Capital One Financial Corp

Positioning statement

Capital One is now the largest credit card issuer in the United States by outstanding card loans, and the only major US bank that simultaneously owns a national card issuing franchise and a general-purpose global payment network. The May 2025 acquisition of Discover Financial Services — a $51.8 billion fair-value transaction, the largest in US card history — converted a monoline-derived consumer lender into a vertically integrated payments company controlling both sides of the transaction economics. The company's structural differentiators are threefold: an information-based underwriting culture dating to 1988; a completed migration to public cloud infrastructure, unique among large US banks; and, since 2025, ownership of the Discover Network, PULSE and Diners Club rails. Capital One remains founder-led by Richard D. Fairbank, one of very few founder-CEOs among America's largest public companies. The investment case now rests almost entirely on execution: converting network ownership and $2.5 billion of targeted synergies into durable earnings before integration costs and consumer credit normalisation erode the thesis.

The company's own description (FY2025 Form 10-K, Item 1)

The 10-K describes Capital One Financial Corporation as "a diversified financial services holding company with banking and non-banking subsidiaries" that operates "as a global payments provider and diversified financial institution, delivering a broad array of financial products and services to consumers, small businesses and commercial clients through digital channels, branch locations, cafés and other distribution channels." The filing states that the company "was the largest issuer of credit cards in the United States of America based on the outstanding balance of credit card loans as of December 31, 2025" — a material change of language from the FY2022 and FY2021 10-Ks, each of which described the company as "the third largest issuer of Visa and MasterCard credit cards in the U.S."

The 10-K further records that "through the acquisition of Discover, we acquired new products including personal loans as well as the Discover Network, the PULSE Network, Diners Club International and Network Partners (collectively, the 'Global Payment Network')."

Independent characterisation

Capital One is a spread-lending business with a payments overlay. The economic engine is unsecured consumer revolving credit funded by retail deposits, and the dominant earnings driver is the gap between the yield on card receivables and the blended cost of deposits, less credit losses and marketing.

The scale of that engine is unusual. In FY2025 the Credit Card segment produced $39.56 billion of the company's $53.43 billion of total net revenue — 74.0% — on average loans of $227.3 billion carrying an average yield of 17.95%. Consumer Banking contributed $10.43 billion (19.5%) and Commercial Banking $3.66 billion (6.8%), with the Other category a $214 million drag. Total net revenue margin for the group was 9.77% of average interest-earning assets, roughly three times the level a conventional regional bank generates, because the asset mix is overwhelmingly high-yield consumer credit rather than mortgages and securities.

Revenue model composition. Capital One's revenue mix is emphatically not product-versus-service-versus-subscription in the industrial sense. For FY2025:

• Net interest income: $42,878 million (80.2% of total net revenue). Interest income of $58,696 million against interest expense of $15,818 million. Loans generated $53,021 million of the interest income at a 13.35% blended yield.
• Non-interest income: $10,556 million (19.8%). Of which discount and interchange fees, net of rewards, were $6,443 million (61% of non-interest income); service charges and other customer-related fees $2,849 million; other $1,264 million.

The Discover acquisition has begun to shift this mix. Discount and interchange income rose 32% year-over-year in FY2025 and service charges 44%, both faster than net interest income's 37% — and both now include network economics that did not previously exist inside the group. The 10-K explains the mechanism precisely: "for transactions on Bank-issued credit and debit cards processed on the Global Payment Network, a portion of the amount that merchants pay to the Global Payment Network is passed through to the Bank." In other words, Capital One now captures both the issuer interchange and the network fee on its own on-us volume. This is the core economic rationale for the deal and it is why the debit-card reissuance onto the Discover Network was prioritised as the first integration workstream.

