U.S. Bancorp USB

Financial Services · Banks - Diversified · Synthos Deep Dive · 2026-08-04

$62.47

Hold

—

The Overview

U.S. Bancorp is a large, ordinary, well-run American bank. It takes in $528 billion of deposits from households and businesses, lends about $400 billion of it out, and keeps the difference. It also runs a sizeable payments business — the machines that take your card in a shop, and corporate purchasing cards.

Three things you want to know about a bank, and this one answers all three well. Does it have enough capital to absorb losses? Yes: it holds 10.8% capital against a 7.1% requirement. Are its loans going bad? No: it wrote off 0.56% of loans last quarter, less than the 0.59% a year earlier. Is it funded safely? Yes: three-quarters of the balance sheet is ordinary customer deposits, which is the safest funding a bank can have.

Profits are growing at about 11% a year, mostly because the bank is earning a slightly wider spread on its loans and holding costs flat. The dividend rises from 52 to 54 cents a quarter this autumn, which is about 3.4% a year on today's price.

The problem is the price. The shares have risen 46% in the last year — nearly twice the market — and now cost $64.23, which is within half a percent of the highest price they have reached in a year. Analysts on average think they are worth $69.33, about 8% more. Our own estimate is $67, about 4% more. And the most pessimistic analyst covering the stock thinks it is worth $61, which is less than it costs today.

So: a good bank, fully paid for. If you own it, the dividend and the earnings growth make it reasonable to keep. If you do not own it, you are being asked to pay a full price for a business growing at 11% a year, with roughly as much room to fall as to rise. That is a Hold, and we would rather say so than dress it up.

One honest note about our own research: the Synthos knowledge base contains 51,928 recorded expert claims and not one of them is about this company. The three that matched "USB" were all about the computer cable.

Downside Risk 5/10. The balance sheet is not the risk; the entry price is.
Growth Quality 5/10. Real, modest, and largely margin rather than volume.
Exponential Potential 2/10. A linear, cyclical, well-managed spread-and-fee business.

Putting a number on it: our fair-value estimate is $67 against a current price of $62.47 — real upside if our numbers are right.

Our summary metrics

Downside Risk (lower = safer)5/10Moderate

"Rated 5 — a conservatively run, deposit-funded, geographically diversified bank whose principal risk today is the price rather than the balance sheet. The supports, all filing-verified from the March 2026 10-Q: a common equity tier 1 ratio of 10.8% at both 2026-03-31 and 2025-12-31 against a requirement of 7.1% (4.5% minimum plus a 2.6% stress capital buffer that the Federal Reserve confirmed on 2026-06-24 would remain unchanged until 2027-10-01); total deposits of $528.2 billion funding approximately 75% of total assets; an allowance for credit losses at 2.00% of period-end loans and 536% of nonperforming loans; and total loan net charge-offs of 0.56% of average loans in Q1 2026, DOWN from 0.59% a year earlier, with the decrease driven by lower credit-card losses. Against that: the equity is priced for continuation. The stock is 0.4% below its 52-week high after a 46.2% twelve-month move; the street's median target of $68 is 5.9% above spot and the LOW target of $61 is 5.0% BELOW it, so the whole published distribution spans roughly minus 5% to plus 17%. Residential mortgages 90+ days past due or nonperforming ran at 4.05% of residential mortgages at 2026-03-31 including GNMA repurchase-option loans, against 3.37% at 2025-12-31 — a deterioration the company attributes to the mechanics of the GNMA pool but which is the one credit series in the filing moving the wrong way. Commercial loan net charge-offs rose. And a pending acquisition of BTIG for up to $1 billion, announced January 2026 and expected to close in Q2 2026, appears nowhere in the vendor payload and is not confirmed closed by anything in this filing archive."

Growth Quality5/10Moderate

"Rated 5 — real, modest, and mostly margin rather than volume. On the basis U.S. Bancorp actually publishes (net interest income plus noninterest income, which is the vendor's revenue line less its own interest expense), total net revenue was $28,013M in FY2023, $27,335M in FY2024 and $28,540M in FY2025 — that is minus 2.4% then plus 4.4%, not the plus 5.1% and plus 0.3% the vendor's gross-up implies. The trajectory has improved sharply in the last three quarters: rebuilt quarterly net revenue runs $6,975M (Q2 2025), $7,300M, $7,337M, $7,260M and $7,686M in the June 2026 quarter — the last of those is plus 10.2% year on year and the highest in the series. Net income followed: $2,177M in Q2 2026 against $1,815M a year earlier, plus 19.9%. Diluted EPS was $1.18 in Q1 2026, plus 14.6%, and $1.35 in Q2 2026 against a $1.28 estimate. The drivers are legible and disclosed: net interest margin 2.77% against 2.72%, average loans up $14.5 billion (3.8%), noninterest expense up only 0.8% year on year in Q1. Consensus wants $5.215 in FY2026 (16 analysts), $5.816 in FY2027 (16) and $6.397 in FY2028 (8) — 11.5% then 10.0%. What holds this at 5: this is a $700 billion balance sheet growing loans at 3.8%, and roughly half the EPS growth is margin expansion and expense discipline, both of which are finite. Payment Services, the segment that should be the growth engine, contributed $231 million of Q1 net income — DOWN $1 million year on year."

Exponential Potential2/10Low

"Rated 2 — there is no exponential here and the dive will not manufacture one. U.S. Bancorp's four reportable segments are Wealth, Corporate, Commercial and Institutional Banking; Consumer and Business Banking; Payment Services; and Treasury and Corporate Support. The only one with a plausible claim to non-linear economics is Payment Services — merchant processing, corporate and purchasing cards, stored-value — where scale should produce operating leverage over a fixed technology base. It is not producing it: Q1 2026 net revenue rose 3.9% and segment net income FELL $1 million to $231 million as noninterest expense rose 3.5%. Card revenue generated from debit cards was moved OUT of Payment Services into Consumer and Business Banking effective 2026, which shrinks the segment's disclosed footprint further. The pending BTIG acquisition (up to $1 billion, institutional trading and investment banking) is a capability purchase in a mature business, not a platform. The genuine strengths of this company — a 75% deposit funding ratio, 2.00% allowance coverage, 68,520 employees running a national franchise — are all the strengths of a well-managed utility. A 2: a linear, cyclical, well-capitalised spread-and-fee business."

Fair value$67 $52–$78

What does “fair value” mean?

Fair value is Synthos’s estimate of what one share is worth today, based on our model of the company’s future cash generation and the risks to it. It is not a price target or a prediction of where the stock trades next quarter — it’s the price where we think risk and reward are balanced. Above it, you’re paying for outcomes better than our base case; below it, the market is offering a margin of safety.

The model’s inputs and weightings are proprietary — they’re the product. What we publish is the output, plus an independent, public-math cross-check further down the page (“Check our number”) so you can judge it for yourself.

What do the 5 tiers mean? (Core · Tactical · Watch · Hold · Avoid)

Buy — CoreOwn it as a foundation — start or add now, size it for years, let dips be gifts.

Buy — TacticalGood price + confirmed trend + a defined exit — buy the setup, not a marriage.

WatchWe want the business, just not at this price/setup — act only when the listed trigger hits.

HoldFine to keep if you own it — no reason to buy more; new money does better elsewhere.

AvoidDon't own it — the problem is the business or the expectations, so a cheaper price won't fix it.

The Road Ahead

What we expect to matter in each window, and the evidence that would prove us wrong.

