Uber reports second-quarter results on August 5, 2026. Our preview of that print tabulates the last reported quarter — revenue, gross bookings, GAAP and non-GAAP EPS, adjusted EBITDA — and carries no margin line at all. Its only mention of margins is forward-looking: whether advertising growth holds is "a real swing factor for margin expansion."
Margin expansion is not a thing to watch for here. It already appeared, in the exact quarter that preview tabulates, and it is the largest single move in this series on record.
It is also, mostly, not what it looks like.
The band, and the break
Every gross margin below is revenue less cost of revenue exclusive of depreciation and amortisation — the line Uber actually files. Say that once and it stays true for the whole piece; a figure computed on a cost line that includes D&A will not match.
For eleven consecutive quarters, from the quarter ended 2023-06-30 through 2025-12-31, Uber's gross margin sat between 39.04% and 40.25% — a band 1.21 points wide. Then:
| quarter ended | revenue | cost of revenue (ex-D&A) | gross margin |
|---|---|---|---|
| 2025-03-31 | $11,533M | $6,937M | 39.85% |
| 2025-06-30 | $12,651M | $7,611M | 39.84% |
| 2025-09-30 | $13,467M | $8,109M | 39.79% |
| 2025-12-31 | $14,366M | $8,681M | 39.57% |
| 2026-03-31 | $13,203M | $7,258M | 45.03% |
That is +5.46 points against the prior quarter and +4.78 against the highest reading of the preceding eleven. The move is roughly four and a half times the entire width of the band it broke.
Uber states the same thing in its own terms: cost of revenue went from 60% of revenue to 55%.
Year over year, revenue rose 14.5% and cost of revenue rose 4.6%. That gap is the whole arithmetic — and the reason for it is in the filing.
What the 10-Q says
This is the part that decides whether the number means anything, and it is not a matter of inference. From Uber's 10-Q for the quarter ended 2026-03-31:
Effective January 2, 2026, we implemented a business model change in certain UK markets, primarily driven by regulatory and tax considerations. As a result of this business model change, we are no longer responsible for the Mobility services in these markets, and accordingly, payments to drivers are recorded as a reduction of revenue instead of cost of revenue.
And, on the revenue line:
The increase in revenue was partially offset by Mobility business model changes in the UK that negatively impacted revenue by $1.0 billion.
So the same payments that used to sit in cost of revenue are now netted against revenue. Both the numerator and the denominator of the margin shrink by the same dollars — which mechanically raises the percentage without anything about the business improving.
The date matters: January 2, 2026 is the first business day of the quarter that broke the band. The break and the change are not merely contemporaneous; the change covers the quarter exactly.
Putting the two presentations back on one basis
Uber gives the revenue impact as $1.0 billion. On the filing's own description — those payments were formerly recorded as cost of revenue — the pre-change presentation of this quarter would carry that $1.0B in both lines:
| quarter ended 2026-03-31 | revenue | cost of revenue | gross margin |
|---|---|---|---|
| as reported | $13,203M | $7,258M | 45.03% |
| reconstructed on the prior basis | $14,203M | $8,258M | 41.86% |
That reconstruction is ours, not Uber's, and it rests on one assumption worth stating: that the netted driver payments equal the $1.0B revenue impact the company discloses. Uber does not publish the cost-side figure for those same markets separately — the itemisation in the MD&A gives a $606M decrease in driver payments within cost of revenue, but that figure is net of growth elsewhere, "despite an increase in Mobility Gross Bookings in certain markets."
On that reconstruction, about 3.2 of the 5.5 points are presentation and the remaining 2.3 are real improvement over the prior quarter. Measured against the band instead of against the prior quarter, the reconstructed 41.86% still sits 1.6 points above its 40.25% ceiling — a different baseline, which is why those two figures do not add to 5.5. Either way the residual is not nothing, but it is an ordinary quarter's worth of movement rather than a step change.
What actually moved in the cost line
The MD&A names the largest movements within that $321M increase — "primarily attributable to", in its own words — and they are worth reading because they show a business whose costs are growing normally:
- +$631M courier payments and incentives, on higher Delivery gross bookings
- +$116M credit card processing costs
- +$85M insurance expense, on more miles driven
- +$81M carrier payments and incentives, on higher Freight gross bookings
- +$78M network costs
- −$606M driver payments and incentives, from the UK business model change
- −$165M legal, tax and regulatory reserve changes and settlements
- +$101M everything else, which the MD&A does not itemise
Those eight add to the $321M the line actually moved; the seven Uber names add to $220M, and the remainder is the balancing figure rather than a disclosed one.
Every increase there is volume-driven. The two decreases are the model change and a reserve movement. There is no line in that list that reads as structural cost efficiency.
Two things this is therefore not. It is not advertising: a $2 billion annualised ad run-rate is roughly $500M a quarter against a $1,423M sequential fall in cost of revenue, and Uber does not attribute any of the move to ads. And the sequential revenue fall of 8.1% is not a business in decline — Uber's first quarter is seasonally its weakest, and gross bookings rose 25% year over year.
What August 5 settles
One quarter cannot distinguish a durable margin from a presentation artefact. The second quarter under the new UK treatment can, because the change will be in both the current and prior periods by the time the comparison is drawn.
- A gross margin near 45% again would confirm the new presentation is the run rate — the level is real, even though the step to it was mostly accounting.
- Anything near 42% would suggest the reconstruction above is the better read of the underlying business.
- The number to watch alongside it is whether the UK change is still being called out. When the company stops quantifying it, the year-over-year comparison becomes clean and the margin series becomes readable again — from the quarter ending 2026-12-31 onward.
Our own coverage should have caught this before the print rather than after. The preview tabulated the quarter in which the largest margin move in Uber's recorded series occurred, and discussed margins only in the future tense.
Revenue and cost of revenue are from SEC XBRL, CIK 0001543151 — revenue filed under Revenues and cost under CostOfGoodsAndServiceExcludingDepreciationDepletionAndAmortization, which excludes depreciation and amortisation. The five quarterly readings reproduce our stored Uber data to the second decimal, and the eleven-quarter band is that stored series. The quotations, the $1.0 billion revenue impact, the 60%-to-55% cost ratio and the itemised cost movements are from Uber's 10-Q for the quarter ended 2026-03-31, accession 0001543151-26-000022. The reconstructed 41.86% is our derivation on the assumption stated above, not a figure Uber publishes. No price, valuation, rating or price-target figure appears here.