Q2 Delivered the Inflection: 8.0% Domestic Margin and 26% More Profit Per Piece, Sold Off on a GAAP Charge Consensus Never Modelled

Results vs. Consensus

Q2 2026 Scorecard

Consensus providers clustered tightly this quarter. The compiled 24-analyst figure was $1.66 on adjusted EPS and $21.84B on revenue; one provider carried $1.65 and $21.75B. The table uses the compiled pair. On the lower EPS figure the beat reads +6.7% rather than +6.0%, so the headline characterization does not change either way. Revenue consensus was materially below the print in every version of it, and the reason is entirely fuel: consolidated revenue per piece rose 11.3% while consolidated volume fell 3.6%.

The GAAP row is the important one, and it is the row that moved the stock. Sell-side models carried roughly $1.64 of GAAP EPS, meaning the $891M after-tax Driver Choice separation charge was not in the published estimate base. That is not a company disclosure failure: the 8-K filed with the first-quarter results sized the Driver Choice program at $1.2B of 2026 excluded operating expense with only $55M incurred through March, which is where our April note derived a roughly $1.09 per share second-quarter wedge. It printed at $1.05. The charge was knowable, and the print collided with an estimate base that had not absorbed it.

Performance Against the Company's Own Q2 Guidance

UPS gave no consolidated second-quarter revenue or EPS guide in April, only three segment ranges. That makes the guide-versus-actual scorecard more informative than the consensus one, because it is the standard management set for itself.

Year-Over-Year Comparison

UPS reports cash flow only on a year-to-date basis in the earnings exhibit, so the two cash lines above are six-month figures against six-month comparatives while every other line is the quarter.

Quarter-Over-Quarter Comparison

The sequential table is where the inflection is visible without any adjustment or interpretation. The first quarter was the trough of the transformation and the second quarter is the first period on the other side of it.

Revenue. Consolidated revenue of $22,834M grew 7.6%, and every segment contributed. Consolidated revenue fell 1.6% in the first quarter and 2.6% for the whole of 2025, so this is the first period in which the top line has worked with the strategy rather than against it. The composition is the reverse of the first quarter: then, price offset a collapsing volume base almost exactly; now, price is running well ahead of a volume base that is declining much more slowly. Consolidated average daily volume fell 3.7% against 7.7% in the first quarter, and consolidated revenue per piece rose 11.3% to $15.96. U.S. Domestic revenue grew 6.0% on volume down 3.3%. International grew 12.5% on volume down 5.8%, an extraordinary 1,830bp of price-and-mix offset that is mostly fuel and geographic mix rather than base rate. Supply Chain Solutions returned to growth at 7.8% after falling 6.5% in the first quarter, which resolves the one line in the March quarter where the revenue trajectory was worse than the strategy explained.

Margins. This is where the quarter was won, and specifically in the segment that is 65% of revenue. U.S. Domestic adjusted operating profit of $1,188M rose 21.0% on revenue up 6.0%, and the segment adjusted margin of 8.0% was up 100bp year over year and 400bp sequentially. That is the number the entire 2026 thesis was resting on when we initiated at Hold in April, and it landed at the midpoint of the range management guided. International went the other way, giving up 280bp to 12.4% with adjusted operating profit down 8.7%, of which management attributes 120bp to fuel because the segment carries more air volume over longer distances and therefore a larger fuel share of its cost base. Supply Chain Solutions expanded 220bp to 10.2%, its third consecutive quarter of year-over-year margin expansion.

EPS. The two EPS lines tell opposite stories and both are true. Adjusted diluted EPS of $1.76 grew 13.5% and beat the consensus by 6.0%. GAAP diluted EPS of $0.71 fell 53.0% and missed a GAAP consensus of roughly $1.64 by more than half, because $1,172M of pre-tax transformation cost ran through the income statement, $891M after tax, $1.05 per share. Net income of $604M fell 52.9%. A reader deciding which number matters should note two things. First, the charge is cash separation cost for people who have already left: roughly 80% of Driver Choice participants departed in the second quarter, and the guided full-year free cash flow of $5.5B already absorbs the payments. Second, it is nearly finished. The 8-K guides full-year excluded operating expense to $1.3B to $1.5B and the first half booked $1,227M of it, so $73M to $273M remains for the back half. The GAAP-to-adjusted wedge substantially closes from here, which means the third and fourth quarters should print GAAP numbers that look much more like the adjusted ones.

Segment Performance

Revenue by Product Line

Next Day Air revenue growing 12.4% while its average daily volume fell 1.3% is the premium-mix claim in a single line. The company's most expensive product is repricing faster than anything else it sells, and it is doing so in a quarter when the cheap end of the book was being deliberately removed.

Volume and Pricing KPIs

U.S. Domestic Package

Revenue of $14,930M grew 6.0% and adjusted operating profit of $1,188M grew 21.0%, more than double the $565M the segment earned in the first quarter. The 8.0% adjusted margin was up 100bp year over year and 400bp sequentially, and it landed at the midpoint of the range management guided in April. Average daily volume fell 3.3%, a sharp deceleration from the 8.0% decline of the first quarter, because the Amazon glide-down finished inside the quarter. Excluding Amazon, total U.S. air average daily volume grew 1.2%, and management said that adjusting for both Amazon and the other low-yield volume it has deliberately de-selected, U.S. volume grew year over year.

Customer mix improved on every disclosed measure and, importantly, in absolute terms rather than only by penetration. SMB average daily volume grew 4.3%, so the record 34.5% SMB share of U.S. volume, up 250bp, is not an artifact of a shrinking denominator this quarter the way it partly was last quarter. B2B average daily volume fell 3.2%, a rate of decline 190bp better than the first quarter, and B2B represented 43.8% of U.S. volume. B2B volume moving through the Digital Access Program grew 34%, and global DAP revenue reached $1.4B, a third consecutive quarter above $1B.

"Strong base rate growth and increased productivity in our reconfigured network contributed to revenue per piece growing 130 basis points faster than the cost per piece growth rate, demonstrating the operating leverage we expected from our network reconfiguration."
— Brian Dykes, CFO

Total segment expense rose 4.9% on an adjusted basis, with more than half of the increase from fuel and purchased transportation. Purchased transportation is the line to watch, because the USPS tender for Ground Saver last-mile now runs at scale and sits inside it.