Short term 0-6 months

Neutral

Driver: "Strong momentum, no entry discount. U.S. Bancorp closed 2026-08-04 at $64.23, up 0.45%, which is 0.4% BELOW the 52-week high of $64.47, 45.7% above the low of $44.09, 7.1% above a rising 50-day average of $59.99 and 17.1% above a 200-day average of $54.84. RSI is 56.5 — unstretched — and MACD is +0.97. Twelve-month return is +46.2% against SPY's +24.3% and QQQ's +30.8%; three-month +16.3% against SPY +7.6%. The fundamental news flow supporting it is real: a 5.5% EPS beat on 2026-07-16, a stress capital buffer held at 2.6% on 2026-06-24 and a dividend increase to $0.54 a quarter from Q3. What makes the stance neutral rather than a tailwind is that all of it is now in the price. The consensus target of $69.33 is 7.9% above spot, the median $68 is 5.9% above, and the LOW target of $61 is 5.0% BELOW — the published street distribution is roughly symmetric around a small gain. Buying a bank at its 52-week high after a 46% year, at 11.0x the second forward year, is not a setup with a margin of safety."
What we’re watching: "The 2026-10-15 print against consensus diluted EPS of $1.32 and net revenue of $7,964M. Within it: the net interest margin, 2.77% in Q1 2026 against 2.72% a year earlier, which is where roughly half the earnings growth comes from; total loan net charge-offs, 0.56% of average loans in Q1 against 0.59%, and specifically whether commercial charge-offs keep rising while credit-card losses fall; the allowance-to-loans ratio, 2.00% at 2026-03-31 against 2.03% at 2025-12-31, since a falling ratio means reserve release is flattering earnings; and the CET1 ratio, 10.8%, against the newly confirmed 7.1% requirement. Separately, whether the BTIG acquisition — up to $1 billion, expected to close in Q2 2026 — is confirmed closed and what it costs in expense and share count (6,600,594 shares were to be issued at closing)."
Confidence: Medium

Medium term 6-24 months

Neutral

Driver: "The medium-term case is a 10-11% earnings compounder paying a 3.4% forward dividend at 11x — respectable, and almost exactly what the price says. Consensus has diluted EPS at $5.215 in FY2026, $5.816 in FY2027 and $6.397 in FY2028, on net revenue rising from $30,972M to $34,839M. The mechanism is not mysterious: net interest margin expansion from fixed-asset repricing (2.77% and rising), loan growth of roughly 4%, noninterest expense growing under 1%, and a modest buyback under a $5.0 billion authorisation dated 2024-09-13. The dividend rises to $0.54 a quarter from Q3 2026, a $2.16 annualised run-rate and a 3.36% forward yield. Capital return was $1.1 billion in Q1 2026 across dividends and repurchases, of which the dividend is roughly $0.8 billion — so the buyback is running near $1.2 billion a year, about 1.2% of the $100.1 billion market capitalisation. Total shareholder yield is therefore roughly 4.6%, well short of what would rescue a thin price return. The honest arithmetic: 10-11% EPS growth on a multiple that is already at the top of this bank's decade range gets you a high-single-digit total return if nothing goes wrong, and considerably less if the multiple normalises."
What we’re watching: "Whether the net interest margin keeps expanding once fixed-asset repricing is exhausted — that is the single biggest swing factor in the FY2027 consensus and it is a rate-regime bet, not a management one. Whether Payment Services stops shrinking in profit: it earned $231 million in Q1 2026, one million dollars less than a year earlier, and it is the only segment with a claim to structural growth. Whether the allowance ratio stabilises at 2.00% or keeps falling, because release is not earnings. Whether BTIG closes, what it costs, and whether an institutional trading and investment-banking business proves accretive inside a commercial bank. Whether commercial-loan charge-offs, explicitly called out as rising in Q1, become a trend. And whether the $5.0 billion repurchase authorisation is actually used at scale now that the stress capital buffer is settled at 2.6% through October 2027 — a genuine buyback would change the return arithmetic materially and is the most plausible route to an upgrade."
Confidence: Medium

Long term 2+ years

Neutral

Driver: "Long-run this is a well-run, scale-advantaged, deposit-funded national bank with a payments business attached, and its long-run return is close to its return on equity less what it pays away for growth. Return on equity is 12.5% on the vendor's trailing measure; return on assets is 1.13%. Those are respectable regional-bank numbers and they have been stable for a decade. The structural question is whether the Payment Services franchise — merchant processing, corporate and purchasing cards — can be defended against payment-native competitors whose cost structures are software rather than branch networks, and the recent evidence is that it is being defended rather than extended: revenue up 3.9%, profit down. The offsetting structural strength is the deposit base. Total deposits of $528.2 billion funding roughly 75% of assets, gathered through 1,056 owned and 1,409 leased facilities, is a genuinely hard thing to replicate and it is why this bank survives rate cycles that damage wholesale-funded lenders. Gunjan Kedia became chief executive in 2025; the filings in this archive contain no succession or strategic-redirection disclosure beyond the BTIG agreement."
What we’re watching: "Whether Payment Services can be made to grow profit again, or whether it is structurally ceding merchant economics — that is the difference between a bank that deserves a premium multiple and one that does not. Whether the deposit base holds its 75% funding share and its cost advantage through the next easing cycle. Whether the loan book's commercial concentration, currently the source of rising charge-offs, behaves through a genuine credit event: the last full test was 2020 and the balance sheet is materially larger. Whether the company converts the settled 2.6% stress capital buffer into a durable, large share repurchase rather than a token one — with 1,554 million shares outstanding and a $5.0 billion authorisation, the difference between using it over two years and over eight is roughly three points of annual shareholder return. And whether U.S. Bancorp remains a consolidator: BTIG is small, but it is the first acquisition since Union Bank and the pattern matters more than the price."
Confidence: Low

Exponential Potential

Exponential Potential2/10Low

What could take this further than the base case — and where the market may be underpricing it. Full forward-growth and acceleration math in Deeper analysis, §4 below.

Deeper analysis

Technicals, fundamentals, valuation, and the full expert-claim evidence panel — the detail behind the numbers above.

Check our number Market-implied growth ≈ 5%/yr This isn’t how we calculate fair value — it’s public math you can verify yourself. To justify today’s $62, earnings would have to compound roughly 5% a year for 10 years (9% discount rate). Analysts forecast ~3%/yr, so the market is pricing in MORE than what the Street expects.

Reference table

Street consensus	$69.33 (+7.9%) · median $68 · high $75 (+16.8%) · low $61 (−5.0%) · 1 strong buy / 24 buy / 22 hold / 2 sell across 49 analysts
Valuation	12.8x trailing diluted EPS ($5.01) · 12.3x FY2026E $5.215 · 11.0x FY2027E $5.816 · 10.0x FY2028E $6.397 · ~1.63x common book and ~2.33x tangible common book after deducting preferred (see Section 8)
Bank metrics — filing-verified, 2026-03-31	CET1 10.8% against a 7.1% requirement (4.5% + 2.6% SCB) · total deposits $528.2B, funding ~75% of assets · loans $399.8B · net interest margin 2.77% (taxable-equivalent) vs 2.72% · net charge-offs 0.56% of average loans vs 0.59% · allowance 2.00% of loans, 536% of nonperforming loans
Capital return	$1.1B returned in Q1 2026 across dividends and repurchases · quarterly dividend rising $0.52 → $0.54 from Q3 2026 ($2.16 annualised, 3.36% forward yield) · $5.0B repurchase authorisation dated 2024-09-13, being used slowly ($489M in all of FY2025) · total shareholder yield ≈ 4.6%
Conviction	None. 3 raw KB hits, 0 used, 3 discarded — all three are the Universal Serial Bus. The lane is empty and is reported as empty
Technicals	−0.4% from the 52-week high of $64.47; +45.7% above the low of $44.09; +7.1% above the 50-DMA ($59.99), +17.1% above the 200-DMA ($54.84); RSI 56.5; MACD +0.97; 12-month +46.2% vs SPY +24.3%

What the experts actually said

No independent expert claims in the Synthos knowledge base yet for USB — this dive is fundamentals- and technicals-driven, not panel-driven.

Price & moving averages 12 months · 50 & 200-day averages · 52-week range

Solid line = price · dashed line = 50-day average · dotted line = 200-day average · the two thin horizontal lines mark the 52-week high and low. Price above both averages is an uptrend.

Bollinger Bands 20-day average ± 2 standard deviations

The shaded band widens when the stock gets more volatile. Riding the upper edge = strong momentum (sometimes stretched); the lower edge = weak / potentially oversold.

RSI (14) momentum gauge · 0–100

Above 70 (overbought zone, shaded) = overbought, below 30 (oversold zone, shaded) = oversold. Currently 45.

MACD 12 / 26 / 9 · trend & momentum

The MACD line crossing above the signal line (bars flip to the up color) = momentum turning up; crossing below (bars flip to the down color) = turning down. Bar height = the size of that gap.

Relative performance vs S&P 500 & its sector (XLF (sector)), set to 100 a year ago

Solid = USB · dashed = S&P 500 · dotted = XLF (sector). A rising line means it is beating that benchmark — the sector line shows whether it is a leader or laggard within its own group.

Forward revenue & earnings actual → estimate · "FY" = fiscal year, "E" = estimate

Darker bars = actual results, brighter = analyst estimates. Taller bars to the right = expected growth.