Value chain position. Pre-2025, Capital One occupied a single node: card issuer, paying network fees to Visa and Mastercard. Post-2025 it occupies three: issuer, network operator (Discover Network for credit, PULSE for debit EFT/ATM, Diners Club for international licensee acceptance), and — via Network Partners agreements — a processing and card-issuing services provider to third-party financial institutions and fintechs. With the April 2026 Brex acquisition it added a fourth node: enterprise spend-management software sitting above the card rail. With the April 2026 in-sourcing of the Hopper technology stack it added a fifth: direct travel merchant distribution.

Customer types and end-markets.

Seasonality. The 10-K notes that Credit Card purchase volume and receivables peak around the winter holiday season and that card net charge-off rates "generally tend to be the highest in the first quarter of the year." This is visible in FY2025: the domestic card net charge-off rate was 6.19% in Q1 2025 falling to 4.63% by Q3 2025, and the pattern repeated in 2026 with 5.05% in Q1 falling to 4.71% in Q2.

Stated strategy — themes from the 2026 proxy and FY2025 10-K

The 2026 proxy sets out the board-approved strategic frame verbatim: management is responsible for a strategy designed to allow Capital One to "(i) invest in long-term capabilities and opportunities; (ii) secure competitive, endgame positions in our key businesses; (iii) attract and retain customers; (iv) grow resiliently; (v) promote ethical behavior and compliance with applicable laws and regulations; (vi) withstand economic stress and market volatility; and (vii) adhere to the Board's established risk appetite programs."

The phrase "endgame positions" is the most revealing in the document. It is the internal justification for buying Discover: the belief that in payments, the terminal competitive structure has room for a small number of vertically integrated issuer-networks, and that a scaled issuer without a network eventually becomes a price-taker.

The proxy also records that the Discover transaction "further positions Capital One as a diversified bank and to operate a global payment network, a rare and valuable asset at a time of significant disruption and innovation across the payments landscape," and states that the company "invested in: (i) marketing and product innovation to support growth; (ii) technology, data, and modern infrastructure to support innovation; and (iii) talent and tools to deliver top tier software development, rigorous risk management, and artificial intelligence capabilities."

Fairbank's own framing on the Q1 2026 call: "we still expect our earnings power on the other side of the Discover integration to be consistent with what we expected at the time we announced the deal, inclusive of the Brex and Hopper travel infrastructure."

Strategic initiatives announced in the last 24 months

Management's medium-term financial targets

Sustainability and ESG commitments are covered in Section 20.

Caveat: Capital One does not publish an exhaustive product register in its filings. The catalogue below is assembled from the FY2025 10-K, the Q2 2026 10-Q, corporate boilerplate and company product pages. Pricing is included only where the company or a primary source discloses it. Where a product's current status could not be verified, it is flagged. The full co-brand partner roster is not individually verified against filings and is therefore not enumerated exhaustively.

5.1 Credit Card segment

Domestic consumer card — travel and premium

Domestic consumer card — cash back and access

Refreshed FICO distribution on the domestic card book at 31 Dec 2025: 73% above 660, 27% at or below 660 — versus 69%/31% at 31 Dec 2024. The Discover portfolio materially improved the credit mix.

Discover-branded card products (acquired May 2025)

The Discover it family — Cash Back, Miles, Chrome, Student and Secured variants — continues to operate under the Discover brand. The migration path is now public: new originations move to Capital One's technology stack by end of Q3 2026; back-book conversion waves run July 2026, October 2026 and January 2027; full technology-stack migration is expected by Q1 2027. Discover cardholders will be required to establish Capital One accounts, which management has flagged as the largest customer-facing impact of the integration. Fairbank characterised the Discover card portfolio as being in a temporary "brownout" during Q2 2026, with balances contracting approximately 1.5% year-over-year while marketing is deliberately restrained ahead of conversion.

Personal loans. Acquired with Discover. $9,499 million period-end balance at 31 Dec 2025; $6,061 million FY2025 average; FY2025 net charge-off rate 3.83%; 30+ day performing delinquency 1.74% at year-end — the best-performing asset class in the Credit Card segment.