Assessment: This is the quarter the April note said was the entire thesis, and it delivered the version of the test that mattered. A 21% increase in segment operating profit on 3.3% fewer pieces is operating leverage produced by structure rather than by volume, and the disclosure is precise enough to verify from outside: the contribution-per-piece math reconciles to the reported profit change within $8M. The caution is the shape of what comes next. Management guides the third quarter to approximately 7.0%, a sequential step down, and the full year to approximately 7.5%, which is below 2025's 7.74%. The segment has proved it can hit a number; it has not yet proved the number compounds.

International Package

International was the only segment to miss its own guide, and the only one where adjusted operating profit fell. Revenue of $5,044M grew 12.5% on a 5.8% volume decline, an enormous price offset driven by an 18.9% increase in revenue per piece with a little more than half of that coming from fuel. Adjusted operating profit of $623M fell 8.7% and the 12.4% adjusted margin fell 280bp from 15.2%, landing below the 13% to 14% range management guided in April.

The operational news underneath was better than the margin. The China-to-U.S. lane returned to year-over-year volume growth in May as the de minimis elimination lapped, and Asia-to-Asia export volume grew 13.6% on investment the company made in the region. The damage was cost, and management was specific about its composition.

"I would say in the U.S., we were able to overcome that and still get to the 8% margin. On the international side, it's more impacted by fuel because you more air volume, you got longer fly distances, it's a bigger proportion of the total cost base."
— Brian Dykes, CFO

Fuel is sized at 120bp of the 280bp decline. The balance is the block hours and leased aircraft cost required to route around Middle East airspace, plus geographic mix that is still rebalancing. Management held the third-quarter guide at 13% to 14% and left the full-year target at mid-teens.

Assessment: Fuel explains 120bp of a 280bp decline, which leaves 160bp that is either mix or genuine cost inflation, and management did not decompose it. The larger problem is arithmetic rather than diagnostic. The first half produced a 12.2% adjusted margin. With the third quarter guided to 13% to 14% and full-year revenue guided to mid-single-digit growth, a 15% full-year margin requires a fourth quarter near 21.7%, which is close to the 21.6% the segment printed in the fourth quarter of 2024. At exactly 14%, the bottom edge of mid-teens, the fourth quarter needs 17.8%, which is achievable because it is essentially last year's 18.0%. So the target is reachable only at its floor and only with a Q4 that repeats a record. This is the guide we expect to be trimmed rather than met, and it is the reason our full-year estimate sits below the company's.

Supply Chain Solutions

The smallest segment resolved the one inconsistency the first quarter left open. Revenue of $2,860M grew 7.8% after falling 6.5% in March, driven by forwarding up 8.1% on higher international air freight rates and logistics up 4.3% on health care. UPS Digital, which houses Roadie and Happy Returns, grew revenue more than 30%. Adjusted operating profit of $291M grew 37.3% and the 10.2% adjusted margin expanded 220bp, a third consecutive quarter of year-over-year margin expansion. Management guides the full year to 10% to 11%, against 10.62% in 2025.

Assessment: In April this segment had delivered the margin half of its guide and none of the revenue half, and we flagged the reacceleration to high-single-digit full-year growth as the least supported line in the guide. It has now delivered both halves in the same quarter, with the top line at 7.8% and margin still expanding. That closes the gap. The one thing to keep in view is the third-quarter comparison: Supply Chain Solutions earned a 21.3% adjusted margin in the third quarter of 2025, so the guided 10% to 11% is a very large year-over-year decline that has nothing to do with the segment's underlying trajectory.

Operating Expense Detail

The GAAP expense table is distorted by the transformation charge and should be read alongside the reconciliation. Compensation and benefits rose 8.8% in a quarter when the company reduced operational positions by nearly 30,000 year over year, because the $1,172M of separation cost sits in that line. Strip it and compensation falls, which is the same direction the line moved in the first quarter. Fuel is the second distortion, up 60.4% on a $639M increase that the surcharge mechanism recovered. Purchased transportation up 26.4% is the line that carries real information and no distortion: it reflects the USPS tender for Ground Saver last mile running at full scale, plus the leased aircraft capacity supporting both the MD-11 replacement and the Middle East reroutes. Depreciation rose only 4.7% against 8.0% in the first quarter, which is the first sign that the automation capex wave is flattening on the P&L.

Key Topics & Management Commentary

Overall Management Tone: The posture shifted from defending a plan to reporting a finished one. Where the April call answered the margin bridge with a list of levers, this call answered it with a delivered number and then moved the conversation to what the network does next, which is a different kind of confidence and a better-earned one. The soft spots were narrow and both were about extrapolation rather than execution: management would not put a longer-term domestic margin target on the record when asked directly for one, and it declined to decompose the International shortfall beyond the fuel component. Analyst pushback concentrated on why the second half is guided to normal seasonality rather than better, and the answer was a restatement of the original plan rather than a walk.

1. The Domestic Bridge Was Built, and It Is Visible in Unit Economics

The single question hanging over UPS in April was whether a 4.0% U.S. Domestic margin could become 7.5% to 8.5% in ninety days. It became 8.0%. What matters more than the level is where it came from, because a seasonal step-up and a structural reset look identical in a margin percentage and completely different in a model. The company gave enough disclosure to tell them apart. Revenue per piece rose 9.3%, adjusted cost per piece rose 8.0%, and the resulting contribution of $1.15 per piece against $0.91 a year ago is up 26.4% on a base of 3.3% fewer pieces.

"But this reconfiguration was never the destination. It was the foundation. We now have a leaner, more automated, more agile network that will deliver operating leverage as volume grows."
— Carol Tomé, CEO

The supporting operational disclosure is the cumulative view management gave for the first time: since the start of 2024 the company has removed roughly 2 million pieces per day of lower-yielding Amazon volume and about $4.5B of related expense.

"we will have eliminated 50 million hours through the course of last year and this year, nearly 78,000 operational [positions] that were associated with that volume, and we'll close nearly 150 building."
— Brian Dykes, CFO

Assessment: Take the two disclosures together and the case is coherent: capacity came out faster than volume, so the remaining pieces carry more contribution. That is the definition of a structural reset, and it is the first quarter in this story where the claim is arithmetic rather than assertion. The honest limit is that a single quarter of a positive spread is a data point, not a trend, and the third-quarter guide immediately narrows the spread again as prior-year pricing actions lap. We are treating the reset as real and the compounding as unproven.

2. The Amazon Glide-Down Is Finished

The 50% Amazon volume reduction completed in June, on the date management committed to in April and in January. The most consequential thing about it is that the risk is now dated rather than open-ended, and the volume comparison stops getting worse.