Key stats an RIA wants

Price$62.47

Market cap$97B

P/E trailing12×

P/E FY26E / FY27E12× / 11×

EV / Sales0.9×*

EV / EBITDA3.5×*

Gross margin63.8%

Net margin18.7%

Dividend yield3.33%

Beta0.977

52-wk range$45 – $65

RSI(14)37

50 / 200-DMA$63 / $56

12-mo return+28% (SPY +19%)

Street target$70 ($66–$75)

Analyst grades24 Buy · 22 Hold · 2 Sell

FMP ratingB

Next earnings2026-10-15 (Q3 2026 earnings, 72 days away; vendor consensus diluted EPS $1.32 and net revenue $7,964M, implying +3.6% on the June quarter's $7,686M rebuilt figure). Second-quarter results were released 2026-07-16, nineteen days before this dive, and beat consensus EPS by 5.5% ($1.35 against $1.28) on revenue of $7,712M against $7,577M estimated.

* Enterprise value recomputed in-house: the data vendor nets cash but omits short-term investments, overstating EV for cash-rich balance sheets. EV multiples marked * use market cap + total debt − cash − short-term investments.

1. What the business is

U.S. Bancorp is a Delaware-incorporated financial holding company headquartered at 800 Nicollet Mall, Minneapolis, with approximately 68,520 employees, 1,056 owned and 1,409 leased facilities, and Gunjan Kedia as chief executive. It is the fifth-largest US bank by assets.

Four reportable segments, per the March 2026 10-Q — and note that these are NOT the names in the vendor payload:

Segment	What it is	Q1 2026 net income	YoY	Q1 2026 net revenue change
Wealth, Corporate, Commercial and Institutional Banking	Core banking, specialised lending, capital markets, asset management, brokerage, tax-advantaged investments	$1.4B	+$229M (+19.0%)	+$351M (+11.2%)
Consumer and Business Banking	Consumer and small-business banking, debit cards, consumer lending	$616M	+$19M (+3.2%)	+$27M (+1.2%)
Payment Services	Consumer and business credit cards, stored-value, corporate/government/purchasing cards, merchant processing	$231M	−$1M (−0.4%)	+$65M (+3.9%)
Treasury and Corporate Support	Investment portfolio, funding, capital management, unallocated items	−$336M	−$11M	−$113M

Two reorganisations landed in 2026 and both matter for reading any historical series. The Impact Finance business unit moved from Treasury and Corporate Support into Wealth, Corporate, Commercial and Institutional Banking; and debit-card revenue moved out of Payment Services into Consumer and Business Banking. Prior periods were recast. Separately, effective 2026-01-01 the company renamed and reclassified several fee lines — "Corporate payment products revenue" became "Corporate payment and treasury management revenue", stored-value card revenue moved out of Card revenue, treasury management services moved out of Lending and deposit-related fees, and Impact Finance syndication fees moved from Other into Capital markets revenue. Any fee-line comparison against a pre-2026 source is comparing different definitions.

The one segment that should be the growth engine is the one that is not growing. Payment Services net revenue rose 3.9% while noninterest expense rose 3.5% and provision rose 9.5%, so segment net income fell by a million dollars. The debit-card reclassification also shrinks the segment's disclosed scope going forward.

The pending BTIG acquisition — announced January 2026, absent from the payload

From the March 2026 10-Q:

> "In January 2026, the Company announced that it entered into a definitive agreement to acquire BTIG for a purchase price of up to $1 billion, consisting of a targeted amount of $725 million ($362.5 million of cash and 6,600,594 shares of the Company's common stock) to be paid at closing and up to an additional $275 million of cash consideration payable over three years, subject to achievement of defined performance targets. BTIG is a global financial services firm specializing in institutional trading, investment banking, research and related brokerage services. The transaction is expected to close in the second quarter of 2026, subject to regulatory approvals and satisfaction of applicable closing conditions."

Nothing in the vendor payload reflects this. bal_a ends 2025-12-31 and cf_a ends FY2025. And nothing in this filing archive confirms the deal actually closed — the June-quarter 8-K carries only the cover page, the earnings release itself being an unextracted exhibit, and there is no Q2 2026 10-Q in the archive. We therefore report BTIG as announced and expected-to-close, not as closed, and say so rather than assuming. At up to $1 billion it is roughly 1% of market capitalisation; the 6,600,594 shares are 0.42% of the count. It is not thesis-moving, but it is a real corporate action that postdates every statement in the payload.

2. The vendor's revenue line is a gross-up the company does not publish

U.S. Bancorp's published top line is total net revenue — net interest income plus noninterest income. The vendor reports gross interest income plus noninterest income, before deducting interest expense.

The reconciliation is exact and runs through the vendor's own interest-expense line:

Basis	FY2025	FY2024	FY2023
Vendor `revenue` (gross-up)	$42,861M	$42,712M	$40,624M
Less vendor `interestExpense`	($14,321M)	($15,377M)	($12,611M)
= Total net revenue (published basis)	$28,540M	$27,335M	$28,013M
Cross-check: `est.revenueAvg` for the same year	$28,598M	$27,462M	$28,216M
Difference	0.2%	0.5%	0.7%

The analyst estimates in the same file are struck on the net basis, and they agree with the rebuild to within 0.7% — the residual being the taxable-equivalent adjustment. So the vendor's revenue line is 50.2% above the basis every analyst in the file is using, and the two are sitting in the same JSON payload.

This inverts the growth story. On the vendor's line, revenue grew 5.1% then 0.3%. On the published line it FELL 2.4% and then grew 4.4%.

The same gross-up runs through inc_q, and there the earnings calendar provides an independent check:

Quarter	Vendor `inc_q` revenue	less interest expense	= rebuilt net revenue	`earn_cal` revenueActual	agreement
Q2 2025	$10,528M	$3,553M	$6,975M	—	—
Q3 2025	$11,005M	$3,705M	$7,300M	$7,300M	exact
Q4 2025	$10,976M	$3,639M	$7,337M	$7,362M	0.3%
Q1 2026	$10,835M	$3,575M	$7,260M	$7,288M	0.4%
Q2 2026	$10,923M	$3,237M	$7,686M	$7,712M	0.3%

Two independent blocks of the same payload disagree by 50% on what this company's revenue is, and the smaller one is right. Every revenue and margin figure in this dive uses the published basis. Consequential rejections: priceToSalesRatioTTM of 2.288x (correct on published revenue: approximately 3.44x), revenuePerShareTTM of $28.15, netProfitMarginTTM of 18.69% (correct: approximately 28.0%), and every margin ratio in ratios_ttm.

The June 2026 quarter's rebuilt net revenue of $7,686M is +10.2% year on year and the highest in the nine-quarter series. That is the single best operating fact in this file, and it is invisible on the vendor's line, which shows +3.8%.

3. The earnings trajectory

Quarter	Rebuilt net revenue	YoY	Net income	Diluted EPS	vs estimate
Q2 2024	$6,838M	—	$1,603M	—	—
Q3 2024	$6,833M	—	$1,714M	—	—
Q4 2024	$6,979M	—	$1,663M	—	—
Q1 2025	$6,928M	—	$1,709M	$1.03	—
Q2 2025	$6,975M	+2.0%	$1,815M	—	—
Q3 2025	$7,300M	+6.8%	$2,001M	$1.22	vs $1.13 (+8.0%)
Q4 2025	$7,337M	+5.1%	$2,051M	$1.26	vs $1.19 (+5.9%)
Q1 2026	$7,260M	+4.8%	$1,945M	$1.18	vs $1.14 (+3.5%)
Q2 2026	$7,686M	+10.2%	$2,177M	$1.35	vs $1.28 (+5.5%)

Quarterly EPS from earn_cal actuals; Q1 2026's $1.18 is filing-confirmed in the 10-Q. Net revenue rebuilt as inc_q revenue less inc_q interestExpense, cross-checked against earn_cal revenueActual above.

Four consecutive beats, none of them large. +8.0%, +5.9%, +3.5%, +5.5%. That is a company the street models well, which is the opposite of the mid-integration chaos seen elsewhere in this programme — and it means there is little scope for a surprise to re-rate the shares.