Small business cards. The Spark family (Spark Cash Plus, Spark Miles, Spark Classic) serves domestic small business card lending, reported within Credit Card. From Q2 2026, Brex corporate cards also contribute to Credit Card purchase volume.

International card businesses. UK portfolios originated through Capital One (Europe) plc and Canadian portfolios through the CONA Canada Branch. Combined period-end loans $7,668 million at 31 Dec 2025 (+11% year-over-year). FY2025 net charge-off rate 5.14%; 30+ day performing delinquency 4.62%.

5.2 Consumer Banking segment

Deposits

Total deposits were $475,771 million at 31 Dec 2025 and $484,257 million at 30 Jun 2026. Interest-bearing deposits rate paid fell from 3.16% in Q4 2025 to 3.00% in Q1 2026 and 2.91% in Q2 2026. The 2026 proxy discloses that approximately 85% of total deposits were FDIC-insured at year-end 2025.

Auto Finance. National indirect and direct auto lending. Period-end loans $83,600 million at 31 Dec 2025 and $89,300 million at 30 Jun 2026 (+4% in the quarter). FY2025 originations $40,996 million, up 19% year-over-year; H1 2026 originations $24,046 million, up 20%. At-origination FICO distribution at 31 Dec 2025: 51% above 660, 19% between 621 and 660, 30% at or below 620 — a three-point shift into the sub-620 bucket over the year, which is the single clearest evidence of deliberate credit-box expansion anywhere in the disclosure. FY2025 auto net charge-off rate 1.54% versus 2.05% in FY2024. Auto Navigator is the pre-qualification and vehicle search platform; Chat Concierge, launched in early 2025, is the generative-AI car-buying assistant covering vehicle research, test-drive scheduling and financing exploration.

Global Payment Network (reported within Consumer Banking). Four components, per the FY2025 10-K:

• Discover Network — "processes transactions for credit and debit cards issued on its network and provides payment transaction processing and settlement services."
• PULSE Network — "operates an electronic funds transfer network, providing financial institutions issuing debit cards on the PULSE Network with access to ATMs domestically and internationally, as well as merchant acceptance throughout the U.S. for debit card transactions."
• Diners Club International — "a global payments network of licensees, which are generally financial institutions, that issue Diners Club-branded charge cards and/or provide card acceptance services."
• Network Partners — agreements with financial institutions, fintechs, networks and commercial service providers "for the provision of card issuing, payments processing and related services."

Global Payment Network volume: $74,014 million in Q2 2025 (partial period from 18 May), $153,117 million in Q3 2025, $174,644 million in Q4 2025, $401,775 million for FY2025, and $189,612 million in Q2 2026 — up 156% year-over-year and 9% sequentially, with H1 2026 at $363,944 million.

Branch, café and ATM network. The 10-K describes servicing "banking customer accounts through digital channels and our network of branch locations, cafés, call centers and automated teller machines." Precise branch, café and ATM counts as of 31 Dec 2025 were not located in the FY2025 10-K extract reviewed and are flagged as not verified in this pass.

Capital One Travel. Brought fully in-house in April 2026 through the acquisition of Hopper's technology stack, licences, servicing contracts, supplier relationships and approximately 150 employees. A dedicated Capital One Travel app was launched. Hopper continues to supply specific fintech products (Flight Disruption Guarantee, Price Freeze) but no longer powers the portal. Capital One Lounges provide airport lounge access for premium cardholders. Capital One Shopping is the browser-extension savings tool. Capital One Cafés combine retail banking with hospitality.

Digital tools. Eno is the virtual assistant and card-security tool. CreditWise is the free credit monitoring product available to non-customers as well as customers.

5.3 Commercial Banking segment

CONA is registered with the CFTC as a swap dealer, subject to OCC capital and margin rules and CFTC business conduct, recordkeeping and reporting requirements. It is not registered with the SEC as a security-based swap dealer.