"Amazon made up 9%. And that's down about 100 basis points from a year ago and certainly down from the which was over 13% during the COVID year."
— Carol Tomé, CEO

The disclosure remains a percentage. Asked directly for a dollar figure now that the glide-down is complete, management gave the same 9% and no number, no floor, and no colour on the pricing of the retained book. What it did give is a forward volume statement.

"we're going to see volume growth ex Amazon in the back half of this year across all segments, right?"
— Brian Dykes, CFO

Assessment: A completed exit is worth more than a promised one, and the U.S. volume decline decelerating from 8.0% to 3.3% in a single quarter is the mechanical proof that the drag is done. The disclosure gap has not closed: for a second consecutive quarter, the customer whose departure defined three years of this story is described only as a percentage that rounds. We continue to treat the post-June Amazon book as a flat contributor rather than a growth vector until UPS discloses enough to model it.

3. Fuel Inflated Both Sides of the Print

Consolidated fuel expense rose 60.4% to $1,697M, a $639M increase driven by the Middle East conflict. That is the largest single distortion in the quarter and it runs through every growth rate in the report. Management was clear about what it does and does not do to profit.

"Our fuel surcharge mechanisms functioned as designed, covering the increase in fuel expense."
— Carol Tomé, CEO

The CFO decomposed the U.S. revenue-per-piece line explicitly and gave the forward path.

"So we were at 9.3% in second quarter for RPP in the U.S., about half of that was fuel and about half of it was base pricing and mix."
— Brian Dykes, CFO

Assessment: Fuel is why revenue beat consensus by 4.6% while adjusted operating profit grew only 12.0%, and it is why the guidance raise carries a 2.6% incremental margin. For modelling purposes the important consequence is directional and slightly favourable to the bull case: because the cost-per-piece base ($12.12) is smaller than the revenue-per-piece base ($13.03), an equal dollar of fuel inflates the cost growth rate more than the revenue growth rate. The reported positive 130bp spread is therefore, if anything, understated relative to an ex-fuel view. The dollar contribution figure of $1.15 versus $0.91 sidesteps the distortion entirely, which is why we lead with it.

4. The Guidance Raise Is $1.5 Billion of Revenue for $39 Million of Profit

UPS raised the full-year revenue outlook to approximately $91.2B from approximately $89.7B and the adjusted operating profit target to approximately $8.65B. In April the company guided approximately $89.7B at an approximately 9.6% adjusted operating margin, which is $8,611M of adjusted operating profit. So the revenue guide rose $1.5B and the profit guide rose $39M, a 2.6% incremental margin, and the implied full-year adjusted operating margin fell about 12bp to 9.48%.

Assessment: Three of the four headline changes are genuine improvements and one is a quiet reduction. The EPS guide moved from a spoken "about flat" to a filed $7.22, which is a real increase in commitment weight as well as in level, and it sits above the roughly $7.13 the Street carried. The revenue raise is fuel and should be discounted as such. The line nobody will lead with is that the full-year U.S. Domestic margin target moved from flat with 2025's 7.74% to approximately 7.5%. That is a 25bp reduction in the segment that is 65% of revenue, made in the same quarter that segment beat, and it is the arithmetic price of the third-quarter step down. It does not change the profit-dollar picture because the revenue base grew, but it is the one place where the guide got worse and the release did not say so.

5. International Missed Its Range and the Full-Year Target Now Needs a Record Quarter

A 12.4% adjusted margin against a 13% to 14% guided range is the only guide UPS missed this quarter, and it is a miss in the direction that matters, because International is where the second-half consolidated improvement is supposed to come from. Asked directly whether normalizing Asia volume brings margins back toward the upper teens, management pointed to trade-lane momentum and the September lap of the global de minimis elimination rather than to a number.

"Look, we're seeing volume recover as trade lanes settle Also, in the back half of this year, we have the wrap of the elimination of global de minimis in September that drives a pretty big year-over-year comp benefit."
— Brian Dykes, CFO

Later in the call the CFO was more explicit that International carries the second-half step-up.

"Where you're seeing the improvement is in the international business."
— Brian Dykes, CFO

The arithmetic of the full-year target is unforgiving. The first half produced a 12.2% adjusted margin on $9,584M of revenue. Taking mid-single-digit full-year revenue growth on 2025's $18,576M and the third-quarter guide at its 13.5% midpoint, the required fourth-quarter margin is 17.8% for a 14.0% full year, 21.7% for a 15.0% full year, and 25.6% for a 16.0% full year. The segment's fourth quarter of 2025 was 18.0%, against 21.6% in the fourth quarter of 2024.

Assessment: Mid-teens is reachable only at its floor, and only by repeating a record. The de minimis lap is real and mechanical, so some improvement is close to certain; the question is the size. We model International at roughly 14% for the full year, which is inside management's language and below where a reader would naturally place "mid-teens," and it is the single largest reason our full-year estimate sits below the company's guide. This is the pillar of the thesis that moved the wrong way this quarter, and it deserves the same quarterly scrutiny the domestic margin got.

6. The Cost Program Cleared Two of the Three Checkpoints We Set

In April we named three verifiable checkpoints for the second quarter. Two of the three were operational and both were met. Forty-five buildings closed in the first half against 23 in the first quarter, so 22 of the 27 remaining closures happened in the second quarter as management said most of them would, with "several additional closures planned in the back half." Operational positions finished down nearly 30,000 versus the first half of last year, ahead of the roughly 25,000 reported at the end of March and consistent with the incremental 5,000 committed for the year. Approximately 80% of Driver Choice participants departed in the second quarter. The 8-K states approximately $1.2B of program benefits achieved in the first half against the $3B full-year target, so the program is at 40% of the target at the halfway mark, against 20% at the quarter mark.

Automation is the compounding piece.

"We know that the cost per piece in the automated building is about 28% lower than nonautomated building."
— Carol Tomé, CEO

By the end of the quarter 68.5% of U.S. volume flowed through an automated building, against 67.5% at the end of March and 64% a year ago, which the CEO framed as roughly 337 million more packages through automation year over year.

Assessment: The third checkpoint, cost-per-piece growth decelerating below 6%, was not met at 8.0%. We are not treating that as disqualifying, and the reason is that the metric it was meant to proxy did pass. The point of the sub-6% test was to confirm that cost growth had fallen below revenue growth by a healthy margin; the actual spread came in at positive 130bp and contribution per piece rose 26.4%, both better than the test was designed to detect. A fuel spike that lifts a $12.12 cost base by more dollars-per-percent than it lifts a $13.03 revenue base makes the absolute cost-growth threshold the wrong instrument this quarter. The spread is the right one, and it passed.