The March quarter, which is the one the filings actually document:

	Q1 2026	Q1 2025	change
Net interest income	~$4.3B	—	+$171M (+4.2%)
Noninterest income	~$3.0B	—	+$161M (+5.7%)
Noninterest expense	~$4.3B	—	+$33M (+0.8%)
Provision for credit losses	$576M	$537M	+$39M (+7.3%)
Net income attributable to U.S. Bancorp	~$1.9B	~$1.7B	—
Diluted EPS	$1.18	$1.03	+14.6%
Net interest margin (taxable-equivalent)	2.77%	2.72%	+5 bps
Average earning assets	—	—	+$13.9B (+2.3%)
Average total loans	—	—	+$14.5B (+3.8%)
Effective tax rate	19.4%	20.5%	−110 bps

Read the operating leverage. Revenue up roughly 5%, expenses up 0.8%. That gap is where the 14.6% EPS growth comes from, and it is the most impressive thing in the filing. It is also the thing that cannot repeat indefinitely: expense growth of 0.8% on a $17 billion cost base is a one- or two-year achievement, not a permanent state.

Note what the vendor's inc_a does to the FY2021-2022 comparison. Reported EPS of $5.11 in FY2021 and $3.69 in FY2022 reflects the Union Bank acquisition and the rate cycle, not a deteriorating franchise. The FY2021 figure is also why the trailing multiple looks unremarkable against history — this bank earned $5.11 five years ago and consensus wants $5.22 this year. Five years of essentially no EPS growth, then a 46% stock move in twelve months, is the shape of a re-rating rather than a compounding.

4. Balance sheet, credit and capital

All figures from the March 2026 10-Q, filing-verified:

	2026-03-31	2025-12-31
Total loans	$399.8B	$391.3B (+2.2%)
Commercial loans	—	+$5.9B (+4.0%)
Total deposits	$528.2B	$522.2B
Deposit funding share of total assets (average)	~75%	~76% (Q1 2025)
Short-term borrowings	$17.9B	$17.2B
Long-term debt	$61.4B	$60.8B
CET1 ratio	10.8%	10.8%
Allowance for credit losses / loans	2.00%	2.03%
Allowance / nonperforming loans	536%	514%
Allowance / annualised net charge-offs	360%	367%
Net charge-offs (quarter)	$546M	$547M (Q1 2025)
Net charge-off rate	0.56%	0.59% (Q1 2025)
Residential mortgages 90+ days past due or nonperforming, incl. GNMA repurchase-option loans	4.05%	3.37%

Capital is settled and the requirement is now known. From the 2026-06-24 8-K: "the Company's Stress Capital Buffer ('SCB') requirement will remain unchanged at 2.6% until October 1, 2027." Against a 4.5% minimum that is a 7.1% CET1 requirement, and the company runs 10.8% — 370 basis points of headroom. The same filing announced the dividend increase from $0.52 to $0.54 per quarter from Q3 2026, subject to board approval.

Credit is good and one series is not. Total net charge-offs of 0.56% are low by any historical standard and improving. But the composition is moving: the improvement came from credit-card losses falling while commercial loan charge-offs rose — and commercial is where this bank grew loans 4.0% in a single quarter. Meanwhile the residential mortgage delinquency ratio including GNMA repurchase-option loans went from 3.37% to 4.05% in three months. The company excludes those loans from headline delinquency statistics because their repayment is insured by the Federal Housing Administration or guaranteed by the Department of Veterans Affairs, which is a legitimate exclusion — but the ratio is disclosed and it moved 68 basis points in a quarter, and we report it.

Reserve coverage is being released, gently. The allowance fell from 2.03% to 2.00% of loans while loans grew 2.2%. A falling coverage ratio contributes to reported earnings; it is not operating performance, and the 2026-10-15 print should be read with that in mind.

Cash flow is not a useful lens for a bank and is not used as one here. The vendor reports FY2025 operating cash flow of $7,970M, capitalExpenditure of exactly zero and therefore free cash flow identical to operating cash flow, a freeCashFlowYieldTTM of 7.56%. A bank with 2,465 facilities does not spend zero on capital expenditure; the line is simply not populated. All free-cash-flow, income-quality and price-to-free-cash-flow measures in this file are rejected. The measures used are CET1, the net charge-off rate, allowance coverage, net interest margin and return on equity.

5. Segment and geographic disclosure — the payload does not reconcile

seg_prod fails the sum test and uses names the company no longer reports. The FY2025 entry contains three lines:

`seg_prod` FY2025 line	Amount
Wealth Management And Investment Services	$12,083M
Payment Services	$9,700M
Consumer And Small Business Banking	$8,873M
Sum	$30,656M
Total net revenue (published basis, Section 2)	$28,540M
Sum as % of revenue	107.4%

Three failures in one block. The sum is 7.4% above the company's total net revenue. Treasury and Corporate Support is missing entirely — it appears in the FY2024 and FY2023 entries (at −$1,031M and +$1,121M) and then vanishes. And the segment names are stale: the company's reportable segments are Wealth, Corporate, Commercial and Institutional Banking; Consumer and Business Banking; Payment Services; and Treasury and Corporate Support. "Wealth Management And Investment Services" at $12,083M is evidently the WCCIB segment under an obsolete label — the actual wealth-management fee business is a fraction of that.

It gets worse further back. The FY2012 entry reads "Consolidated Company $20,288M, Segment Five $2,451M, Segment Four $4,740M, Segment One $3,323M, Segment Three $1,476M, Segment Two $8,298M" — anonymised placeholders of the USFD "Product One–Seven" class, mixed with a consolidated total that would double-count if summed.

seg_geo contains exactly one entry, for FY2012, and it is the anonymised placeholder set: Segment Four $4,740M, Segment Three $1,476M, Segment Two $8,298M. There is no geographic revenue disclosure for this company in the payload at all, for any year in the last decade. The only geographic datum available is from the 10-Q: "Revenue generated from sources in Europe represented approximately 2 percent of the Company's total net revenue for the three months ended March 31, 2026," with approximately $6.2 billion on deposit with European banks, predominantly the Central Bank of Ireland and Bank of England. All segment figures in this dive come from the 10-Q, in percentage-change form, because the 10-Q's tables were not preserved either — see Section 8.

6. Valuation — priced in or room?

At $64.23 (market capitalisation $100.05B, implied 1,557.7M shares):

	Trailing	FY2026E	FY2027E	FY2028E
Diluted EPS	$5.01	$5.215 (16)	$5.816 (16)	$6.397 (8)
EPS growth	—	—	+11.5%	+10.0%
P/E	12.8x	12.3x	11.0x	10.0x
Net revenue (published basis)	$28,540M (FY2025)	$30,972M (8)	$33,062M (10)	$34,839M (7)
Revenue growth	—	+8.5%	+6.7%	+5.4%
Price / common book	~1.63x	—	—	—
Price / tangible common book	~2.33x	—	—	—
Return on equity (vendor TTM)	12.5%	—	—	—

Trailing diluted EPS of $5.01 is the sum of the four earn_cal actuals ($1.22 + $1.26 + $1.18 + $1.35) and is confirmed by the vendor's own priceToEarningsRatioTTM of 12.82x. It is NOT the vendor's netIncomePerShareTTM of $5.26, which does not deduct preferred dividends — see Section 8.

Estimate coverage is good on the near years and thin beyond. Sixteen analysts on both FY2026 and FY2027 EPS; eight on FY2028; three on FY2029, which is excluded from every conclusion. The FY2027 EPS range is $5.442 to $5.968, a 9.7% spread — narrow, which is what you expect for a bank the street models to within 5%.

est.ebitdaAvg and est.ebitAvg are rejected outright. In FY2026, FY2027 and FY2028, ebitdaAvg is exactly 27.556% of revenueAvg and ebitAvg is exactly 25.427% — the fixed-ratio fabrication signature, present here to five decimal places across three years. The rows are also conceptually void for a bank whose principal cost is interest. All forward valuation uses epsAvg, cross-checked against earn_cal actuals.

Enterprise value is rejected as a concept. The vendor prints enterpriseValueTTM of $129.57B and netDebtToEBITDATTM of 2.72x for a company funded by $528.2 billion of customer deposits. Deposits are the raw material of this business, not leverage. evToSalesTTM (2.96x), evToEBITDATTM (11.93x), interestCoverageRatioTTM (0.74x) and netCurrentAssetValueTTM (−$591,532M) are all discarded.

Peer context. The vendor set is Barclays, CIBC, Deutsche Bank, HDFC, ING, Itaú, Lloyds, Mizuho, PNC and Truist — eight of ten are non-US banks operating under different capital regimes, accounting standards and rate environments. Only PNC ($101.4B market capitalisation, near-identical size) and Truist ($65.6B) are genuine comparables. No peer multiple comparison is drawn from a set that is 80% inapplicable, beyond noting that PNC at $254.14 and a nearly identical market capitalisation is the single most useful line in the block.

