the year’s winning style factors and are still positive over the last month; meanwhile the Financials sector factor is −11.7% over 252d (z −1.58) — financials have lagged the market over the year, and WFC’s relative strength is −5.1% YTD vs. the index despite the absolute recovery. Positioning read: an un-crowded, factor-supported recovering uptrend inside a wide ~$73–$96 2026 range — price above rising 21/50/200-day EMAs, +19% off the May low, −8.2% off the January high, 6-month return flat. Neither a one-way street nor a falling knife; the residual disagreement between bull and bear is therefore fundamental (is 17.7% durable?), not positional. (INTERPRETATION — loadings/returns are FACT; the durability question is the market’s open bet.)

The 3–5 assumptions that decide it (INTERPRETATION).

1. Is 17.7% the run-rate or the peak? H1-26 is 16.1%; the bear says Q2’s markets windfall and tax items are non-recurring; the bull says efficiency at 60% is structural. This is now the load-bearing assumption — the price underwrites ~17%.
2. NII ex-Markets. The reiterated ~$50B guide and the Q2 NII miss: does post-cap balance-sheet growth turn NII positive in 2027, or does easing compress NIM first?
3. Efficiency below 60%. Q2 printed exactly 60%. A sustained break into the 50s validates the bull’s operating-leverage path.
4. Credit, esp. CRE/office. Provisions benign so far; a normalization into reserve build is the bear’s fastest route to 15% ROTCE.
5. Post-cap deployment discipline. Growth at target returns vs. volume at marginal returns — the seven-years-pent-up temptation.

What would change our mind, each way. Bull confirmation: Q3–Q4 ROTCE holding ≥17% with efficiency <60% and NII ex-Markets inflecting positive — at which point the 2.08–2.15x justified band and the BAC/USB relative-discount argument take over. Bear confirmation: ROTCE fading toward 15% on normalizing markets revenue with NII missing the ~$50B guide — the multiple then has ~14% of air beneath it to ~1.6x. Scoreboard, quarterly: ROTCE print, efficiency ratio, NII vs. guide, CRE/office charge-offs.

Verdict (INTERPRETATION). The variant perception has moved: June’s gap — a market paying for ~16% against a 17–18% target — has largely closed, both because Q2 delivered 17.7% and because the price now embeds ~17%. What remains genuinely non-consensus is narrower and sharper: the tape says “target achieved, hold it,” the cross-section (BAC/USB at higher multiples for equal returns) says a modest relative discount persists, and the factor read says positioning is not the risk — no momentum crowd to unwind, value/dividend factors still at its back. The honest variant question is no longer “will the market believe?” but “is Q2’s 17.7% the new floor or the cycle’s ceiling?” — and the price now leaves roughly symmetric room for each answer (~+5–13% re-rating vs. ~−14% relapse), carried either way by high-single-digit TBV compounding.

12. Fact vs. Interpretation Table

13. Open Questions

1. Is 17.7% the floor or the ceiling? Two or more quarters of ≥17% ROTCE excluding equity-gain noise would settle it; H1’s 16.1% is the honest current marker. (The load-bearing question of the whole report.)
2. Does sub-62% efficiency hold without $847M of venture marks? Q2 printed 60% headline, ~62.8% adjusted; the FY26 ~$55.7B expense guide implies continued discipline — verifiable at Q3.
3. NII trajectory vs the ~$50B guide. Q2 NII missed street expectations and needs ~$12.8B/quarter in H2 vs $12.3B printed; does volume + fixed-rate repricing + day count deliver, and when does NII ex-Markets inflect?
4. The markets-financing trade’s true profitability. Does the low-spread build-out (~$200B added since end-2024) produce the claimed NIB-deposit and fee cross-sell at client level, or just NIM-dilutive volume? Management’s defense is asserted, not yet banked.
5. NIB deposit stabilization. Mix fell to ~24% (from 38% in 2021); management guides to “stable” with costs “inching up” — where is the floor, and what does it imply for through-cycle NIM?
6. HTM unrealized losses. Not disclosed in the Q2 supplement and the 10-Q is not yet filed; AOCI widened to $(8.3)B as rates backed up. The fully-marked tangible book is modestly overstated by an unknown amount until the 10-Q.
7. Card/auto vintage seasoning. New card accounts +46–60% and auto originations +41% season into losses with a lag — do 2026 vintages perform like the 8.5%-reserved existing book?
8. Basel III endgame finalization. WFC expects ~7% RWA relief (a potential ~80bp CET1 release) — but finalization could also tighten; management explicitly sets capital plans “absent” it. The 2027 swing factor for the buyback.
9. Board response to the 65.5% say-on-pay vote. A sub-70% vote obliges a responsiveness review; the 2027 proxy (and any succession signal for the combined Chair/CEO) is the dated checkpoint.
10. OCC AML/sanctions formal agreement (Sept-2024). Still open with no status news — the last enforcement item; timing and terms of termination unknown.
11. Durability of CIB share gains. LevFin #3, ECM +74bp, US M&A #4 from #9 — but won in the hottest capital-markets quarter in years; do they survive a normal one?
12. Peer comp refresh. USB/PNC/TFC Q2-26 ROTCE and BAC Q2 TBVPS landed around 2026-07-16/17 but were not primary-sourced at writing; the comp table carries them at FY25/Q1 values (footnoted in section 10).