7. The GAAP Charge Landed Where We Sized It, and the Wedge Now Closes

The April note observed that roughly $1.2B of Driver Choice separation cost still sat ahead, almost all of it landing in the second quarter, worth about $1.09 per share of GAAP-to-adjusted wedge at the guided tax rate on 850M shares, against $0.05 in the first quarter. The actual wedge was $1.05, on $1,172M of pre-tax and $891M of after-tax transformation cost. The sell-side estimate base did not carry it: published GAAP consensus was near $1.64 against a $0.71 print.

The forward point is more useful than the retrospective one. The 8-K guides full-year excluded operating expense to $1.3B to $1.5B, and the first half booked $1,227M. That leaves $73M to $273M for the second half, and the release states that the initiatives are expected to conclude by 2027.

Assessment: The largest single reason UPS has screened badly on GAAP metrics for two years is about to stop. If the remaining excluded cost lands at the midpoint of what is left, third and fourth quarter GAAP EPS will sit within roughly a nickel of adjusted, and the full-year GAAP figure converges toward the adjusted one for the first time since the transformation began. For an investor base that has been told to look through adjustments for eight quarters, the disappearance of the adjustment is a quiet but real catalyst.

8. Capital Allocation: Still No Buyback, Still 98% Payout of Guided Free Cash Flow

UPS repurchased no shares in the second quarter, as it did none in the first, against $1,000M of repurchases in the first half of 2025. Repurchases were not mentioned in the release or on the call. Six-month free cash flow of $1,573M more than doubled from $742M, which is genuine improvement, but it covered only 58% of the $2,708M of dividends paid over the same period. Cash fell $1,234M to $4,653M and retained earnings fell $1,321M. Total debt and finance leases stand at $24,484M against $15,099M of equity, and interest expense rose 14.3% in the quarter and 17.0% in the half.

The full-year plan is unchanged: approximately $3.0B of capex, a $1.3B pension contribution, around $5.4B of dividends, and approximately $5.5B of free cash flow.

Assessment: The seasonal shape is normal, so the 58% first-half coverage is not the alarm it looks like in isolation. The relevant test is whether the second half can produce the $3,927M of free cash flow the guide implies, and last year's second half produced $4,728M, so the requirement is 17% below what the business has just demonstrated it can do. That is comfortable. What has not changed is the structure: a $5.4B dividend against $5.5B of guided free cash flow is 98% of everything the business expects to generate, with no repurchase and a debt load compounding interest expense at a high-teens rate. The dividend is safe this year. The optionality is gone, and it will stay gone until free cash flow moves materially above $6B.

9. The Third Quarter Steps Down, and the Volume Decline Accelerates

The forward guide is the part of this print the market disliked, and it deserves to be read precisely rather than as a headline. Management guides third-quarter U.S. Domestic revenue to approximately flat, third-quarter domestic operating margin to approximately 7.0%, and U.S. average daily volume to decline mid-single digits.

"we expect average daily volume to decline mid-single digits, reflecting a seasonal decline as well as the impact of this year's Amazon glide down, which completed in June."
— Brian Dykes, CFO

A mid-single-digit third-quarter decline is a faster rate of decline than the second quarter's 3.3%, in the first quarter after the Amazon exit completed. The margin path is also lumpy: approximately 7.0% in the third quarter against a second half guided to approximately 8.8% implies a fourth quarter near 10.3%, against 10.18% in the fourth quarter of 2025.

"we're very excited that we're going to be expanding margin in both Q3 and Q4 in our domestic business."
— Carol Tomé, CEO

Assessment: Both halves of that are true and the market weighted the first. On a year-over-year basis the third-quarter guide is an improvement: 7.0% against 6.36% a year ago is 64bp of expansion and roughly a 10% increase in segment operating profit dollars on flat revenue. Sequentially it is a step down from 8.0%, and management attributed the difference to the lapping of prior-year pricing actions plus ordinary seasonality. The volume line is the genuine concern. Mid-single-digit decline after the Amazon exit finished means the non-Amazon base is not yet growing enough to offset the residual comparison, and it is the reason we are not underwriting a domestic margin above the guide.

10. Capital Intensity Resets to 3.5% of Revenue

The most quietly important disclosure of the call came in response to a question about where capital spending settles over three to five years. UPS spent $3.7B in 2025 and $3.9B in 2024 against a historical range closer to $4.5B to $5.0B, and now guides to about $3.0B.

"About 3.5% of revenue, Jeff, I think it's about the right level to think about."
— Brian Dykes, CFO

On approximately $91.2B of revenue that is roughly $3.2B, and the CEO framed it as structurally lower because the network build is complete rather than deferred, while naming two recent growth investments: 27 temperature-controlled cross-dock facilities and $50M into North American air freight for automotive and industrial customers.

Assessment: A 3.5%-of-revenue capital intensity is roughly $3.2B, close to the ~$3.0B already embedded in the guided $5.5B of free cash flow, so the payout ratio falls only if capex runs below that or profit recovers. It is the single most important number for anyone underwriting the dividend beyond 2026, and it received one sentence at the end of the call. The caveat is that maintenance capital on an automated network is not yet proven at the lower level, and depreciation is still rising, so the reported earnings benefit lags the cash benefit.

11. The 2028 Teamsters Contract Enters the Modelling Window

The current labour agreement expires in 2028, which puts the negotiation inside a two-year forecast horizon for the first time. Asked how investors get comfortable that the next round does not repeat the 2023 disruption, management answered on customer relationships rather than on labour cost.

"the contract renewal is 2 years out. We are today renewing contracts with our customers that go long past 2028."
— Carol Tomé, CEO

Assessment: This is the second consecutive quarter in which a direct question about the labour step-up was answered with something other than the labour step-up. In April the answer was about headcount when the question was about rate; this quarter it is about customer contracts when the question was about negotiation risk. The contractual wage escalator applies to the roughly 78,000-position-lighter workforce that remains, and it has never been quantified in dollars or in cost per piece. It remains the largest known, dated, unquantified item in the model, and it is now close enough to matter to a terminal-value assumption rather than only to a headline risk.

Guidance & Outlook

The full-year guidance changes are laid out in the table under Key Topic 4 above. This section covers the segment ranges, the implied ramp, and where the guide sits against consensus.

Segment Guidance

Management also guided the second half of the year for U.S. Domestic specifically: an operating margin of approximately 8.8%, with year-over-year expansion in both the third and fourth quarters. It gave no consolidated third-quarter revenue or EPS number, which is the second consecutive quarter in which the consolidated forward guide has to be assembled by the reader from segment ranges.