6a. What today's price assumes (the inversion)

At $64.23 — 12.3x FY2026 consensus EPS, 11.0x FY2027, roughly 1.63x common book — the price embeds:

Diluted EPS reaches $5.215 in 2026 and $5.816 in 2027. (Consensus; 16 analysts each.) The first half of 2026 delivered $2.53 ($1.18 + $1.35), so the FY2026 figure needs $2.69 in the second half — a 6.3% step, which the September guide of $1.32 is consistent with. FY2027 then requires 11.5% growth on top.
The net interest margin keeps expanding past 2.77%. (Our derivation; the 10-Q supplies the rate and the driver.) This is the most fragile assumption in the price. The Q1 expansion was attributed specifically to "fixed asset repricing" — the roll-off of low-yielding legacy assets into current rates. That is a finite, exhaustible source, and roughly half the consensus earnings growth rests on it continuing.
Noninterest expense growth stays near 1%. (Our number, from the 10-Q's +0.8%.) On a cost base near $17 billion, expenses growing at 1% while revenue grows 5-8% is the entire operating-leverage story. Two points of extra expense growth is roughly $340 million pre-tax, about 4% of consensus net income.
The market keeps paying roughly 11-12x forward earnings and 1.6x book. (Our number.) At 9.5x FY2027E the stock is $55; at 13x it is $76. This bank has spent most of the last decade between 9x and 13x, and it is now at the upper end of that band.
Credit stays benign at a 0.56% charge-off rate. (Filing-verified rate; our extrapolation.) The allowance is 2.00% of loans and falling. A move to a 1.0% charge-off rate — still historically unremarkable — is roughly $1.8 billion of additional annual provision, about a fifth of pre-tax income.

6b. The return bridge (why the multiple moves)

Expected return over the next twelve months decomposes as: EPS growth (+11.5%, from FY2026E $5.215 to FY2027E $5.816) + multiple drift (COMPRESSION, roughly 12.3x to about 11.5x on the then-forward year, −6 to −7%) + shareholder yield (+4.6%) ≈ +9%.

Our base assumes mild multiple compression, and the reason is the level rather than the business. U.S. Bancorp has re-rated from roughly 9x to roughly 12x forward earnings over twelve months without a corresponding change in growth rate, return on equity or credit. A 46% twelve-month move on 11% earnings growth is, arithmetically, about three-quarters multiple and one-quarter earnings. We are not willing to underwrite that continuing, and we are equally unwilling to assume it reverses violently — a bank earning 12.5% on equity with settled capital requirements and improving credit does not obviously deserve 9x either.

What this makes the return: almost entirely earnings growth and dividend, with the multiple working mildly against it. That is a defensible 8-10% total return and it is why this is a Hold rather than an Avoid. It is not a 15%+ return and there is no honest arithmetic that makes it one.

If the multiple compressed to 9.5x FY2027E the price is $55 (−14.0%). If it expanded to 13x, $76 (+18.3%). The bull case at $78 is 13.1x the FY2027 consensus high of $5.968.

6c. Variant perception (where we differ, what would surprise)

We are BELOW the street and the gap is small and honest. Our $67 against consensus $69.33 is a 3.4% difference — inside the noise of a single multiple point. We have no meaningful variant perception on this name and say so plainly. Where we differ is in what we do with the absence of one: a name where our number and the street's agree, at a full multiple, at a 52-week high, is a Hold by construction. A no-edge name defaults toward Hold, and this is the cleanest example of that in the batch.
We differ from the vendor data on three points, all documented in Section 8, and one of them changes the growth narrative. On the vendor's revenue line U.S. Bancorp grew 5.1% and then 0.3%; on the basis the company publishes and the analysts in the same file use, it FELL 2.4% and then grew 4.4%. Anyone screening this name on revenue or any margin ratio is looking at a company that does not exist.
We think the June quarter is better than the headline and the market has already noticed. Rebuilt net revenue of $7,686M is +10.2% year on year — the strongest quarter in the nine-quarter series and roughly double the Q1 rate. Watchable number: whether the 2026-10-15 print sustains net revenue above $7.9 billion, which is what consensus implies and what a continuing margin expansion would produce. The stock's 16.3% three-month move suggests this is understood.
Positive surprise that would force a re-rate: aggressive use of the $5.0 billion repurchase authorisation now that the stress capital buffer is settled at 2.6% through October 2027. FY2025 saw only $489 million of repurchase against a $100 billion market capitalisation. Running the full authorisation over two years would add roughly 2.5 percentage points a year to shareholder return and would be the single most plausible route from Hold to Buy. A second surprise: Payment Services returning to profit growth, which would justify the fee-business premium the current multiple implies.
Negative surprise that would break the thesis: net interest margin flat or down at the October print, which removes half the consensus growth; commercial-loan charge-offs continuing to rise on a book that just grew 4.0% in a quarter; an allowance ratio below 1.95%, confirming that release rather than performance is carrying earnings; or a BTIG closing on worse terms than the January agreement. Any of these hits a stock 0.4% below its 52-week high with a street low target 5% below spot.

Synthos fair values

All three anchors are multiples of the FY2027 consensus diluted EPS distribution (mean $5.816, low $5.442, high $5.968, 16 analysts), each cross-checked against common book value.

Bear ~$52 — 9.5x the FY2027 consensus LOW of $5.442, approximately 1.32x current common book. Cross-check: 10.0x FY2026E; 18.0% above the 52-week low of $44.09. The scenario: net interest margin expansion stops, commercial credit deteriorates toward a 1.0% charge-off rate, the allowance has to be rebuilt, and the multiple returns to the bottom of the decade band. −19.0%.
Base ~$67 — 11.5x the FY2027 consensus MEAN of $5.816, approximately 1.70x common book. Cross-check: 12.8x FY2026E; 10.5x FY2028E. Sensitivity, stated openly: 11.0x gives $64 — spot — and 12.0x gives $70. The whole answer lives inside one multiple point. The scenario: margin expansion continues modestly, expenses stay near flat, credit stays benign, the dividend rises as announced, and the multiple gives back a little of its twelve-month gain. +4.3%.
Bull ~$78 — 13.1x the FY2027 consensus HIGH of $5.968, approximately 1.98x common book. Cross-check: 12.2x FY2028E; 21.0% above the 52-week high. The scenario: the buyback goes to scale under the settled 2.6% buffer, Payment Services re-accelerates, credit stays where it is, and the market pays a fee-bank multiple. +21.4%.

Base is 4.3% above spot; asymmetry roughly 1.13:1 (19.0% down, 21.4% up), before a shareholder yield of approximately 4.6%. A base case inside 5% and a payoff ratio near one-to-one is the definition of fairly valued. This is a Hold and the arithmetic is the whole argument.

7. Knowledge base — three hits, all of them a computer cable

Raw hits: 3. Entity matches: 0. Used: 0. Discarded: 3, all homograph noise.

The sweep ran case-sensitive entity tokens USB, U.S. Bancorp, US Bancorp and U.S. Bank, plus case-sensitive free-text patterns for the same, across all 51,928 distilled claims. It returned zero entity matches and three free-text hits. All three are the Universal Serial Bus.

> 2024-04-24 · neutral · conviction 80 · horizon: principle · entities: [none] · categories: sleep science, memory consolidation, neuroscience · channel: huberman_lab

> "Hippocampus acts as a limited 'USB stick'; sleep transfers memories to cortical long-term storage, clearing capacity to acquire new information."

> 2025-09-27 · bullish · conviction 85 · horizon: principle · entities: Bitcoin · channel: lyn_alden

> "Like Ethernet, TCP/IP, USB, Bitcoin is a dominant communication protocol (of value); it gets displaced only by its own updates, not competitors."

> 2026-04-26 · bullish · conviction 72 · horizon: thesis · entities: Bitcoin · speaker: Lyn Alden · channel: lyn_alden

> "Bitcoin is like an entrenched communication protocol (USB, TCP/IP); after 17 years his network effect makes it nearly impossible to displace — you bet on its properties."

All three are discarded and the sweep is discarded entirely. This is the homograph class the programme has repeatedly caught — "PH" as petahash, "Vertex" as a cloud platform, "ETN" as exchange-traded notes — and here it is total: the three-letter ticker of the fifth-largest bank in the United States is also the name of the most widely deployed hardware interface standard in history, and the knowledge base contains only the latter. Note that a case-INSENSITIVE or substring sweep would have been far worse, catching every instance of the string inside other words; the case-sensitive word-boundary discipline is what reduced this to three.