14. What Must Be True

For the BULL case to be right (from here):

• Q2 2026 is the start of the run-rate: ROTCE holds ≥17% ex-one-offs through H2 2026 and 2027, with efficiency sustained ≤62% ex-equity-gains.
• The asset-cap deployment earns at least the corporate average return: the markets-financing build-out converts into NIB deposits and fee cross-sell at client level, and NII ex-Markets inflects positive in 2027.
• Fee engines (IB, markets, WIM) keep taking share through a normal cycle, not just a boom one; NIM stabilizes after Q3 as guided.
• Credit stays benign as card/auto vintages season; the CRE release remains vindicated.
• The multiple re-rates to ~2.1–2.2x (closing the residual BAC/USB relative discount) as 17–18% prints stack up.
• Falsification test: two consecutive quarters of ROTCE below ~15.5% excluding one-time items, or efficiency back above 64%, or FY26 NII missing the ~$50B guide, or a CRE/card reserve rebuild — any confirms Q2 was the flattered cyclical peak and breaks the bull case.

For the BEAR case to be right (from here):

• Q2’s 17.7% was composition, not structure: markets revenue normalizes, venture marks don’t repeat, the tax rate normalizes, and ROTCE fades to ~15–16% by 2027.
• The low-spread growth proves to be marginal-return volume; NIM keeps compressing toward 2.3% and NII stalls near ~$50B into 2027.
• Card/auto vintages and/or the shrunken office book force reserve rebuilds as the cycle turns; the Warsh Fed’s hawkish hold becomes hikes into a credit turn.
• The multiple de-rates toward ~1.6x (what ~15% justifies), erasing ~14% of the price.
• Falsification test: two consecutive quarters of ≥17% ROTCE with efficiency ≤62% excluding equity-gain noise and NII ex-Markets growing — this demonstrates the improvement is structural run-rate, not cyclical composition, and defeats the bear.

Both cases now turn on the same single fulcrum — whether 17.7% is the new floor or the cycle’s ceiling — and the scoreboard is quarterly and concrete: ROTCE ex-items, efficiency ex-gains, NII vs. guide, NIB stabilization, and vintage-level credit. The June tests resolved in the bull’s favor on all three legs; the tests above are their forward successors, re-aimed at the one question June could not ask: is it durable?

15. Source Appendix

See Appendix B below for the full, tiered, deduplicated source list and the 27-row claim-verification table. The trailing 60-month SEC filing corpus is maintained as a local SEC corpus mirror, extended with the 2026-07-14 Q2 earnings 8-K and Exhibits 99.1 (release), 99.2 (supplement), and 99.3 (deck). All Q2 2026 figures are from those 8-K exhibits — the Q2 2026 10-Q was not yet filed as of 2026-07-18 and figures should be re-tied to it when it lands. Primary sources take precedence: the Q2 2026 8-K exhibits; the Q2 2026 earnings-call transcript; the FY2025 10-K; the Q1 2026 10-Q; the 2026 DEF 14A; the 2026-04-30 say-on-pay 8-K; the Form 4 corpus; and the Federal Reserve’s 2025-06-03 asset-cap-removal release. Quantitative data was reconciled to EDGAR XBRL; ROIC.ai, AZI, FactorsToday, and yfinance are flagged unofficial aggregators. Peer money-center framing is cross-read from prior reports on JPMorgan, Bank of America, Citigroup, U.S. Bancorp, PNC, and Truist. Prior internal materials were not re-accessed for this update and are treated as historical context only.

Independent research. No buy/sell recommendation and no price target outside the labeled Kimi’s Take block; valuation is framed as embedded expectations and scenarios.