Implied Ramp

This is the arithmetic that should drive positioning, and it moved decisively in the company's favour this quarter. The full-year guide of approximately $91.2B and approximately $8.65B of adjusted operating profit, less the $44,036M and $3,424M delivered in the first half, leaves $47,164M of revenue and $5,226M of adjusted operating profit for the second half. The comparable second half of 2025 produced $5,022M on $45,894M. So the guide now requires 2.8% revenue growth, 4.1% adjusted operating profit growth, and 14bp of margin expansion over six months.

Compare that with where the same calculation stood three months ago. In April the reaffirmed guide required the remaining nine months to grow adjusted operating profit 5.7% and expand margin 36bp, immediately after a quarter in which both had fallen about 25%. Today it requires the remaining six months to grow adjusted operating profit 4.1% and expand margin 14bp, immediately after a quarter in which adjusted operating profit grew 12.0%. The ask got smaller and the evidence got larger in the same print. That combination, more than the raise itself, is what changes the rating.

The EPS line looks harder than the operating line and is not. Management flagged the reason directly.

"we had anticipated an earnings decline in the first half and an earnings increase in the second half when you normalize for the sale-leaseback transactions in the third quarter of last year, now in mid-teens EPS growth in both the third quarter and the fourth quarter."
— Brian Dykes, CFO

Third-quarter 2025 diluted EPS included a $0.30 benefit from sale-leaseback transactions. Normalized for it, the second half of 2025 earned $3.82 and the implied second half of 2026 is $4.40, which is 15.2% growth and matches management's "mid-teens" framing. The free cash flow line is the one that requires no heroics at all: $3,927M against $4,728M delivered in the same period last year.

Street at: the raised guide sits above consensus on both lines. Published full-year consensus entering the print was approximately $90.38B of revenue and $7.13 of adjusted EPS, against the new guide of approximately $91.2B and approximately $7.22. The gap is small on EPS and mostly fuel on revenue, so the raise is best read as a modest positive on the top line and a genuine positive on the commitment weight, since the EPS number moved from a spoken remark to the filed release.

Guidance style: the same conservatism as April, now with a quarter of delivery behind it. Management raised, but raised operating profit by $39M on $1.5B of extra revenue and framed the second half as "normal seasonality" three separate times under questioning. The pattern across two quarters is a team that banks its cushion rather than distributing it, which is worth roughly the difference between the guide and where the first half implies the year could land. Our estimate treats that cushion as real on the domestic line and does not extend it to International, where the full-year target looks stretched on its own arithmetic.

Analyst Q&A Highlights

Whether the Domestic Margin Reset Is Structural, and Where It Goes

The first question of the call asked for the thing management has never given: a longer-term domestic margin target. The answer supplied a framework instead, and the framework is the same 50 to 100bp spread commitment made three months ago, now restated as the permanent operating algorithm rather than as a year-end target to be reached.

Q: "can you maybe go into a little bit more on your confidence level around the structural change on domestic margin? And what that could mean from a longer-term domestic margin perspective? Is there a bogey you'd like to see over the next couple of years?"
— Jordan Alliger, Goldman Sachs

A: "So Jordan, I think the way to think about it is about 50 basis -- 50 to 100 basis point spread between RPP and CPP will help us drive margin accretion in our U.S. business as we go forward."
— Brian Dykes, CFO

Assessment: The refusal to name a multi-year margin bogey is the right call from a management team that has been held to one before, and the spread framework is a better disclosure anyway because it is checkable every ninety days. What it quietly concedes is that the second quarter's positive 130bp is above the algorithm rather than the new normal. A business running a steady 50 to 100bp spread compounds margin slowly, not in 400bp sequential steps, and that is the correct base case to carry.

Why Domestic Margin Steps Down in the Third Quarter Before Reaccelerating

The dominant line of questioning on the call, raised in three separate exchanges, was why the sequential path dips to approximately 7.0% in the third quarter when the transitional costs are gone, the Driver Choice departures are complete and the network is rightsized. Management answered the same way each time and the answer never moved beyond seasonality and the lapping of prior-year pricing.

Q: "So it seems like domestic margin improvement moderates a little bit in Q3, then reaccelerates again in Q4. Maybe just give a little more color to sort of talk about that?"
— Multiple analysts incl. Scott Group, Wolfe Research; Brian Ossenbeck, JP Morgan

A: "So first, Scott, if you think about the sequentials between Q2 to Q3, it's really normal seasonality, right? So we're still in that point -- 50 to 100 basis point RPP to CPP spread. It is consistent with where we set out at the beginning of the year."
— Brian Dykes, CFO

Assessment: The answer is internally consistent and does not fully satisfy the question. If the second quarter carried a positive 130bp spread and the third quarter returns to the 50 to 100bp algorithm, the sequential margin decline follows arithmetically, which is what management is saying. What it does not explain is why the completed removal of a large fixed-cost block does not push the spread above the algorithm rather than back to it. The most likely reconciliation is that fuel flattered the second-quarter spread and the third quarter is closer to the underlying rate, which would make 7.0% the honest run-rate and 8.0% the flattered print. That reading is more conservative than management's and is the one embedded in our estimate.

Where Revenue Per Piece and Cost Per Piece Go in the Second Half and into 2027

The most useful forward disclosure of the call came in response to a question about the cost-per-piece trajectory and how much product mix adds on top of base pricing. Management gave a specific deceleration path for both lines rather than a direction.

Q: "So you've had strong execution on resizing the network. Just want to get your sense on kind of how you think that plays out. And then how should we think about mix as a factor when we look at, let's say, 2027?"
— Thomas Wadewitz, UBS

A: "So we were at 9.3% in second quarter for RPP in the U.S., about half of that was fuel and about half of it was base pricing and mix. That will come down closer to 4% -- 4% to 4.5% as we wrap some of those impacts from last year. On a 2-year stack, it still shows really strong pricing. The same dynamic happens in CPP, right? So as we brought down the cost of the network, we expect CPP to come down as well. So we'll maintain that kind of 50 to 100 basis point spread."
— Brian Dykes, CFO

Assessment: This is the answer to build a model on, and it is materially more specific than anything given in April. Revenue per piece decelerating to 4% to 4.5% with cost per piece decelerating alongside and the spread held at 50 to 100bp is a coherent, modellable second half, and it is roughly consistent with the third-quarter margin guide. It also frames 2027: base pricing of 250 to 350bp plus 50 to 100bp of product mix, against a cost line the company believes it now controls. The unaddressed variable in that 2027 sketch is the labour rate escalator, which appears in neither the pricing nor the cost half of the answer.