The correct report is that the lane is empty. Breadth 0, claim count 0, net conviction none. No expert in the Synthos knowledge base has expressed a recorded view on U.S. Bancorp, and this dive is therefore built entirely on filings and the vendor payload. That is a limitation and it is stated rather than papered over with adjacent macro claims about banks or consumer credit, which would not name this company.

Attribution note: two of the three discards carry speaker: null and are channel-attributed; the third names Lyn Alden. No management voice appears anywhere in this lane, and no concentration sensitivity test is meaningful across an empty pool.

8. Data integrity — what we rejected and why

Nine findings. The most consequential is the revenue basis, which is wrong by 50% and inverts the growth narrative. The most unusual is that this company's filing archive contains no preserved statement tables at all and no financial statements in the 10-K.

1. revenue is a gross-up 50.2% above the published basis — REJECTED and rebuilt. Detailed in Section 2. FY2025 vendor revenue $42,861M against a rebuilt total net revenue of $28,540M; the reconciliation runs exactly through the vendor's own interestExpense of $14,321M, and the rebuild agrees with the payload's own est.revenueAvg for the same year ($28,598M) to within 0.2%. Two blocks of the same file disagree by 50% and the smaller is right. The same defect runs through all nine quarters of inc_q, verified against earn_cal revenueActual to within 0.4%. Consequential rejections: priceToSalesRatioTTM 2.288x (correct ≈ 3.44x), revenuePerShareTTM $28.15, netProfitMarginTTM 18.69% (correct ≈ 28.0%), grossProfitMarginTTM 63.8%, ebitMarginTTM 23.3%, ebitdaMarginTTM 24.8%, assetTurnoverTTM, fixedAssetTurnoverTTM.

2. seg_prod sums to 107.4% of net revenue, omits a segment, and uses obsolete names — REJECTED. Detailed in Section 5. FY2025's three lines total $30,656M against $28,540M of total net revenue. Treasury and Corporate Support is present in FY2024 and FY2023 and absent in FY2025. The names ("Wealth Management And Investment Services", "Consumer And Small Business Banking") are not the company's current reportable segments. The FY2012 entry additionally mixes a "Consolidated Company" total with anonymised "Segment One" through "Segment Five" placeholders. No segment dollar figure in this dive comes from the payload.

3. seg_geo contains one entry, dated FY2012, and it is anonymised placeholders. Segment Four $4,740M, Segment Three $1,476M, Segment Two $8,298M. There is no usable geographic disclosure for this company in the payload for any year. The 10-Q supplies the only geographic datum used: Europe ≈ 2% of Q1 2026 total net revenue.

4. est.ebitdaAvg and est.ebitAvg carry the fixed-ratio fabrication signature — REJECTED. ebitdaAvg is exactly 27.556% of revenueAvg and ebitAvg exactly 25.427% in FY2026, FY2027 and FY2028 alike. The historical rows are worse: FY2023 shows ebitdaAvg of $11,658M on revenueAvg of $28,216M (41.3%) against FY2024's 27.6% — a 14-point step with no corporate event. EBITDA is in any case not a concept that applies to a bank whose principal cost is interest. All forward valuation uses epsAvg.

5. netIncomePerShareTTM does not deduct preferred dividends — a 5% overstatement, and the payload contradicts itself. The vendor reports netIncomePerShareTTM of $5.260. The four earn_cal diluted EPS actuals sum to $5.01, and the vendor's own priceToEarningsRatioTTM of 12.8204 implies exactly $5.010 at the current price. So the P/E ratio is right and the per-share earnings figure feeding it is not. U.S. Bancorp has $6,808M of preferred stock outstanding; the ~$250M annual difference is the preferred dividend. We use $5.01.

6. bookValuePerShareTTM and tangibleBookValuePerShareTTM do not deduct preferred equity. The vendor reports $43.690 and $31.958; shareholdersEquityPerShareTTM of $43.393 is essentially identical to book value, confirming no preferred deduction. Deducting $6,808M of preferred across ~1,554M shares removes $4.38 per share, giving common book of approximately $39.31 and tangible common book of approximately $27.58. Price-to-book is therefore ~1.63x, not the vendor's 1.480x, and price-to-tangible-book is ~2.33x. Both corrected figures are used in this dive and both are labelled as derived.

7. Enterprise value, net debt and every derived multiple are structurally inapplicable to a deposit-funded bank — REJECTED as a class. enterpriseValueTTM $129.57B, netDebtToEBITDATTM 2.72x, evToSalesTTM 2.96x, evToEBITDATTM 11.93x, interestCoverageRatioTTM 0.74x, netCurrentAssetValueTTM −$591,532M, grahamNetNetTTM −$380.65. None is used. The measures used are CET1 (10.8%), the net charge-off rate (0.56%), allowance coverage (2.00% of loans, 536% of nonperforming loans), net interest margin (2.77%) and return on equity.

8. capitalExpenditure is exactly zero in all four years of cf_a, so free cash flow equals operating cash flow by construction. A bank operating 1,056 owned and 1,409 leased facilities does not spend zero on premises and equipment; the line is unpopulated rather than measured. freeCashFlowYieldTTM (7.56%), capexToRevenueTTM (0), capexToDepreciationTTM (0), priceToFreeCashFlowRatioTTM (13.23x) and incomeQualityTTM (0.926) are all rejected. Note the direction of this defect is the opposite of the COP case flagged in this programme — there capex was understated by 76% and inflated the FCF yield; here it is simply absent.

9. The BTIG acquisition is absent from the payload, and its completion is unconfirmed by the archive. Announced January 2026 at up to $1 billion including 6,600,594 shares, expected to close in Q2 2026. bal_a ends 2025-12-31; cf_a ends FY2025; there is no FY2026 row anywhere. Nor does the archive settle it: the 2026-07-16 8-K is a cover page only and there is no June-quarter 10-Q. We report it as pending.

The filing archive itself is a finding, and it is the most restrictive in this batch. No file in the USB archive contains a single [TABLE] marker — a grep -c returns zero on both the 10-K and the 10-Q. And the 10-K contains no financial statements at all: U.S. Bancorp's Form 10-K incorporates the income statement, balance sheet, segment review, capital management section and risk factors by reference to its separately-published 2025 Annual Report ("Additional information regarding the Company's business segments can be found on pages 53 to 54 of the Company's 2025 Annual Report... which is incorporated herein by reference"), and that document is not in this archive. Consequently every filing-sourced figure in this dive is either a prose statement in the March 2026 10-Q's MD&A or a percentage change quoted there. We could not verify a single line item of the balance sheet or income statement directly, and we say so rather than working around it. Where a dollar figure appears above with a "~", it is the 10-Q's own rounded prose figure ("net interest income was $4.3 billion").

Also checked and recorded: quote.yearHigh/yearLow ($64.85 / $43.83) against tech.hi52/lo52 ($64.47 / $44.09) — a 0.6% discrepancy at both ends; tech is used throughout for internal consistency, which is the more conservative choice here since it makes the stock look slightly further from its high. effectiveTaxRateTTM of 19.7% reconciles well to the 10-Q's 19.4% for Q1 2026. dividendPerShareTTM of $2.08 matches lastDividend; the forward run-rate from Q3 2026 is $2.16 per the 2026-06-24 8-K, and we use the forward figure where labelled. Implied share count (market capitalisation ÷ price = 1,557.7M) is consistent with revenuePerShareTTM (1,553.8M) to within 0.25%.

returnOnEquityTTM — which block we used. The figure appears in km_ttm at 0.12491 (12.49%) and is absent from ratios_ttm entirely. This is the inverse of the null-in-one-block defect flagged across seven names in the last batch; here there is no conflict, only a single source. We use km_ttm's 12.49% and note that it is struck on total equity including the $6,808M of preferred, so return on COMMON equity is roughly 12.7% and return on tangible common equity materially higher.

Vendor composite rating — noted, not used. B+ / 3 overall, with a discounted-cash-flow sub-score of 5 out of 5. That sub-score is built on the free-cash-flow figure rejected in finding 8 and carries no weight. The debt-to-equity sub-score of 2 penalises a bank for having a balance sheet.

Non-equity tripwire — checked and passed. USB is common stock, $0.01 par value, NYSE-listed. Price of $64.23 is not par-like; beta is 0.979; volume was 9.42M shares (~$605M of turnover); the 52-week band of $44.09 to $64.47 is a 46% range; the dividend is variable and rising. Note the separately-registered Series A, B, K, L, M and O preferred stock and two series of senior notes listed on the 8-K cover pages — none of those is this security. This is common equity.