APPENDIX A — Standard Diligence Questionnaire

Supplement to the WFC research memo (UPDATE, 2026-07-18). Answers a standard fundamental diligence checklist; supplemental and excluded from the ~20-page length target of the main memo. Applies the Competition Demystified (Greenwald) barriers-to-entry lens and Capital Returns (Marathon) capital-cycle lens. Update-mode: answers emphasize what changed versus the 2026-06-04 baseline. No price target, no BUY/SELL. Facts/Interpretation/Assumption labeled where it matters. For a bank, “free cash flow” questions are answered with the correct sector analogs (ROTCE, NII, CET1, tangible book, capital return).

General

What thoughtful questions have other investors asked about this company?

• Is Q2 2026’s 17.7% ROTCE the new run-rate or the flattered peak of the cycle? (the post-print debate — the stock dipped on the NII miss even as returns hit the target band)
• How much of the beat was venture-capital marks and tax items? (~two-thirds of the ~$0.27 headline beat; core EPS ~$1.80–1.85)
• Is the low-spread markets-financing build-out actually profitable at client level, or is WFC buying share with its balance sheet? (Scharf: “we’re either gonna get paid for it or we’re not gonna do it” — asserted, not yet banked)
• When does NII ex-Markets inflect, and is the ~$50B FY26 guide safe after the Q2 miss?
• Why do BAC and USB now trade at higher P/TBV multiples for equal returns — is WFC’s governance/mix discount still being applied?
• Has the cost program structurally broken the efficiency ratio below 62%, or does it drift back without equity gains?
• What does the board do about a 65.5% say-on-pay vote, and who succeeds the combined Chair/CEO?
• Does the buyback keep decelerating as post-cap RWA grows — and what does Basel endgame do to 2027 capital plans?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? (INTERPRETATION.) Cyclically elevated — this is the load-bearing caveat of the update. Q2 2026 was arguably the best industry quarter in years: record capital-markets activity (industry IB fees up ~26% per KBW’s estimate), benign credit everywhere (WFC NCOs 0.34%, BAC 0.47%, PNC 0.25%), and a hot IPO/issuance tape (SpaceX’s $75B debut). WFC’s 17.7% ROTCE includes ~1–1.5 points of one-offs (venture marks, discrete tax, CRE release); the underlying ~16–16.5% against H1’s 16.1% is the cleaner marker. The June baseline called earnings “mid-cycle and turnaround-driven”; the update shifts that to upper-mid-cycle with the turnaround doing most of the lifting — the idiosyncratic recovery (efficiency 66%→60%, post-cap growth) is real, but it is now being flattered by the cycle rather than fighting it.

Are earnings driven primarily by the external environment or internal company actions? Still more internal than most banks — the dominant swings (asset-cap deployment, the cost program, ROTCE remediation) are WFC-specific. But the mix shifted this quarter: the external environment (markets windfall, hot IB tape, benign credit) contributed materially to Q2’s optics, which is precisely why the durability question now dominates.

How stable are revenues? Improving in quality on the trend, more cyclical in the quarter. The durable streams (deposit spread, WIM asset-based fees, card/payments) all grew double digits; fee mix rose to ~46% of revenue from ~43% — structurally stabilizing. But Q2’s fastest-growing lines (Markets +24%, IB +35%, +$728M venture-gain swing) are the most cyclical, and NII ex-Markets grew only 2%.

Outlook for products and services? Broadly positive and now unconstrained: balance-sheet growth (loans +12%, deposits +10%), cards (new accounts +46–60%), auto (originations +41%), WIM ($2.69T client assets, +15%; Advisor Gateway + AI copilot shipping), CIB share gains (LevFin #3, US M&A #4 from #9). Mortgage remains deliberately de-emphasized.

How big will this market be — growing, shrinking, domestic or international? US-domestic, GDP-plus at best structurally — but WFC’s relevant “market” expanded idiosyncratically: the asset cap’s removal returned ~$200B+ of balance-sheet capacity it had been forced to ration, so WFC’s addressable opportunity is currently share recapture, a one-time normalization layered on a mature industry (Marathon: favorable for the incumbent inside the regulated perimeter; the oversupply forms outside it — private credit, non-bank lenders).