The Amazon Disclosure After the Glide-Down Completed

With the glide-down finished, a direct request was made for the Amazon revenue dollar figure the company had partially disclosed a quarter earlier. The answer repeated the percentage and reframed the relationship as a partnership without adding a number, a floor or any pricing colour.

Q: "In the last quarter, you gave the Amazon revenue. I know it wasn't a year-end number. But I was wondering if you could maybe offer that considering we're kind of at the end of the glide down."
— Christian Wetherbee, Wells Fargo

A: "as Amazon is and will continue to be an important customer to UPS as a percent of our total revenue, Amazon made up 9%. And that's down about 100 basis points from a year ago and certainly down from the which was over 13% during the COVID year. As we think about going forward, it's all about optimizing the volume that comes into our network, and we are working together as partners to make that happen."
— Carol Tomé, CEO

Assessment: The question was for a dollar and the answer was a percentage, for the second consecutive quarter and now at the moment the transition is complete. Nine percent of approximately $91.2B is roughly $8B, so this is not a rounding item. Management has told the market that Amazon revenue will grow every quarter this year while declining to size the base it grows from, which makes the claim unfalsifiable. We continue to treat the retained book as flat and would read a first dollar disclosure as a signal of management's own confidence in it.

Whether International Margin Returns to the Upper Teens

The segment that missed its guide drew the most forward-looking question of the call, framed as a range rather than a number so that management could pick. Management named the mechanisms and did not pick a level.

Q: "And if international is normalizing now with Asia volume, should we see the margins return to upper teens, mid-teens -- where do you think international pans out as well?"
— Ken Hoexter, Bank of America

A: "Look, we're seeing volume recover as trade lanes settle Also, in the back half of this year, we have the wrap of the elimination of global de minimis in September that drives a pretty big year-over-year comp benefit. I would also say, look, our international business is going to perform a little bit better than seasonality because of the volume momentum we have going into the fourth quarter, especially with surge fees and fees that we typically see in the market."
— Brian Dykes, CFO

Assessment: An answer built entirely from mechanical comparisons rather than from underlying margin. The de minimis lap and the fourth-quarter surge fees are both real and both temporary, and neither speaks to why the segment gave up 160bp beyond the fuel explanation. Offered the words "upper teens" and "mid-teens," management took neither. On the segment's own first-half margin of 12.2% and the guided third quarter, the honest answer is low-to-mid teens for the year, and the absence of a direct answer to a directly posed question is the strongest signal in the exchange.

Whether the Second-Half Step-Up Exceeds Normal Seasonality

The sharpest analytical challenge of the call granted management its own first-quarter adjustment and then showed the implied second-half step-up is still larger than the company has historically produced. The answer located the entire excess in a single segment.

Q: "But even if we add back the $350 million in onetime costs that you talked about in the first quarter, it still looks like the consolidated operating profit in the second half versus first half is a bit more than you've done in recent years. Could you just broad strokes high level kind of [indiscernible] to what we're missing on the cost side or the revenue side that gets you comfortable with that maybe slightly more than typical momentum there?"
— Bascome Majors, Stephens

A: "So Bascome, if you look across the segments, Domestic and SES are kind of following oral seasonality. Where you're seeing the improvement is in the international business."
— Brian Dykes, CFO

Assessment: The most consequential exchange on the call, and it points the risk squarely at the segment that just missed. Management has conceded that the domestic and Supply Chain trajectories are ordinary seasonality and that the above-seasonal portion of the second-half guide is International. That is precisely the segment whose full-year target the arithmetic will only support at its floor and only with a record fourth quarter. The domestic story is now the safe half of the guide and International is the risk half, which is an exact inversion of how this thesis was framed in April.

Whether Real Estate Gains Flattered the Quarter

Given how large the sequential margin step was, a direct question tested whether property transactions had contributed to it. The answer was a qualified no, and it also drew the only clean disclosure of the call about the size of the offsetting write-offs inside the Amazon plan.

Q: "can you confirm that whether you had any real estate gains in 2Q and either run rate for the rest of the year."
— Ravi Shanker, Morgan Stanley

A: "Ravi, thanks for the question. On the real estate gain, so we didn't have anything that was out of ordinary course in the quarter. Look, as part of the Amazon drawdown plan, right, that we included in our guide that we've been executing over the last couple of years, there's buy and sale of real estate."
— Brian Dykes, CFO

Assessment: This is the exchange that makes the quarter clean, and it deserves more weight than it received. The single most common way a parcel network manufactures a margin inflection is by selling buildings into it, and management ruled it out on the record while confirming that property gains and asset write-offs inside the reconfiguration net to approximately zero. Combined with the contribution-per-piece reconciliation, there is now no plausible non-operational explanation for the domestic profit increase.

Where Capital Intensity Settles

The last substantive question of the call was the only one about the long run, and it produced the number that matters most for the dividend beyond this year.

Q: "I know we're spending about $3 billion right now, and that's been the case for the last year too. But if I go back time, spend in the $4.5 billion to $5 billion range has kind of been more normal for the business. If we look out 3 to 5 years, where do you think capital spend should settle?"
— Jeffrey Kauffman, Citizens Bank

A: "About 3.5% of revenue, Jeff, I think it's about the right level to think about."
— Brian Dykes, CFO

Assessment: A one-sentence answer with more valuation content than most of the call. Historic spend near 5% of revenue versus a stated forward 3.5% is worth roughly $1.3B of annual free cash flow at the current revenue base, which is the entire difference between a dividend that consumes almost all of free cash flow and one that leaves room for repurchases. It was volunteered in response to the fifteenth question rather than in the prepared remarks, and it is the least-priced disclosure of the quarter.