9. Technicals

Price $64.23. −0.4% from the 52-week high of $64.47; +45.7% above the 52-week low of $44.09. Position within the annual range: 99th percentile.
Above both moving averages and both are rising: +7.1% above a 50-day average of $59.99; +17.1% above a 200-day average of $54.84, with the 50-day well above the 200-day.
RSI 56.5 — notably unstretched for a stock at its 52-week high, which suggests the advance has been steady rather than a spike. MACD +0.97.
Maximum drawdown from peak over the trailing year: −0.4% — i.e. the stock is at its peak. This is the cleanest uptrend in the batch and it is also the reason there is no entry discount.
Relative performance: 3-month +16.3% against SPY +7.6% and QQQ +7.7%; 6-month +12.0% against SPY +11.1%; 12-month +46.2% against SPY +24.3% and QQQ +30.8% — a 22-point outperformance against the S&P in a year in which EPS grew about 15%. The gap between those two numbers is the re-rating, and it is the substance of the Hold.
Sentiment: 1 strong buy, 24 buy, 22 hold, 2 sell across 49 analysts; consensus target $69.33 (+7.9%), median $68 (+5.9%), high $75 (+16.8%), low $61 — 5.0% BELOW spot. Roughly half the covering analysts are at hold or worse, and the published distribution spans −5% to +17%. That is a market that thinks this is fairly valued, and we agree.

Today's move and what it does to the entry

USB closed 2026-08-04 at $64.23, up 0.45% or $0.29 from $63.94. It opened at $63.53, traded $63.51 to $64.85 — printing a new intraday 52-week high — and closed 0.96% below the day high on 9.42M shares. No company-specific filing is dated 2026-08-04; the last events were the 2026-07-24 insider filings and the 2026-07-16 earnings release.

The honest read: there is no entry here, only a decision about whether to keep holding. The stock made a new high today. It is 17.1% above its 200-day average. RSI at 56.5 says the move is orderly, not exhausted — which argues against calling a top, and equally against paying up. A Hold verdict on a stock at a 52-week high is not a bearish call; it is a refusal to buy a 4% base case with 19% of downside because the chart looks good.

10. Insiders — eight transactions, and the pattern is exercise-and-sell

Date	Person	Role	Type	Shares	Price	Held after
2026-04-23	Richard P. McKenney	Director	A-Award (deferred comp)	2,560	$0	34,342
2026-04-23	Richard P. McKenney	Director	A-Award (RSUs)	3,355	$0	3,355
2026-05-05	Dilip Venkatachari	SEVP & Chief Information and Technology Officer	S-Sale	34,522	$55.52	51,292
2026-07-20	Stephen L. Philipson	Vice Chair	S-Sale	36,906	$63.077	74,969
2026-07-21	James L. Chosy	Senior EVP and General Counsel	G-Gift	3,969	$0	250,941
2026-07-23	James L. Chosy	Senior EVP and General Counsel	M-Exempt (option exercise)	24,711	$55.01	275,652
2026-07-23	James L. Chosy	Senior EVP and General Counsel	S-Sale	22,968	$63.28	252,684
2026-07-23	James L. Chosy	Senior EVP and General Counsel	M-Exempt (option disposition)	24,711	$55.01	0

Three open-market sales, two option exercises, one gift and two director awards. Not one purchase.

The reading, with its qualifications. The Chosy sequence on 2026-07-23 is the mechanical one: exercise 24,711 options at $55.01, sell 22,968 shares at $63.28, retain the balance. That is an exercise-and-sell to fund the strike and the tax, and it carries little signal — he still holds 252,684 shares, a position worth roughly $16 million.

The two that carry more signal are the standalone sales. Stephen Philipson, Vice Chair, sold 36,906 shares at $63.077 on 2026-07-20 — four days after the earnings beat, at a price 1.8% below today's — reducing his holding to 74,969, a 33% reduction. Dilip Venkatachari, the Chief Information and Technology Officer, sold 34,522 at $55.52 on 2026-05-05, reducing to 51,292, a 40% reduction. Both were straight S-Sales with no accompanying option exercise, and both are the largest proportional reductions in the file.

What the file does not contain: any transaction by Gunjan Kedia, the chief executive, or by the chief financial officer. No purchase, no sale. Nor is there a single open-market purchase by anyone. On a stock at its 52-week high after a 46% year, an insider file containing three sales, two of them large proportional reductions, and zero purchases is mildly negative and consistent with the Hold.

11. Verdict, kill-criteria and flip conditions

Hold.

The case for owning it. U.S. Bancorp runs a 10.8% CET1 ratio against a newly-confirmed 7.1% requirement, funds 75% of a $700 billion balance sheet with $528.2 billion of deposits, charges off 0.56% of loans a year with the rate falling, carries an allowance at 2.00% of loans and 536% of nonperforming loans, and grew diluted EPS 14.6% in the March quarter on net interest margin expansion to 2.77% and noninterest expense growth of 0.8%. The June quarter beat consensus by 5.5% on rebuilt net revenue of $7,686M, +10.2% year on year — the strongest quarter in the series. The dividend rises to $0.54 a quarter from Q3, a 3.36% forward yield. At 11.0x FY2027 consensus that is a perfectly sensible thing to own.

The case against buying it, which is the operative one. Our base fair value of $67 is 4.3% above spot. The street's $69.33 is 7.9% above and its low target of $61 is 5.0% below. Asymmetry is 1.13:1. The stock is 0.4% from its 52-week high after a 46.2% twelve-month move against 15% EPS growth — three-quarters of that move was multiple, not earnings. The bank earned $5.11 per share in FY2021 and consensus wants $5.22 in FY2026: five years of flat earnings, then a re-rating. Roughly half the forward growth depends on net interest margin expansion driven by "fixed asset repricing", which is by definition exhaustible. Payment Services, the only segment with a structural growth claim, earned less than a year ago. And the knowledge base contributes nothing — three hits, all a computer cable.

We are not going to manufacture a Buy out of a 4% base case. Nor is there anything here that justifies Avoid: this is a well-capitalised bank with improving credit and a settled capital requirement, and the downside case requires an actual credit turn, not merely a disappointment. Hold is the honest tier.

Pre-registered KILL criteria — what would take this to Avoid:

Net interest margin below 2.72% in any quarter, reversing the expansion that carries half the consensus growth.
Total net charge-offs above 0.85% of average loans, roughly half again the current 0.56%, or a second consecutive quarter of rising commercial-loan charge-offs.
Allowance for credit losses below 1.90% of loans, confirming that reserve release rather than operating performance is carrying reported earnings.
CET1 below 9.5%, or any change to the 2.6% stress capital buffer before its 2027-10-01 expiry.
Noninterest expense growth above 4% year on year, which would end the operating-leverage story on its own.
A BTIG closing on materially worse terms, or evidence that the institutional trading business is dilutive to returns.

Pre-registered UPGRADE conditions — what would take this to Buy — Tactical:

A price below $56, which is 9.6x FY2027 consensus and roughly 1.42x common book. At that level the same forecast produces a 20% base-case return and the verdict changes on arithmetic alone.
Full-scale use of the $5.0 billion repurchase authorisation. FY2025 saw $489 million. Running the authorisation over two years adds roughly 2.5 points a year to total shareholder return and would be the single most effective thing management could do.
Payment Services net income growing again, which would support the fee-bank multiple the current price implies.
A second and third quarter of net revenue above $7.9 billion, confirming the June quarter's +10.2% was a trend and not a rate-cycle artefact.
Insider buying, or at minimum the end of the current pattern of standalone executive sales.

Where USB fits in the Synthos Framework Portfolio. The financials sleeve, held at existing weight if already owned, no new capital today. Sizing note: the correct response to a fairly-valued, well-run bank at a 52-week high is to have a price at which you would buy it, not to buy it at the price it is. Ours is $56. Logged as a tracked Synthos call (Hold) as of 2026-08-04 at $64.23, with the fair-value anchors, kill criteria and the $56 upgrade trigger all gradeable.

Single biggest risk: multiple normalisation. Nothing in this bank's disclosure is alarming. Everything about its price is full. At 11.0x FY2027 consensus and roughly 1.63x common book it sits at the upper end of the band this bank has occupied for a decade, having got there by re-rating rather than by earning more. A return to 9.5x on the same, unchanged FY2027 estimate is $55 — a 14% decline with no operational failure required. The credit book is clean, the capital is settled and the dividend is rising; none of that prevents a multiple from giving back what it took. The kill criteria above are set on the operating series precisely because the valuation risk is not the kind that announces itself.