5.4 Capital One Software (enterprise B2B)

Based in McLean, Virginia. Positioned by the company as "backed by 25 years of data innovation," addressing "cost performance, infrastructure management, and data security" in the cloud. Two commercial products:

Capital One Software revenue is not separately disclosed in the segment reporting and is therefore not publicly quantified.

5.5 Brex (acquired 7 April 2026)

Described in the acquisition release as "a modern, AI-native software platform offering intelligent finance solutions that make it easy for businesses to issue corporate cards, automate expense management and make secure, real-time payments. The company also leverages AI agents to help customers automate complex workflows to reduce manual review and control spend." Founded 2017 by Pedro Franceschi and Henrique Dubugras. Approximately 25,000–35,000 business clients (sources differ) and approximately $13 billion of customer deposits held at partner banks or in money market funds at the time of the deal. Franceschi continues to lead the business. Brex began contributing to Capital One's domestic card purchase volume in Q2 2026; Q2 2026 Brex integration expenses were $96 million pre-tax ($0.12 per diluted share).

Methodological note on banks. Gross profit, EBITDA, EV/EBITDA, working capital, inventory-based asset turnover and the cash conversion cycle are not meaningful measures for a bank holding company and are not reported by Capital One. The analogous measures used throughout are total net revenue (revenue net of interest expense), pre-provision earnings (total net revenue less non-interest expense), the efficiency ratio and the operating efficiency ratio. Where the brief requests a non-applicable metric, this is stated rather than a proxy being silently substituted.

Comparability warning. FY2025 includes Discover from 18 May 2025 only. Any FY2024-to-FY2025 growth rate is acquisition-inflated. FY2026 will be the first full-year comparative, and even that includes Brex only from 7 April 2026.

Table 6.1 — Income statement, five years (USD millions except per-share and %)

FY2021 basic EPS of $27.11 and diluted share count of 444.1 million are analyst-derived from net income available to common and reported EPS; the reported diluted EPS of $26.94 is source-verified. FY2022–FY2025 share counts are as reported.

Table 6.2 — Profitability and efficiency ratios (%)

FY2021 and FY2025 efficiency ratios are company-reported (54.44% and 57.08%). FY2022–FY2024 are analyst-computed as non-interest expense divided by total net revenue; FY2024 company-reported is 54.93%, confirming the method.

Table 6.3 — Company-reported non-GAAP adjusted measures

Table 6.4 — Balance sheet, five years (USD millions, period-end)

Balance-sheet series sourced from S&P Global Market Intelligence via StockAnalysis, cross-checked for FY2024 and FY2025 against the Q4 2025 financial supplement (total assets, deposits, equity, goodwill, intangibles, allowance all reconcile exactly). Working capital is not a meaningful bank metric and is not reported.

Table 6.5 — Cash flow, five years (USD millions)

Table 6.6 — Key ratios (%, analyst-computed on period-end balances unless noted)

Company-reported comparatives where available: return on average assets 0.99% (FY2024) and 0.35% (FY2025); return on average common equity 8.08% (FY2024) and 2.03% (FY2025); return on average tangible common equity 11.18% (FY2024) and 3.16% (FY2025). Net interest margin FY2021 6.21% and FY2024–FY2025 as reported; FY2022 and FY2023 are indicative and should be re-verified against the respective 10-Ks. Net charge-off rates FY2024 3.39% and FY2025 3.30% are as reported; earlier years are indicative. Return on invested capital, interest coverage, current ratio, asset turnover and the cash conversion cycle are not meaningful for a bank holding company and are not reported by Capital One.

