US Bancorp (USB)

Business Overview

US Bancorp pairs super-regional scale with money-centre-like returns on tangible equity. The company is not a G-SIB, one of the globally systemically important banks required to carry extra capital, yet FY2025 return on tangible common equity was 18.1%, second only to JPMorgan Chase's 20% among the six banks covered here. Market capitalisation was roughly $87.0 billion as of 9 June 2026, a fraction of JPM's ~$834B, but its returns still cluster with the large-bank names.

The P/TBV vs ROTCE guide explains how an 18.1% return is read against the cost of equity.

The operational differentiator is cost. FY2025 efficiency ratio was 58.6%, below our 60% screen that separates good cost discipline from average large-bank execution. Only JPM's 52% overhead ratio is lower. The efficiency guide holds USB up as the super-regional benchmark against Wells Fargo at 66%.

How the Franchise Earns

FY2025 NIM was 2.72%, solid if not top of the ladder: above BAC's 2.01% net interest yield and JPM's 2.50% net yield, below PNC's 2.83%. Net interest margin alone does not explain 18.1% ROTCE; spread, fee businesses, and a sub-60% efficiency ratio together do.

On credit, the net charge-off ratio was 0.57% for FY2025, middle of the pack: below JPM's 0.74% and just under the 0.62% industry figure for the same year, above BAC's 0.50% and WFC's 0.43%. Nothing there looks like stress. Charge-offs lag, though, so the thing to watch if commercial books soften is the provision, which moves first.

Capital is where not being a G-SIB pays. On the Basel III standardised approach the CET1 ratio ended FY2025 at 10.8%. USB's requirement is a 4.5% minimum plus a 2.6% stress capital buffer, so 7.1% in total, with no G-SIB surcharge on top because USB is not designated one. That leaves +3.7 percentage points of headroom, wider than any money-centre bank here despite a lower headline ratio, which is the point the CET1 guide makes: read the ratio against its own bar.

Valuation Framework

For a bank like USB, value tracks return quality more than balance-sheet heft. The steady-state identity sets ROTCE against the cost of equity, 10.5% in the model, and at 18.1% USB clears it comfortably on efficiency leadership and sub-industry charge-offs.

TBVPS of $29.12 at 31 December 2025 is the per-share capital anchor, built by retained earnings. Buying stock back above tangible book takes tangible book per share down, so repurchases add value only if the price paid is below what an 18.1% return is worth. On that test USB has more room than the mid-teens names.

What to Watch in the Financials

Efficiency holding below 60%. At 58.6%, only JPM runs leaner; drift toward the low 60s without revenue growth would compress ROTCE toward BAC/WFC territory.

ROTCE versus 18.1%. High-teens returns are what separate USB from the mid-teens pack; a slide to mid-teens would erase that.

CET1 at 10.8% against a 7.1% bar. The buffer is set at 2.6% until October 2027, because the Fed has frozen stress capital buffers while it reworks its stress models. So the requirement is unlikely to move, and it is USB's own risk-weighted asset growth that would narrow the gap.

NIM at 2.72%. Spread is healthy but not PNC-leading. Deposit beta disclosures in quarterly decks matter more than the static FY average as rates move along the planning path.

Key Risks

Scale versus G-SIBs. USB competes on efficiency; it cannot match a G-SIB balance sheet. Large-bank pricing on deposits and payments can compress fees even when costs stay disciplined.

Commercial credit normalisation. NCO at 0.57% is benign; a regional commercial book can turn faster than the lagging ratio suggests. Provisions lead charge-offs under CECL.

Capital rules can move against a non-G-SIB. The 3.7 pp cushion rests on a 2.6% stress capital buffer that reflects one year's model run. When the Fed finishes revising its stress framework, a harsher result for USB's commercial book would tighten distribution capacity without anything changing in the business.

Peer catch-up. If money-centre efficiency programmes succeed (WFC in particular), USB's relative advantage narrows even if absolute ROTCE holds.