Provenance & disclosures

Traceability: ZERO knowledge-base claims name U.S. Bancorp out of 51,928 distilled claims (raw hits 3, entity matches 0, used 0, discarded 3; breadth 0, net conviction none). All three discards are the Universal Serial Bus, caught by a case-sensitive free-text pattern on the three-letter ticker: huberman_lab 2024-04-24 (hippocampus as a "USB stick"), lyn_alden 2025-09-27 and 2026-04-26 (USB alongside Ethernet and TCP/IP as entrenched protocols, both with Bitcoin as the entity). The lane is reported as empty rather than padded with adjacent macro claims about banks or consumer credit, none of which name this company. No management voice appears in the lane; no concentration test is meaningful on an empty pool. All quotes are verbatim from the stored claim text.
Data as-of: fundamentals — loans, deposits, capital ratios, credit metrics, net interest margin and segment performance through 2026-03-31, from the 10-Q filed 2026-05-04; second-quarter results from the vendor payload only, since the 2026-07-16 8-K carries no extractable exhibit; the stress capital buffer and dividend increase from the 8-K filed 2026-06-24 · estimates 2026-08-04 · prices 2026-08-04, quote timestamp 1785873602 = 2026-08-04T20:00:02Z ($64.23, +0.45%; 50-DMA $59.99; 200-DMA $54.84; RSI 56.5; MACD +0.97) · knowledge-base claims 2026-08-04. U.S. Bancorp's fiscal year is the calendar year. All figures come from the Synthos vendor data file for USB or from the SEC filings in the USB archive; no figure comes from memory, recall or external retrieval.
Filing archive contents and its limits — this is itself a finding. 10-K filed 2026-02-23 (fiscal 2025); 10-Q filed 2025-11-05 (September 2025 quarter); 10-Q filed 2026-05-04 (March 2026 quarter — the source of every filing-verified figure in this dive); 8-K filed 2026-04-23 (annual meeting voting results); 8-K filed 2026-06-24 (stress capital buffer held at 2.6% until 2027-10-01; quarterly dividend rising $0.52 → $0.54 from Q3 2026); 8-K filed 2026-07-16 (Item 2.02 cover page announcing second-quarter results — the earnings release is Exhibit 99.1 and is not in the extracted text). NO file in this archive contains a [TABLE] marker — statement tables were not preserved for this name, verified by grep -c returning zero on both the 10-K and the 10-Q. Furthermore the 10-K contains no financial statements: U.S. Bancorp incorporates its income statement, balance sheet, segment review, capital management discussion and risk factors by reference to its separately-published 2025 Annual Report, which is not in the archive. Every filing-sourced figure above is therefore a prose statement or percentage change quoted in the March 2026 10-Q's MD&A, and no line item of either primary statement was independently verified. There is no June-quarter 10-Q.
Where the filings contradicted or corrected the vendor: the revenue basis — the vendor reports FY2025 revenue of $42,861M where the company publishes total net revenue of $28,540M, a 50.2% gross-up reconciling exactly through the vendor's own interest-expense line and confirmed against the payload's own est.revenueAvg ($28,598M, 0.2% apart) and earn_cal.revenueActual (four quarters, all within 0.4%); the growth direction — on the vendor's line FY2024 revenue GREW 5.1%, on the published basis it FELL 2.4%; the segment structure — the 10-Q's four reportable segments (Wealth, Corporate, Commercial and Institutional Banking; Consumer and Business Banking; Payment Services; Treasury and Corporate Support) bear little relation to seg_prod's three stale names summing to 107.4% of net revenue, with Treasury and Corporate Support present in FY2024 and absent in FY2025; seg_geo, which holds a single FY2012 entry of anonymised "Segment Two/Three/Four" placeholders against the 10-Q's actual disclosure that Europe is ~2% of total net revenue; and the BTIG acquisition — announced January 2026 at up to $1 billion including 6,600,594 shares, absent from every vendor field and not confirmed closed by anything in the archive. Where vendor and filing AGREED — recorded: effectiveTaxRateTTM of 19.7% against the 10-Q's 19.4% for Q1 2026; earn_cal Q1 2026 EPS actual of $1.18 exactly matching the 10-Q's reported diluted EPS; and the vendor's own priceToEarningsRatioTTM of 12.82x implying trailing EPS of $5.010, which matches the sum of the four earn_cal actuals to the cent.
Internal vendor contradictions recorded separately from filing conflicts: netIncomePerShareTTM ($5.260) is inconsistent with priceToEarningsRatioTTM (12.8204, implying $5.010) because the former does not deduct preferred dividends on $6,808M of preferred stock — we use $5.01; bookValuePerShareTTM ($43.690) and tangibleBookValuePerShareTTM ($31.958) likewise omit the preferred deduction, confirmed by shareholdersEquityPerShareTTM ($43.393) being essentially identical to book value — corrected to approximately $39.31 and $27.58, giving 1.63x and 2.33x rather than the vendor's 1.48x; and quote.yearHigh/yearLow ($64.85 / $43.83) differ from tech.hi52/lo52 ($64.47 / $44.09) by 0.6% at both ends, with tech used throughout.
Rejected as a class: all enterprise-value measures (enterpriseValueTTM $129.57B, netDebtToEBITDATTM 2.72x, evToSalesTTM 2.96x, evToEBITDATTM 11.93x, interestCoverageRatioTTM 0.74x, grahamNetNetTTM −$380.65, netCurrentAssetValueTTM −$591,532M) as structurally inapplicable to a bank funded by $528.2 billion of customer deposits; all cash-flow measures (freeCashFlowYieldTTM 7.56%, incomeQualityTTM 0.926, priceToFreeCashFlowRatioTTM 13.23x) because capitalExpenditure is exactly zero in all four years of cf_a for a company operating 2,465 facilities; and the entire est EBITDA/EBIT block for the fixed-ratio signature (ebitdaAvg exactly 27.556% and ebitAvg exactly 25.427% of revenueAvg in every forward year).
returnOnEquityTTM — source stated: present in km_ttm at 12.49% and absent from ratios_ttm. We use km_ttm's figure and note it is struck on total equity including $6,808M of preferred; return on common equity is approximately 12.7%.
Estimate coverage: 16 analysts on both FY2026 and FY2027 diluted EPS — the anchor for all three fair values, with an FY2027 range of $5.442 to $5.968 (9.7%); 8 on FY2028; 3 on FY2029, which is excluded from every conclusion. Revenue coverage is thinner (8, 10 and 7 analysts) and is used only for context.
Peer note: the vendor peer set is Barclays, CIBC, Deutsche Bank, HDFC, ING, Itaú, Lloyds, Mizuho, PNC and Truist. Eight of ten are non-US institutions under different capital regimes and accounting standards; only PNC and Truist are genuine comparables. No peer-multiple comparison is drawn.
Fair-value caveat: the $52 / $67 / $78 anchors are multiples of the FY2027 consensus diluted EPS distribution — 9.5x the low of $5.442, 11.5x the mean of $5.816, and 13.1x the high of $5.968 — each cross-checked against derived common book value of approximately $39.31 per share (1.32x, 1.70x and 1.98x). Stated arithmetic, not a discounted cash flow. The base is sensitivity-disclosed: 11.0x gives $64, which is spot; 12.0x gives $70. The entire verdict lives inside one multiple point, and that fact is the reason it is a Hold.
Bank-metric note: the measures used throughout are CET1 (10.8% against a 7.1% requirement), the total loan net charge-off rate (0.56% of average loans, down from 0.59%), allowance coverage (2.00% of loans, 536% of nonperforming loans), net interest margin on a taxable-equivalent basis (2.77%, up from 2.72%), the deposit funding share (~75% of total assets), and return on equity (12.49%). Cash-flow and enterprise-value measures are not used anywhere in this dive.
Timing: second-quarter 2026 results were released 2026-07-16, nineteen days before this dive, beating consensus EPS by 5.5% ($1.35 against $1.28) on revenue of $7,712M against $7,577M estimated. The next print is 2026-10-15, 72 days away (consensus EPS $1.32, revenue $7,964M). The most recent insider filings are dated 2026-07-22 and 2026-07-24. 2026-08-04 carried no company-specific filing, and the stock printed a new intraday 52-week high of $64.85.
Accessibility note: no information in this dive is conveyed by colour. All emphasis is carried by bold text, table structure and explicit labelling.
Not investment advice. Independent research, educational and informational only, never personalised. No recommendation to buy, sell or hold any security is made to any person.
Version: 2026-08-04-full.