Table 6.7 — Growth (%)

Commentary on trends, inflections and drivers

Revenue. The four-year revenue CAGR of 15.1% is dominated by a single step: FY2025's 36.6% increase, of which the overwhelming majority is Discover consolidation. Organic revenue growth through FY2022–FY2024 ran at 12.5%, 7.4% and 6.3% — decelerating, and driven principally by domestic card balance growth and rate-cycle net interest margin expansion rather than volume in the banking businesses. Consumer Banking revenue actually contracted in FY2023 and FY2024. This deceleration is a substantial part of the strategic logic for buying Discover.

Net interest margin is the standout operating story. NIM expanded from 6.88% in FY2024 to 7.84% in FY2025 (+96bp), reached 8.36% in Q3 2025, and stood at 8.01% in Q2 2026 (+39bp year-over-year). The FY2025 expansion is mix — replacing lower-yielding assets with Discover's high-yield card book — but the 2026 progression is genuine: deposit costs are falling (interest-bearing rate paid 3.16% in Q4 2025 to 2.91% in Q2 2026) faster than asset yields. Management guided on the Q2 2026 call to a further NIM catch-up in Q3 2026 as average cash balances align with lower ending cash and day-count works in the company's favour.

The FY2025 earnings collapse is accounting, not economics. Pre-provision earnings — the cleanest read on underlying franchise capacity — rose 30% to $22,936 million. Provision rose $8.9 billion to $20,655 million, of which the day-two Discover build accounts for the great majority. GAAP net income fell 48% to $2,453 million while adjusted net income available to common nearly doubled to $10,615 million. The Q2 2025 quarter alone recorded a $4,277 million net loss and a $(8.58) diluted EPS. Any multi-year model that treats FY2025 GAAP EPS as a run-rate base is wrong by roughly a factor of five.

Credit is the genuinely encouraging line. Group net charge-offs of 3.30% in FY2025 were nine basis points better than FY2024 despite absorbing a substantially larger and initially unseasoned book. The 30+ day delinquency rate fell 39bp to 3.59%. Domestic card charge-offs fell from 5.91% (FY2024) to 5.12% (FY2025), and continued to improve into 2026 — 5.05% in Q1 2026 and 4.71% in Q2 2026, with the domestic card 30+ day delinquency rate at 3.39% in June 2026, down 31bp sequentially and 21bp year-over-year. The Q2 2026 quarter delivered a $662 million allowance release. Management stated credit trends were similar across the legacy Capital One and legacy Discover portfolios — an important confirmation that the acquired book is not underperforming.

The one credit line moving the wrong way is auto. The auto net charge-off rate rose from 1.25% in Q2 2025 to 1.82% in Q4 2025 and 1.48% in Q2 2026 (versus 1.30% a year earlier), while the at-origination sub-620 FICO share rose from 27% to 30% of the book across 2025. Auto balances grew 4% in Q2 2026 alone. This is deliberate credit-box expansion into a weakening lower-income consumer, and it is the most under-discussed risk in the story.

Expenses are the live controversy. Non-interest expense rose 42% in FY2025; operating expense rose 45%. The efficiency ratio deteriorated 215bp to 57.08% even as the adjusted efficiency ratio improved 219bp to 51.81%. Integration costs have already exceeded the original $2.8 billion guidance — American Banker reported in July 2025 that Capital One had spent nearly $10 billion on Discover integration on a cumulative basis, and Fairbank confirmed costs would exceed the original figure. FY2025 Discover integration expenses were $1,109 million; Q2 2026 added $298 million of Discover and $96 million of Brex integration expense. Management has also flagged that technology and investment spending will rise further as Brex and the in-sourced travel platform are absorbed.

Capital and funding. The balance sheet is the strongest it has been in the period. CET1 rose from 13.5% (Dec 2024) to 14.3% (Dec 2025) and 14.4% (Mar 2026), before falling 70bp to 13.7% in Q2 2026 on $2.7 billion of buybacks and the Brex close. Tangible common equity ratio improved 210bp to 10.7%. Net debt turned to a $5.3 billion net cash position in FY2025 for the first time in the period. Q4 2025 average LCR was 173%; Q2 2026 average LCR 165% and NSFR 136%, with $144.1 billion of total liquidity reserves at 30 June 2026. Management stated on the Q2 2026 call an intention to run toward an 11% CET1 need over the long term — implying roughly 270bp, or approximately $14 billion, of currently excess capital at Q2 2026 risk-weighted assets.

