Capital One (COF) Q1 FY26 earnings call: guidance and Q&A

With energy-price concerns rising, why do card and auto credit results remain so strong, and what are you seeing across the consumer?

Fairbank said the U.S. consumer remained healthy and the economy resilient: unemployment improved slightly, income growth outran inflation, spending was robust, and tax refunds are higher. Card metrics improved YoY, delinquencies ran a bit better than seasonality, and auto losses (slightly higher YoY on subprime mix) remained near pre-pandemic levels. He flagged the new Persian Gulf conflict and a ~6-week energy-price spike as a real potential headwind, noted March inflation ticked up on gas prices, but said no adverse portfolio effects have appeared and elevated macro risk was judgmentally added to the allowance via qualitative factors.

“There's a lot of noise in the external environment, but the consumer is showing quite a bit of resilience.”

asked by Terry Ma · answered by Richard Fairbank

How should the efficiency ratio migrate as Brex/Hopper enter the run rate, and how do elevated liquidity levels trend and affect NIM going forward?

Fairbank said Brex and Hopper are not yet in the efficiency ratio and marketing will be heavier over the year, all reflected in efficiency across multiple line items, but reiterated post-integration earnings power is intact. Young framed NIM: Discover added ~85 bps and back-half 2025 set a new structural level; Q1's 39 bps drop was fewer days (~18 bps), seasonal card paydown and elevated cash (home-loan sale, strong deposits, favorable tax flows). Cash should trend down (~$8B Q2 debt maturities plus tax payments), day-count adds ~9 bps in Q2 and again into Q3/Q4, and the structural post-Discover NIM should persist.

“if you look at the back half of the year, that on a seasonally adjusted basis is a pretty good indication of where you should expect NIM to kind of structurally be.”

asked by Sanjay Sakhrani · answered by Andrew Young

Why not put out efficiency parameters like in 2019, and can you help the market understand what 'consistent with deal expectations' is measured against?

Fairbank declined explicit efficiency guidance, framing Capital One as an organic-growth, opportunity-driven company with a large investment agenda that generally does not guide. Pressed on the starting point, he clarified that 'earnings power' means ROTCE at a constant capital level, that the deal model assumed 12.5% capital, and that the guidance is normalized to that 12.5% level — adding the guidance would still hold at somewhat higher capital levels (current CET1 is 14.4%), though he would not quantify the breakeven.

“When we're talking in the form of this guidance when we say earnings power, we're talking about ROTCE... The capital level assumed in the deal model was 12.5%.”

asked by Ryan Nash · answered by Richard Fairbank

How should we think about card-business growth over the next year given the Discover volume, and is there a way to think about Brex payback periods?

Fairbank said the legacy branded card book is growing strongly (top-of-league among booked-up-market customers), while the 'elephant in the room' is the Discover brownout: following Discover's 2022–23 expansion and subsequent pullbacks, plus Capital One's trims to high-balance revolvers, Discover outstandings fell 1.2% YoY and the brownout will deepen until tech integration completes. New originations convert by end of Q3 2026 and the back book by Q1 2027, unlocking growth thereafter (personal loans also in temporary brownout). On Brex, he described an enablement strategy — lower cost of funds and brand benefits arrive quickly, marketing and travel benefits phase in later — but warned that more traction means more investment and thus deferred vertical financial impact.

“these brownouts are a natural and temporary part of the deal and have been accompanied by a better credit and even some margin strength along the way.”

asked by Moshe Orenbuch · answered by Richard Fairbank

How could the Basel III reproposal (standardized vs ERBA) affect RWA, and how does that shape capital allocation? Also confirm the ROTCE/constant-capital point.

Young said the firm is ~$680B in assets, ~$20B below the $700B Category 2 threshold (a 4-quarter trailing average, possibly indexed up), so triggering is likely far off. The reproposal would add ~20 bps to CET1 fully phased-in on standardized: ~140 bps RWA tailwind (RWA down 8–9%, similar under ERBA given its ops-risk charge) versus a ~120 bps AOCI headwind on $5.2B AOCI; they have begun using held-to-maturity and don't plan to elect ERBA (DTA threshold falls 25%→10%). On capital actions, he cited conservatism and asymmetrical value of capital behind the $2.5B repurchase. Fairbank confirmed 12.5% deal-model capital and that guidance can hold at somewhat higher capital.

“the effect under the standardized approach for us if it were enacted on a fully phased-in basis today would increase our CET1 by something like 20 basis points.”

asked by L. Erika Penala · answered by Andrew Young

How are you thinking about AI-driven job-loss risk, and do you build anything into credit underwriting for that?

Fairbank gave a wide-ranging view, comparing today's AI anxiety to the industrial and digital revolutions and arguing people underestimate economic dynamism and job elevation; he leans optimistic on employment implications but declined to prognosticate. On credit, he acknowledged credit is tightly linked to employment and any large unemployment shift would matter, but said Capital One is not making credit-policy choices now in anticipation of AI job loss, relying instead on 3–4 decades of modeling history and a deliberate resilience buffer to adapt.

“We certainly are not making credit policy choices now in anticipation of things like that.”

asked by Donald Fandetti · answered by Richard Fairbank

Learnings from the debit conversion, timing of credit conversion and issuing on the Discover network, and when do integration expenses wind down and synergies come through?

Fairbank said the debit conversion is complete and reinforced confidence in doing these conversions; card-side network work is in early testing (originating on the Discover network now, moving a portion of the credit book next year), with investment in international acceptance, brand and technology to turn the network flywheel. Young said revenue synergies are front-loaded (tied to the substantially complete debit conversion, meaningful portion already in Q1, fuller in Q2), while expense synergies are back-loaded to tech-platform conversions/decommissioning and won't be fully realized until conversions complete in first half of 2027 — with the full $2.5B of synergies still expected by mid-2027.

“we still feel very good about achieving the full $2.5 billion of synergies by the time we complete integration in the middle of '27.”

asked by Mihir Bhatia · answered by Andrew Young

Why not increase buybacks above $2.5B/quarter given a comfortable 14.4% CET1 and excess capital even after Brex and standardized/AOCI?

Young reiterated the decision factors (current/projected capital, balance-sheet growth, regulatory environment, valuations, macro) and stressed a conservative, resilience-focused posture and the asymmetrical value of capital, noting nearly $12B of remaining authorization and SCB flexibility while committing only to using flexibility in the moment. Fairbank added that despite the conservative framing, share repurchases are an important part of Capital One's value-creation equation. Separately, Young confirmed the Discover fair-value mark amortization schedule is final (~$1M NIM impact this quarter, inconsequential) and directed analysts to the previously provided tables.

“we're always going to err on the side of conservatism and a focus on resilience... we are well aware that capital has asymmetrical value in certain environments.”

asked by Saul Martinez · answered by Andrew Young