Independent M&A Teardown · Edition 01 · 2026

Capital One / Discover: 1-Year Retrospective

Capital One / Discover. One year on, the case has shifted.

The $35B all-stock deal closed May 2025. Twelve months of post-close disclosures now sit next to the announce-day case. This is a retrospective on whether management's story is still intact, built from the S-4, the Fed and OCC approval orders, the Q2 2025 10-Q, and four post-close earnings calls.

Deal value$35.3B
Consideration100% stock
Closed18-May-2025
Pro forma assets~$660B
Premium paid26.6% (above 2022-23 avg)
P/E paid7.8x FY24 net income
P/TBV paid2.0x
Pro forma CET114.0% (up from 13.6% standalone)

01

The headline

Management guided to >15% Y2 EPS accretion. My realistic case is 3.3%.

The Feb 2024 deck guided to $2.7B in run-rate synergies by 2027 and more than 15% accretion to adjusted EPS in Year 2. My three-case model, anchored to a normalized 2027 standalone COF net income of ~$8.1B (i.e., FY2024's $4.8B normalized for credit roll-off), says even management's own announce-day case only delivers ~13% accretion. Layer in what Year 1 actually showed (a $4.3B GAAP loss in Q2 2025, the synergy framing quietly shifting from "$2.7B by 2027" to "$2.5B net by mid-2027," and an open-ended OCC remediation overhang) and my realistic case lands at ~3% accretion in 2027.

Case2027 A/(D) %Assumptions
Mgmt (announce)+13.0%$2.7B · 100% phase
Base (revised)+8.6%$2.5B · 85% phase
Haircut (mine)+3.3%$1.9B · 65% phase

Source: author's model. COF standalone 2027 NI anchored to consensus of approximately $8.1B.

Mgmt's headline >15% number was almost certainly built on a more conservative standalone denominator than consensus. 2025 and 2026 are dilutive in all three cases because integration cost runs through both years.

02

Why this deal

The first major U.S. bank to acquire a combined issuer and payment network.

For most of the U.S. consumer card market, the network sits between the issuing bank and the merchant. Visa and Mastercard rails carry the volume for Cap1, Citi, Chase. American Express has always been the closed-loop exception, owning both ends. Discover was the rarer case: a card-issuing bank that also operated its own payment network. COF/DFS is the first time a major U.S. bank holding company has acquired that combination.

What pushed me to actually build the model rather than just write about the announcement is that the deal closed twelve months ago. You can put the announce-day case next to a year of post-close disclosure and see how the story has held up. Most teardowns are done at announce. This one can be a retrospective.

03

Methodology

How an announce-day case becomes a retrospective in five steps.

01. The announce-day case comes off the deck.

$2.7B run-rate synergies, >15% Y2 EPS accretion, ROIC 16%, IRR >20%, $2.8B integration cost. Lifted verbatim from the Feb 19, 2024 joint investor presentation.

02. The standalone projections get anchored to a normalized baseline.

COF FY2024 net income was $4.8B, depressed by a $10.9B provision build. Normalizing for credit roll-off implies a 2027 baseline of ~$8.1B. DFS continuing-ops at $4.5B to $5.0B over the same window. These are stated inputs, not tuned to flatter the answer.

03. The pro forma balance sheet comes from the Q2 2025 10-Q.

Reported goodwill $13.2B, card-relationship intangible $10.3B, network intangible $3.1B, and a ~$9B Day-1 fair-value markdown to DFS's loan book. These are not estimates.

04. The synergy realization is split into three cases.

Management's announce case ($2.7B, 100% phased by 2027). Base case using the revised $2.5B net framework with a one-year slip. My Haircut case at $1.9B (~30% off announce) with another year of slip, reflecting OCC remediation drag.

05. Year 1 actuals do the testing.

$4.3B Q2 2025 GAAP loss, $9.4B in Discover-related charges that quarter, $409M of H1 integration spend, the $1.225B merchant restitution settlement landing post-close, and Fairbank's "somewhat higher than $2.8B" admission on integration cost.

04

The deal in numbers

100% stock, but not what the headlines suggested for capital.

COF issued ~252.8M new shares (DFS shares × 1.0192 exchange ratio) to acquire Discover. The 26.6% premium against DFS's $110.49 unaffected price sits above the post-pandemic U.S. bank M&A average; Mercer Capital reports ~20% in 2022 and ~17% in 2023 on five-day announce premiums for >$100M deals.

The conventional read on a $35B all-stock deal is that pro forma CET1 takes a beating from goodwill and intangible deductions. Per COF's Q2 2025 10-Q, purchase accounting produced $13.2B of goodwill, $10.3B in purchased credit-card relationships, and $3.1B in network intangibles, all of which are large deductions from regulatory capital. And yet pro forma CET1 came in at 14.0%, up from COF's standalone 13.6% the prior quarter. The equity issued to DFS shareholders added more to the capital base than the deductions consumed. Unusual for a deal this size, and largely absent from the sell-side notes I read at close.

Purchase price allocation and capital$B / %
Equity consideration to DFS$35.3B
Goodwill$13.2B
Card-relationship intangible$10.3B
Network intangible$3.1B
Day-1 loan FV markdown (feeds Q2'25 allowance build)~$9.0B
Pro forma CET1 (Q2 2025)14.0%
Standalone CET1 (Q1 2025)13.6%

Source: COF Q2 2025 10-Q and earnings release.

05

The synergy bridge

From "$2.7B by 2027" to "$2.5B net by mid-2027." Then my haircut.