Operating cash flow rose 52.6% to $27,718 million in FY2025 and to $31,452 million on a trailing-twelve-month basis at 30 June 2026. Capital expenditures rose 31% to $1,578 million, consistent with heavy technology investment; computer software gross carrying value rose from $3,357 million to $4,902 million over the year.

Management guidance

Capital One does not provide formal EPS guidance. The disclosed commitments are:

• $2.5 billion of total net synergies, approximately one-third realised at Q2 2026, remainder by the second half of 2027. Composition: roughly $1.3–1.5 billion of expense synergies (targeting elimination of 25% of Discover's operating expenses and 10% of its marketing expenses) and roughly $1.2 billion of revenue synergies.
• Approximately 15% adjusted EPS accretion in 2027 as the deal-model outcome, calibrated to a normalised CET1 of 12.5%.
• Discover front-book migration complete by end of Q3 2026, enabling full-spectrum underwriting on the Discover brand for the first time. Back-book conversion waves July 2026, October 2026 and January 2027. Full technology-stack migration by Q1 2027.
• Long-term CET1 need of approximately 11%, with capital viewed as having "asymmetric value for both defensive and offensive flexibility."
• Net interest margin catch-up expected in Q3 2026 as average cash balances align with lower ending cash and day-count works favourably.
• Rising technology and investment spending, including efficiency ratio pressure, as Brex and the travel platform are integrated.
• Integration costs above the original $2.8 billion guidance, with cumulative spend reported near $10 billion.

Consensus expectations

Twenty-four analysts carry a consensus Buy rating with an average price target of $255.19–$256.50, a 14–17% premium to the current price. Forward P/E is 10.18x. The dispersion — UBS at $280, Citi at $295, consensus at $255 — reflects genuinely divergent synergy-timing assumptions rather than disagreement about the franchise.

Bull case

1. The market is capitalising a CECL artefact as if it were earnings power. FY2025 GAAP EPS was $4.03; adjusted was $19.61. The trailing twelve months to 30 June 2026 produced $16.16 of GAAP EPS and $31.5 billion of operating cash flow. At $223.83 the stock trades at 10.2x forward earnings and 4.6x free cash flow — pricing that assumes the synergy target substantially fails.

2. Two-thirds of the synergy target is still ahead, and the delivered third worked. The debit conversion completed on schedule and is now producing its full quarterly run-rate. Consumer Banking revenue rose 26% year-over-year in Q2 2026 as a direct consequence. If the credit-card leg delivers proportionally, the remaining $1.7 billion arrives across 2027 against a current pre-provision earnings base of roughly $27 billion annualised.

3. Credit is improving into a supposedly weakening consumer. Domestic card charge-offs fell from 6.19% (Q1 2025) to 4.71% (Q2 2026); delinquencies fell 21bp year-over-year; Q2 2026 delivered a $662 million allowance release; and management confirmed legacy Discover credit is behaving like legacy Capital One. Payment rates remain above pre-pandemic levels. The bear thesis on the middle-income consumer has not yet shown up in the data.

Bear case

1. Costs are running well ahead of plan and the guidance has already been revised once. Original integration guidance of $2.8 billion; cumulative spend reported near $10 billion; the efficiency ratio deteriorated 215bp in FY2025; management has pre-announced further increases for Brex and travel. The $2.5 billion synergy target is gross of an integration bill that is now open-ended.