Business Quality & Competitive Moat

Is the industry getting more or less competitive? (INTERPRETATION.) More competitive at the top in the short run — every large bank is adding markets balance sheet, bankers, and traders into the hot tape, and WFC itself is re-entering businesses it rationed — while the regulated-oligopoly structure (top-4 hold ~40%+ of deposits) remains intact. Marathon’s read: capacity is being added late-cycle, which seeds the next downturn’s losses; the discipline question is who stays in-lane (Scharf claims WFC is declining the frothier AI/data-center credit).

How profitable is the business (ROIC, ROE)? (FACT/INTERPRETATION.) Q2 2026: ROE 15.0%, ROTCE 17.7% (H1 16.1%; FY25 14.6%), ROA 1.15%. Underlying (ex-venture/tax) ~16–16.5%. Against a ~10.5% cost of equity (ASSUMPTION), the franchise now clears its hurdle mid-cycle — narrowly on H1 figures, comfortably on the Q2 run-rate. Segment ROACs: WIM 32.4% (best), CB&L 27.1% (flattered by an allocated-capital methodology change), CIB 19.2% (from 14.9%), Commercial 17.2%.

How profitable is the industry — how many competitors, what barriers to entry? The protected tier earned Q2-26 ROTCEs of 13–23% (JPM 23% ex-items, USB 18.7%, PNC 17.9%, WFC 17.7%, BAC 17.0%, TFC 15.4%, C 13.0%) while the ~4,000-bank tail earns far less. Barriers are regulatory (charters, capital rules, G-SIB surcharges) plus deposit-franchise captivity — the highest entry barriers of any major industry, which is exactly why a below-average industry houses above-average franchises.

Can the business be easily understood? Yes at the model level (borrow cheap, lend dear, plus fees); the complexity lives in the accounting (ACL, AOCI/HTM marks, fair-value venture gains) and in the wholesale businesses now driving growth (markets financing, derivatives, trading), which are less transparent than the retail core.

Can it be undermined by foreign low-cost labor? No — domestic regulatory perimeter and relationship banking are not labor-arbitrageable. The real “low-cost” threats are non-bank: private credit, fintech payments, and stablecoin/tokenized deposit rails — slow, structural, and currently second-order to WFC’s own execution.

Do brands matter? Moderately. The Wells Fargo brand survived a franchise-damaging scandal (deposits held through 2016–2025 with no mass attrition — a fact that upgrades the captivity assessment), but in wholesale/IB, brand = league-table position, where WFC is a challenger, not an incumbent.

What is the nature of competition? Customers’ switching costs? Oligopolistic rivalry at the top, commoditized at the product level. Switching costs are real in primary checking (13 consecutive quarters of consumer checking growth — Greenwald demand-side captivity), treasury management, and advised assets; weak in markets/IB, where share follows balance-sheet commitment and pricing — the businesses WFC is now growing fastest.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? (INTERPRETATION.) The deposit franchise itself (~$1.47T at ~1.51% average cost) is the great unbooked asset. Against it: AOCI losses widened to $(8.3)B as rates backed up, and HTM unrealized losses are not disclosed in the Q2 supplement (10-Q not yet filed — OPEN QUESTION); tangible book is modestly overstated on a fully-marked basis by an unknown amount until the 10-Q lands.

Off-balance-sheet liabilities? Standard large-bank set: unfunded commitments, derivatives notionals, operating leases, and the residual regulatory/litigation tail (the Sept-2024 OCC AML agreement is the live item). Nothing unusual surfaced in the Q2 exhibits.

How conservative is the accounting? Mixed-to-conservative. Positives: ACL 1.40% of loans with 1.69x-ish coverage of nonperforming assets was maintained while releasing into falling charge-offs; the office book was allowed to run off −20% rather than being warehoused; SBC is modest and more than offset by buybacks. Watch items: the CRE reserve cut (2.50%→1.77%) is a directional bet that has so far been right; the $847M venture-gain line is real but Level-3-adjacent and lumpy; CB&L’s ROAC was flattered by an allocated-capital methodology change ($45.5B→$33.0B) — a reminder that segment returns are partly allocation artifacts.

How CapEx-hungry is the business? Not capital-expenditure-hungry in the industrial sense — the analogs are (i) regulatory capital intensity: RWAs grew +9% YoY and consumed CET1 (11.1%→10.3%), so growth “costs” capital at roughly 10 cents of equity per dollar of RWA; and (ii) technology spend: tech expense +13% YoY, including the >$1B multi-year WIM platform modernization. The honest statement: WFC’s growth is now capital-hungry by design — earnings retention is being redirected from buybacks into balance-sheet growth.

Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy? FCF is N/A for a bank; the analog is net income minus capital consumed by RWA growth. Philosophy under Scharf: (1) fund profitable balance-sheet growth first, (2) maintain a ~26–28% dividend payout (just raised 11% to $0.50/qtr — back to the pre-2020-cut run-rate), (3) return the excess via buybacks inside a disclosed 10–10.5% CET1 operating range (new this quarter — resolves the June report’s open question #3), (4) divest sub-scale/non-core businesses (12 since 2019; the rail JV closed Q1 2026), (5) no acquisitions. 1H26: ~$9.8B returned ($7.0B buybacks + ~$2.8B dividends, ~88% of net income).

Significant acquisitions recently? None — WFC remains a divestor, not an acquirer. This remains a genuine relative strength versus historical money-center behavior.

Buying back shares? Yes, but deliberately decelerating: $3.0B in Q2 (~$14B annualized vs ~$17.5B FY25), at ~1.9x tangible book. Refinement to the June framing: above 1x TBV, buybacks are mildly TBVPS-dilutive (~0.4% per $1B) but EPS- and ROTCE-accretive — the accretion runs through returns math, not tangible-book compounding, and it reverses sign if ROTCE stalls.

Issuing large amounts of new shares to insiders? No. SBC ~7% of net income, more than offset by repurchases; share count −6% YoY to 3,028.5M (−25%+ over five years).

Compensation policy of directors/management? ROTCE-aligned in architecture (PSAs: 3-yr absolute ROTCE 75% + relative ROTCE 25% + relative-TSR modifier; absolute target raised 100bp for 2026 grants) — but the governance ledger deteriorated this window: 2026 say-on-pay passed with only 65.53% support (a failed grade obliging a board responsiveness review), the June report’s 92.4% figure was a stale 2025 number; the ~$60M one-time CEO special award, eliminated pre-set comp targets, the combined Chair/CEO role, and no quantified risk-remediation metric in the pay formula all stand.

Motivations of management? (INTERPRETATION.) Financially: ROTCE-target attainment and relative-TSR. Reputationally: Scharf’s legacy is “the man who got Wells Fargo out of the penalty box” — achieved; the forward incentive is proving the franchise can grow responsibly, which cuts both ways (discipline so far; the seven-years-pent-up growth temptation the bear case warns about). The insider tape is not encouraging: zero open-market buys in 110 filings, and a February 2026 cluster of ~$10M discretionary (non-10b5-1) sales by three senior executives near the 52-week high.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? No — US domestic common (NYSE: WFC), standard 1099.

Dividend policy? $0.50/qtr expected from Q3 2026 (+11%), ~$2.00 annualized ≈ 26–28% payout; ~2.3% forward yield at $87.51. Policy: conservative payout with the residual via buybacks, all inside the 10–10.5% CET1 range.

How profitable is the business? See moat section above: Q2 ROTCE 17.7% headline, ~16–16.5% underlying, H1 16.1% — versus a 14.6% FY2025 base and a 17–18% medium-term target reaffirmed on the Q2 call.

Is net income diverging from cash from operations? For a bank the meaningful divergence checks are different, and two are live: (i) one-offs in earnings — ~$1B of Q2’s pre-tax upside (venture marks + tax) is non-run-rate, so annualizing the headline overstates run-rate earnings; (ii) unrealized marks — AOCI $(8.3)B and undisclosed HTM losses mean tangible book overstates fully-marked economics by an amount pending the 10-Q. Provisions (non-cash) running below charge-offs is the benign direction of divergence, but it is also a reserve-release tailwind that cannot persist indefinitely.

Risks & Downside

What factors would cause the stock to decline? In order of thesis-impact: (1) ROTCE fading toward 15% as markets revenue and venture marks normalize — the multiple paid for ~17% has ~14% of air beneath it to ~1.6x; (2) NII missing the ~$50B guide with NIM compressing through 2.4%; (3) a credit turn exposing the seasoned card/auto vintages or the shrunken office book; (4) a governance event (board response to the pay vote mishandled, a compliance relapse under the open AML agreement); (5) sector de-rating — the whole complex re-rated since June, and WFC’s absolute rise is partly beta.

Risk of a catastrophic loss? Low. A G-SIB with 10.3% CET1 (8.5% minimum+buffers), 0.34% NCOs, 1.40% coverage, and a frozen 2.5% SCB does not fail absent a systemic event; the realistic catastrophe is regulatory, not financial — a new enforcement action re-imposing growth restrictions, which is why the open AML agreement matters beyond its size.