$2.7B
Announce headline
$1.5B expense + $1.2B network synergies. Run-rate by 2027 per the Feb 2024 deck.
$2.5B
Q2 2025 framing
"Total net synergies as previously discussed." Same target re-described as net of investments. The 2027 finish line slipped to mid-2027.
$1.9B
My case
~30% haircut from announce. Reflects regulatory remediation displacing realization, network buildout capex offsetting revenue uplift years 1 to 3, and integration drag timing.

Whether the move from $2.7B to $2.5B is a re-statement or a quiet haircut depends on how you treat the additional integration investments now sitting outside the synergy box. Either way, the headline number got smaller and the timeline slipped.

06

Year 1 actuals

What the post-close disclosure has actually said.

$4.3B Q2 2025 GAAP loss

The first consolidation quarter. Total Discover-related charges that quarter ran to $9.4B (loan loss reserve build, integration costs, deal costs combined).

Synergy target shifted, no new finish line

From "$2.7B in expected synergies by 2027" (Feb 2024 deck) to "$2.5B in total net synergies by mid-2027" (Q2 2025 transcript). The 2027 framing has dropped out of the disclosures.

$1.225B merchant restitution at close

Tied to DFS's historical card misclassification. Preliminary court approval 30-Jul-2025. Plus a $100M Fed civil money penalty and a $150M FDIC penalty, all costs that landed on COF's books post-close.

Integration cost "somewhat higher than $2.8B"

Fairbank's Q2 2025 wording. No revised number given. H1 2025 spend was $409M.

OCC remediation overhang

The OCC's conditional approval requires COF to submit a remediation plan for outstanding Discover Bank enforcement actions. Open-ended cost, not in the announce-day synergy math. The $265B Community Benefits Plan (announced Jul 2024) layers five more years of opex commitments on top.

The capital surprise

Pro forma CET1 expanded to 14.0% at Q2 2025, up from COF's standalone 13.6%. Almost nobody wrote about this. Equity issuance was larger than goodwill drag, which is unusual for a deal this size.

07

My view

The deal is accretive. By less than what was guided.

The synergy framing has moved.

"$2.7B by 2027" became "$2.5B net by mid-2027," and the additional investments needed to unlock the network-side revenue are now positioned outside the synergy target. That is a real reframing: a piece of what used to be the cost stack of synergy realization is now a separate, open-ended bucket. The 2027 finish line slipped.

DFS came with a regulatory tail.

The merchant restitution alone is $1.225B, the Fed and FDIC penalties add another $250M, and the OCC remediation plan is open-ended. None of these were sized in the announce-day synergy math, and they sit in opex rather than in one-time charges.

Integration cost is running hot.

Fairbank's "somewhat higher than $2.8B" is the most candid signal. The cost is front-loaded into 2025 and 2026, which is why my model shows both years as dilutive even before any synergy haircut.

The bull case, and why I still disagree.

Network volume migration is structurally durable once it happens. COF began reissuing debit cards onto the Discover network in pilot in June 2025, with phased rollout into early 2026. The expense bucket has reasonable precedent: BAC/MBNA targeted $850M in after-tax expense synergies by end-2007 and was tracking on plan through Q3 2006 ($357M pre-tax run-rate). Similar consolidation logic. If integration runs cleaner than I assume from 2027 onwards, accretion above 10% by 2028 is plausible. I just do not see the path to a 2027 print above 10% given what has already been disclosed.

The buyback counterfactual.

$35B at the announce price would have repurchased ~60% of COF's float. Not a realistic single-step move past the SCB and 4(c), but the right framing for whether the deal compounded value or just bought scale. The straight reading is that it bought scale, and the synergy story has to do most of the value-creation work for that to have been the right trade.

08

What to watch and what I would have flagged

Three forward signals. Four risks I would have raised at announce.

Forward signals

Network volume migration

Quarterly disclosure of COF debit and credit volume migration onto Discover's rails. The single hardest synergy to fake on a slide.

Synergy guidance refresh

Whether the $2.5B net target gets re-affirmed, lowered, or quietly dropped on 2026 calls.

OCC remediation milestones

Open-ended cost crystallizes when COF files completion plan progress.

What I would have flagged at announce

DFS regulatory exposure

The card-misclass settlement (which turned into $1.225B post-close) and DFS Bank's standing FDIC/Fed/CFPB enforcement actions. Both disclosed in the 10-K notes pre-deal.

Approval gauntlet

The deal was announced in Feb 2024 under hostile FTC scrutiny of bank consolidation. Approval required Fed + OCC + DOJ + state AGs, all needed and none guaranteed at announce. The deal arguably got through because the regulatory environment shifted between the announce and the approval, a timing dependency worth flagging.

Network re-platforming complexity

First time a major U.S. bank has owned both an issuer and a payment network as a combined entity (AmEx aside).

Goodwill impairment risk

$13.2B of goodwill plus $13.4B of intangibles against a Q2 2025 GAAP loss; one bad quarter from a real impairment headline.

09

Resources

The note, the model, the sources.

2-page research note

The recruiter-shareable PDF.

Open PDF →
20-sheet M&A model

Cover toggle, standalone P&Ls for COF and DFS, combined NI build, sources & uses, purchase price allocation, intangible amortization, full pro forma balance sheet, capital walk-down to 2028, quarterly synergy build, A/D output with Earnings Impact Footnotes 1-5 (including breakeven synergies), Analysis at Various Prices, stacked data tables, contribution analysis, two-way sensitivity, comparing structures, precedent transactions, trading comps, football-field valuation, methodology.

Download XLSX →

Primary sources