What They're NOT Saying

1. The Amazon revenue dollar, now that the glide-down is complete: asked directly for the figure at the moment it became a final number rather than a moving one, management repeated the 9% share. At approximately $91.2B of guided revenue that is roughly $8B, and it comes with a promise that it will grow every quarter and no base from which to verify the promise. Second consecutive quarter the same question got the same non-answer.
2. Why International lost 160bp beyond fuel: the 280bp margin decline was explained as 120bp of fuel and then a list of qualitative items, with no decomposition of the remainder into mix, rate or cost. The segment carrying the above-seasonal portion of the second-half guide is the segment whose margin shortfall was left undiagnosed.
3. That the full-year domestic margin target was lowered: April framed the full year as flat with 2025's 7.74%. Management said approximately 7.5% on the call; the release does not carry it. That is a 25bp reduction in the segment that is 65% of revenue, disclosed only as a number in the outlook commentary with no reference to the prior framing.
4. Share repurchases, for the second consecutive quarter of silence: UPS bought back no stock in either quarter of 2026 against $1,000M in the first half of 2025, and the word did not appear in the release or on the call. The full-year plan of $3.0B of capex, $1.3B of pension and $5.4B of dividends against $5.5B of free cash flow explains why. The silence is the disclosure.
5. Any consolidated third-quarter guide: three segment ranges and no consolidated revenue, margin or EPS figure, for the second consecutive quarter. Every reader has to construct the number from segment weights, which is a step down in specificity from a company whose second half carries most of its year.
6. EPS and free cash flow guidance remain partly unfiled: the adjusted EPS guide made it into the written release this quarter, which is an improvement. The approximately $5.5B free cash flow figure still exists only in spoken remarks, as does the $1.3B pension contribution and the 3.5%-of-revenue capital intensity framing.
7. The 2028 labour rate step-up, for the second consecutive quarter: asked how investors get comfortable with the negotiation, management answered with customer contract durations. The wage escalator applies to the workforce that remains after roughly 78,000 operational positions came out, and it has never been quantified in dollars or in cost per piece by a company that has now put a 2027 operating-leverage framework on the record.
8. The economics of the USPS Ground Saver arrangement: purchased transportation rose 26.4% to $3,187M, and the USPS tender sits inside it at full scale. No per-piece cost, no contract term, and no disclosure of what portion of that 26.4% is the tender versus leased aircraft.
9. What the non-Amazon volume base is actually doing: management said U.S. volume grew year over year after adjusting for Amazon and other de-selected volume, and that back-half volume growth ex-Amazon will appear across all segments. Neither claim came with the adjusted figure, and the reported third-quarter guide of a mid-single-digit decline is a faster fall than the second quarter's 3.3%.
10. Leverage, ratings and the maturity profile: total debt and finance leases of $24,484M against $15,099M of equity, with interest expense up 17.0% in the half, appear only as balance sheet lines. Neither the release narrative nor the call gave a leverage ratio, rating or maturity profile, in a year when the dividend absorbs 98% of guided free cash flow.

Market Reaction

• Pre-print setup: UPS closed at $112.95 on July 27, up 13.9% year to date against the S&P 500's 8.3%, up 11.2% over trailing twelve months, and up 4.4% over the trailing thirty days. The 52-week closing range entering the print was $82.58 to $120.00, so the stock came into the event at roughly the 81st percentile of its own year, having recovered the whole of the April sell-off and more.
• Reaction session (before-open reporter, same-day reaction): opened at $108.60, down 3.9% on the gap, traded a range of $103.09 to $109.20, and closed at $105.53, down 6.6% or $7.42 on the day. The S&P 500 rose 0.2% in the same session, so essentially the entire move was stock-specific.
• Volume: 14.3M shares against a 4.8M thirty-day average, 3.0 times normal, and heavier than the 2.2 times that accompanied the April print.
• Intraday shape: unlike April, the stock did not recover. The low of $103.09 came 8.7% below the prior close and the close of $105.53 sat only 2.4% above it, in the lower half of the day's range and 2.8% below the opening print. The selling was persistent rather than front-loaded.

The move was not a verdict on the operating result. It was a collision between a print that beat on the number management manages to and a consensus base that was anchored to a different number. Published GAAP consensus stood near $1.64 and the company printed $0.71, a shortfall of more than half, because the $891M after-tax Driver Choice charge was not in the models. Screens, wire headlines and the first thirty minutes of trading all read GAAP. By the time the call clarified that the charge was the final bill for a program that completed inside the quarter, the gap had opened and did not close.

The second driver was the third-quarter guide, and it is the one with genuine content. A sequential domestic margin step from 8.0% to approximately 7.0%, on approximately flat segment revenue and U.S. volume guided down mid-single digits, is a worse near-term shape than a stock up 13.9% year to date was positioned for. The market had spent three months pricing the inflection; the company delivered the inflection and then guided the next quarter below it. That the third-quarter figure is still 64bp of year-over-year expansion is the kind of point that gets made on the second day, not the first.

The absence of a recovery off the low is the difference from April and the more informative half of the session. In April a momentum unwind met a bid at a 6.3% yield and the stock closed where it opened. Here the stock closed near the bottom of its range on three times normal volume, which suggests the seller was a holder repositioning around the forward guide rather than a fast-money exit. That is a more durable kind of selling and it is also, for a buyer, a more attractive entry, because it prices the near-term shape rather than the terminal economics.

Street Perspective

Debate: Is the Domestic Margin Reset Structural or a Cost Cut a Shrinking Volume Base Will Erode?

Bull view: The bull case being made on the Street is that this quarter settles it. A 21% increase in segment operating profit on 3.3% fewer pieces cannot come from volume, seasonality or property gains, all three of which are ruled out by the disclosure or by management on the record. Capacity came out faster than volume: roughly 78,000 operational positions, 150 buildings and 50 million hours across 2025 and 2026, with automation now handling 68.5% of U.S. volume at a 28% cost-per-piece advantage. Contribution per piece rose 26.4%, which is the definition of operating leverage arriving.

Bear view: The bear camp contends that a network that has removed 2 million daily pieces and 78,000 positions has spent its ammunition, and what remains is a business whose volume is still falling and is guided to fall faster next quarter. Cost per piece still grew 8.0%. The margin algorithm management put on the record is a 50 to 100bp spread, which compounds margin at a few basis points a quarter, not in 400bp steps. Strip fuel and the second quarter's outperformance against that algorithm shrinks toward nothing.

Our take: The bulls have the better read and the bears have the better forward number. The reset is real: the contribution-per-piece arithmetic reconciles to the reported profit change within $8M and there is no non-operational explanation left standing. But the correct run-rate is the 50 to 100bp algorithm, not the 130bp print, and management's own third-quarter guide says so. Underwrite a business that has permanently repriced its unit economics and then compounds them slowly, which is a materially better business than the one priced into the stock in April and a materially worse one than a 400bp sequential step implies.

Debate: Was the Sell-Off a GAAP Optics Problem or a Real Downgrade of the Second Half?

Bull view: A growing consensus view holds that the session was mispriced mechanics. The GAAP shortfall was a $891M separation charge for people who have already left, disclosed in April, fully absorbed in the guided free cash flow, and with only $73M to $273M of excluded cost left for the entire second half. Meanwhile the company raised revenue, operating profit and EPS guidance above consensus and cut the second-half ask from 5.7% profit growth over nine months to 4.1% over six. A 6.6% decline on that combination is a gift.