Chance of a total loss? De minimis for the equity over any normal horizon — this is a too-big-to-fail franchise with a century of operating history; the realistic bear case is a ~15–25% drawdown (de-rating to ~1.5–1.6x TBV on a ROTCE fade), not impairment.

Recent News & Events

Has the business environment changed recently? Yes — three ways: (i) Q2 2026 print (2026-07-14) — the thesis-confirming event covered throughout this report; (ii) the rate/inflation regime inverted — from two-cuts-priced in January to a hike priced after the spring Iran/Hormuz oil shock (May CPI 4.2%), then June CPI cooled sharply to 3.5% YoY / −0.4% MoM under a new, hawkish Fed Chair (Kevin Warsh’s first testimony was 2026-07-14); net NII-supportive versus January assumptions, which is why management held the ~$50B guide; (iii) capital-markets boom — record industry IB/trading quarter, the SpaceX IPO, and WFC share gains within it; plus the early-July collapse of the Iran ceasefire as a live oil/inflation tail risk.

Significant acquisitions? None (rail JV divestiture closed Q1 2026 — the opposite direction).

Change in accounting policies? None identified. One disclosure change of note: the new explicit 10–10.5% CET1 operating range, and a Q1 2026 allocated-capital methodology change that flattered CB&L’s ROAC.

Recent changes — new markets, facilities, management? No management changes since the 2026 proxy (the combined Chair/CEO role stands; no succession signal). Facility footprint continues to shrink (branches −56 YoY to 4,079; headcount −7% YoY, 24 consecutive quarterly declines). Product-side: Advisor Gateway (May 2026) and the AI advisor copilot (July 2026) shipped in WIM; the card and auto origination pushes are the consumer growth engines; the 2026 stress test left the SCB frozen at 2.5% through 2027, and the dividend rises 11% in Q3 (pending board approval at writing).

APPENDIX B — Source Appendix — Wells Fargo & Company (NYSE: WFC) — UPDATE 2026-07-18

Report date: 2026-07-18 (update follow-up to the 2026-06-04 report) Target: Wells Fargo & Company — SEC CIK 0000072971 — NYSE: WFC — GICS Diversified Banks — FYE December Baseline: carried forward from the 2026-06-04 source appendix; entries still relevant are retained, stale entries dropped or relabeled.

Limitations (explicit):

• Q2 2026 Form 10-Q was NOT yet filed as of 2026-07-18 (EDGAR sweep since 2026-06-04 shows no 10-Q; latest 10-Q remains Q1 2026, filed 2026-04-29). All Q2 figures herein are from the 2026-07-14 earnings 8-K exhibits, not the 10-Q.

No price target. No BUY/SELL anywhere outside the labeled Kimi’s Take block.

TIER 1 — SEC Filings & Regulatory (Primary)

TIER 2 — Earnings Releases / Transcripts / Decks

TIER 3 — Data Feeds (aggregator / unofficial — reconcile to filings)

TIER 4 — News / Press (secondary; publisher + date + URL)

SEC corpus notes

• Incremental EDGAR sweep since 2026-06-04 (181 rows incl. debt-shelf noise). Non-noise adds: 1× 8-K (Q2 earnings), 6× Form 4, 1× 13D/A (fund preferred — see 1.8), 1× 13F-HR/A, 1× 11-K. No 10-Q filed since 2026-04-29 (Q2’26 10-Q pending).
• Insider read since 2026-06-04 remains NEUTRAL: zero open-market purchases or sales of WFC common; activity is gifts (G), annual director grants (A), option exercise + withholding (M/F), and a fund-level J-code item.
• A full 5-year local SEC corpus mirror is maintained, extended with the 2026-07-14 8-K + Exhibits 99.1/99.2/99.3.

Carried-forward caveats (still binding)

• ROTCE/efficiency/CET1/TBVPS are filer-reported headline metrics (8-K exhibits), not standalone XBRL line items — verified-by-issuer; NI/EPS/revenue reconcile to the release’s income statement.
• Peer ROTCE/TBVPS figures from the 2026-06-04 report that were secondary-sourced (BAC, USB, PNC) remain footnoted-secondary unless re-pulled from primary sources.
• All yfinance/ROIC/AZI/FactorsToday outputs are unofficial aggregates; the filing wins any conflict.
• Prior internal materials (a Deutsche Bank 2011 industry primer, a prior Bank of America memo, saved transcripts) were not re-accessed for this update and are treated as historical context only.