2. The credit-card network migration — where the majority of remaining synergy value sits — is the hard part, and it has barely started. Debit was straightforward because acceptance is near-universal for PIN debit. Credit is not: management has explicitly acknowledged domestic acceptance gaps and named four international markets requiring work. Meanwhile the Discover card book is in a self-inflicted "brownout," contracting 1.5% year-over-year, and three conversion waves through January 2027 each carry attrition risk requiring customers to establish new Capital One accounts.

3. Concentration plus credit-box expansion into a fragile borrower base. 74% of revenue from one segment; 27% of the domestic card book at or below 660 FICO; and the sub-620 share of auto originations rose from 27% to 30% during 2025 while auto charge-offs rose year-over-year. Layer on regulatory tail risk — the Credit Card Competition Act, proposed rate caps, Regulation II amendments, and the possibility that three-party network debit becomes subject to Regulation II, which the 10-K explicitly flags as capable of undermining the deal's benefits — and the downside distribution is fatter than the multiple implies. The Fitch Rating Watch Negative, unresolved since May 2026, is a signal from a party with access to the same data.

Catalysts and monitorables for the next 12 months

Analyst verdict

Capital One is, at this moment, one of the most mispriced-looking large-cap financials in the US market — and one of the hardest to underwrite with confidence, which is precisely why it looks mispriced.

The facts are not in dispute. The company is the largest US card issuer by outstandings. It owns the only general-purpose payment network in American hands outside American Express. It has $484 billion of cheap, overwhelmingly insured deposits, a 530-basis-point CET1 cushion above requirement, roughly $14 billion of capital above its own stated long-term need, and $31.5 billion of trailing operating cash flow. Credit is improving, not deteriorating. Net interest margin is at a record 8.01% and management expects it to rise again. Adjusted return on tangible common equity was 18.5% in FY2025. The stock trades at 10.2 times forward earnings.

The reason it trades there is that nobody can yet distinguish between two futures. In the first, the credit-card network migration works the way the debit migration worked, the acceptance gaps close, the $2.5 billion arrives on schedule, integration spending rolls off in 2028, and a business earning close to $20 of adjusted EPS with an 18% ROTCE re-rates toward the mid-teens multiple that franchise deserves. In the second, acceptance proves to be the structural constraint it has always been for Discover, conversion attrition compounds the brownout, integration costs keep exceeding guidance the way they already have once, and a portfolio with 30% of auto originations below 620 FICO meets a genuinely weakening lower-income consumer.

Two things should discipline the judgment. First, the reported financials are close to uninformative right now — FY2025 GAAP EPS of $4.03 against adjusted EPS of $19.61 is not a rounding difference, and any analysis built on the GAAP series is simply wrong. Second, management's track record on this specific question is good: Fairbank has made large asymmetric bets before, on marketing in the early 2000s, on the premium card push, and on the cloud migration, and has largely been right. Against that, integration cost guidance has already been missed by a wide margin, which is the one place where the promise-versus-delivery record on this deal is already established.

The honest position is that the next four quarters resolve it. The front-book migration in Q3 2026, the October and January conversion waves, and the first production routing of credit volume onto the Discover Network are not incremental data points — they are the thesis. Investors who require the answer before committing will pay a materially higher price. Investors who buy here are underwriting execution risk at a discount, with a fortress balance sheet, an accelerating buyback and improving credit as the compensation for being early.

END OF DOSSIER

Prepared 15 August 2026. All financial data traceable to the FY2025 Form 10-K (filed 19 February 2026), the Q1 2026 and Q2 2026 Forms 10-Q, the quarterly 8-K earnings releases and financial supplements, the 2026 DEF 14A (filed 25 March 2026), and market data current to 14 August 2026. Items flagged as "not verified in this pass" require retrieval from the underlying primary source before being used in a transaction, credit or investment context. Where third-party aggregator data conflicts — notably institutional ownership percentages, tangible book value per share on a Q2 2026 basis, and ESG emissions data — both the conflict and the sources have been disclosed rather than a single figure being selected.

Covering approximately February 2025 to August 2026.