Bear view: The bear argument is that the guide, not the charge, did the damage and did it correctly. Third-quarter domestic margin steps down to approximately 7.0%, U.S. volume is guided to decline mid-single digits in the first quarter after the Amazon exit finished, and the full-year domestic margin target was quietly cut from flat-with-2025 to approximately 7.5%. The revenue raise is fuel with a 2.6% incremental margin and the implied full-year adjusted operating margin actually fell.

Our take: Both effects were present and they were not equally weighted. The first two hours of the session were GAAP mechanics; the failure to recover was the guide. The bears are right that the third-quarter shape is worse than positioning assumed and right that the domestic full-year target was trimmed without commentary. They are wrong that this makes the year harder, because the same guide requires less of the second half than the April version did. The market repriced a quarter and left the year unchanged, which is the kind of dislocation that resolves in favour of the fundamentals.

Debate: Is International a Temporary Fuel and Geopolitics Problem or a Structural De-Rating?

Bull view: The bull framing is that every element of the shortfall is dated. Fuel is 120bp of the 280bp decline and reverses when crude does. The Middle East reroute costs disappear when airspace reopens. The China-to-U.S. lane returned to growth in May, Asia-to-Asia grew 13.6%, and the September lap of the global de minimis elimination is a mechanical fourth-quarter comparison benefit. The segment earned 15.80% for the whole of 2025 with worse trade conditions than it faces now.

Bear view: The bear view is arithmetic. The first half delivered 12.2% and the third quarter is guided to 13% to 14%, which means a mid-teens full year needs a fourth quarter between 17.8% and 21.7% depending on where inside mid-teens you land. Only the bottom of that band is achievable and only by matching a level the segment last printed in 2024. Management, offered "upper teens" and "mid-teens" in a direct question, declined to take either. And this is the segment the CFO identified as carrying the entire above-seasonal portion of the second-half guide.

Our take: The bears have this one and it is the reason our full-year estimate sits below the guide. Some recovery is close to certain because the de minimis lap is mechanical, but the gap between 14% and 15% is roughly $195M of operating profit, and the full-year target as worded implies the upper end. We model approximately 14%, treat any print above 15% as upside to our number, and regard a second consecutive quarter below the guided range as the event that would send this thesis back to Hold.

Debate: Does a 6.2% Yield With No Buyback Signal Value or Signal Constraint?

Bull view: The bull argument has strengthened this quarter. Six-month free cash flow more than doubled to $1,573M, the second half needs $3,927M against $4,728M produced in the same period last year, and the guided figure already absorbs the whole of the Driver Choice cash. Underneath that, management put forward capital intensity at 3.5% of revenue against a historical 5%, which holds capex near the level already embedded in the free-cash-flow guide.

Bear view: The bear argument is unchanged and unanswered. A $5.4B dividend against $5.5B of guided free cash flow is 98% of everything the business generates. Repurchases have been zero for two consecutive quarters and unmentioned for two consecutive calls. Cash fell $1,234M in the half, retained earnings fell $1,321M, and interest expense on $24,484M of debt is compounding at a high-teens rate. A yield that high in a mega-cap industrial usually prices a probability.

Our take: The yield is compensation for tightness that is now visibly loosening rather than for a risk that is building. The dividend is not at risk in 2026 and the second-half cash requirement is below what the business produced last year. The capital-intensity disclosure is the piece the market has not priced, because it converts a 98% payout into something closer to 80% at an unchanged dividend within two years. We would still not own this for the yield alone, but the yield is no longer the reason to stay away.

Model Update & Valuation Framework

The table below sets our revised estimates against both the company's raised guidance and the estimates we published with the first-quarter recap.

Valuation. At the July 28 close of $105.53, UPS trades at 14.6 times the company's raised full-year adjusted EPS guide of $7.22 and 14.9 times our $7.10, with a 6.2% dividend yield on the declared $1.64 quarterly rate. We are moving our multiple range from 14 to 16 times to 15 to 17 times, which on $7.10 produces a $107 to $121 fair-value range with a midpoint near $114, about 7.6% above the current price. Add the dividend and the twelve-month total return we underwrite is roughly 14%.

Why the multiple moves. In April we used 14 to 16 times explicitly because the margin recovery had not produced a single quarter of evidence and because the low end of the range reflected a business with declining volumes, a fully committed dividend and rising interest expense. Three of those four conditions are unchanged. The one that changed is the decisive one: the recovery has now produced a quarter, and it produced it in the form that is hardest to fake. A full point of multiple is what that evidence is worth. We are deliberately not going above 17 times, because volume is still declining, is guided to decline faster next quarter, and the dividend still absorbs the whole of guided free cash flow.

What would change the rating. Back to Hold: a third-quarter U.S. Domestic operating margin below the guided 7.0%, a second consecutive International quarter below its own guided range, or any evidence that the 50 to 100bp spread algorithm is not holding once the fuel distortion washes out. To Underperform: a fourth-quarter domestic margin materially below the approximately 10.3% the second-half guide implies, or any softening of the $5.4B dividend commitment. To add to the position: a first disclosure of the Amazon revenue dollar, a repurchase authorization funded by the lower capital intensity, or a fourth quarter that puts International back inside mid-teens on the full year.

Thesis Scorecard Post-Earnings

The table grades the seven pillars established when we initiated coverage in April, in the same order and against the same wording.

Overall: Thesis strengthened. Five of seven pillars moved in the company's favour, including both of the bear points that carried the most weight in the April rating. The single pillar that moved against is Bull 4, and it did so by inverting rather than breaking: the half of it that was supposed to be the safe half is now the risk half. That is a meaningful change in where the analytical attention goes and a smaller change in the value of the business, because U.S. Domestic is 65% of revenue and International is 22%.

Action: Upgrade to Outperform. The quarter we said in April was the entire thesis delivered the test at its midpoint, the cost-per-piece reversal we called the most consequential unsupported commitment on the prior call arrived two quarters early and above its committed range, and the second-half ask fell in the same print. The stock de-rated 6.6% into that evidence on a GAAP charge that was disclosed three months ago and is now nearly finished. At 14.6 times a raised guide with a 6.2% yield and a fair-value midpoint near $114, the risk/reward has moved. Size it for a business whose volume is still falling and whose International segment just missed, and revisit on the third-quarter print, where the 7.0% domestic margin guide and a second International quarter are the two things that matter.