GitLab reported the quarter ended 31 July 2026 after the US close on 1 September. Revenue was $286.254 million, 4.9% above the midpoint of the company's own $272–274 million guide and 4.8% above the roughly $273.1 million consensus; non-GAAP diluted earnings were $0.24 against $0.18 expected. The shares closed the session at $45.09 and traded at $52.97 at 7:59 p.m. Eastern, up 17.5% on the print.
Our preview argued that the guide was the story: a seven-point deceleration in one step, with the operating margin guided down in the quarter GitLab removed a seventh of its workforce. The quarter answered that in the most awkward way available. It beat the guide by $13.3 million, converted 14.9% rather than the guided 10.9–11.8% — and then GitLab raised the full year by $16.0 million, of which $2.7 million lands in the half that has not happened yet.
That subtraction is the piece. Everything else the print did, it did in one quarter.
What the preview asked, and what the print answered
Revenue against $272–274 million, and the full-year range — beat by $13.3 million, and the raise is that beat. Fiscal 2027 guidance went from $1,112–1,118 million to $1,129–1,133 million, a $16.0 million raise at the midpoint. First-half revenue is now $550.412 million against the $537.158 million the old guide implied, so the beat alone is $13.3 million of it. Take the new midpoint less the reported half and the implied second half is $580.6 million against $577.8 million before — up 0.5%, or +15.0% year over year against +14.5%. That subtraction is ours, on GitLab's own two guides. The second half was raised by less than a fifth of a point of growth.
Non-GAAP operating margin against the guided 10.9–11.8% — 14.9%, and nobody had to explain the guide. Non-GAAP operating income was $42.566 million against a guided $30–32 million, 37% above the midpoint. The margin the preview called the least-explained number in the guide simply did not appear; what appeared instead was a margin down year over year, 16.77% to 14.87%, on 21.3% revenue growth. And the third quarter is guided straight back to it: $35–37 million on $281–283 million is 12.4% to 13.2%.
Customers above $5,000 of ARR against 10,831 and +7%, and net retention against 117% — the seven-quarter slide stopped. The count rose to 11,114, +8% year over year, the first acceleration in the series, and the dollar-based net retention rate held at 117% after 121% a year earlier. Held, not improved: the company reports it to the whole percent, so the CFO's "dollar-based net retention accelerated sequentially for the first time since 2024" is not checkable from the disclosure. The bear case in our model extrapolates both of those lines downward, and this quarter is the first that does not feed it.
Current RPO against $724.1 million and +24% — $744.7 million and +20%, and the metric now has a hole in it. Growth decelerated four points, but not for the reason a decelerating book would. Flex, the new consumption contract, is a committed dollar amount that sits in total RPO and is excluded from current RPO entirely: "This quarter, CRPO saw a three-point headwind relative to RPO purely from the absence of Flex commitments in that metric." Calculated billings, which carry no such exclusion, grew 24%, double the 12% of the April quarter. Two forward indicators moved in opposite directions and the one that fell is the one with the exclusion in it.
Restructuring inside $30–35 million with about $19 million this quarter — $19.4 million of charges, $23.3 million on the wider definition, and the country exits did not show up in the customer counts. The release puts $19.422 million of restructuring charges through operating expenses and cost of revenue; on the call the CFO sized it as "approximately $23.3 million in restructuring charges, in line with what we outlined last quarter", the difference being $3.9 million of accelerated share-based compensation the release books separately. GitLab still gives no cumulative-to-date figure and no remaining estimate. The exit from 22 countries, which the preview flagged as the removal of the coverage that refills a funnel, is not visible in either customer bucket this quarter.
What actually changed: a second revenue model arrived, and it is not in the guide
Flex has been on sale for ten weeks and it changes three things at once, none of them in the press release. The company set them out in an investor letter posted only to its own site, and on the call.
Revenue recognition. About 15% of a traditional self-managed licence is recognised upfront; under Flex none of it is. The CFO's rule of thumb: "for every $50 million of our self-managed available to renew that converts to Flex in FY 2027, we estimate approximately $5 million of revenue that would otherwise have landed in FY 2027 instead shifts to be recognized over future periods," with a ceiling of about $13 million for the year. That is timing, not demand — bookings, billings, cash and total RPO are untouched. But it is not in the raised guide: "we have not incorporated the potential impact of Flex in our guide, but we do intend to quantify the revenue recognition impacts through the rest of the year." So the $1,129–1,133 million range has up to $13 million of deferral sitting on top of it, and a company that is switching its default contract for at least three fiscal years has just made its own headline metric partly an accounting artefact.
A new metric arrived with the quarter it flatters. Paid Consumption Run Rate exited the quarter above $40 million, from $15 million a quarter earlier, with a stated objective "to exceed $100 million of paid CRR by the end of this fiscal year". GitLab says plainly that it is not revenue, not ARR and reconciles to no quarter. Treat it as an adoption gauge; it is the only number attached to the consumption business at all.
And the gross margin kept going. GAAP gross margin fell to 84.06% from 87.92%, because subscription cost of revenue grew 75.9% against 21.5% subscription revenue growth — $21.8 million to $38.3 million — taking subscription gross margin from 89.77% to 85.19%. The obvious reading is inference cost, and the CEO rejected it: "for many of our customers, the token or inference cost is actually not embedded in the GitLab agreement… a lot of the margin changes that we've seen in the business have been driven more by the mix shift to SaaS than the early AI adoption." SaaS is about 34% of revenue and growing 36%, against 22% at the IPO. Both explanations were given the same night; only the mix one has a disclosed number attached.
One more thing worth reading twice, because the free cash flow on our quarter page is negative and GitLab's is not. Operating cash flow was −$3.092 million and capital spending $0.213 million, so cash flow after capital spending was −$3.305 million. GitLab's "adjusted free cash flow" of +$9.750 million adds back a $14.036 million non-recurring payment tied to the formation of its China joint venture and nets off a $0.981 million tax refund — a $13.055 million wedge that flips the sign. Both figures are in the same table of the same release. The negative one is the plain arithmetic; the positive one is the company's definition.
What the quarter does to the model
Our GitLab model was published on 31 August, one day before the print, against a $46.54 close. Its first projected quarter — this one — was $273.9 million, inside the guide. It came in 4.5% higher.
The model's value is overwhelmingly terminal: 77% of enterprise value sits in an exit at 4.5 times revenue five years out, and net cash is $8.04 a share. So carrying the whole path up by the quarter's 4.5% and changing nothing else lifts the base case from $44.57 to roughly $46.2. That solve is ours, arithmetic on the model's own disclosed exit assumption rather than a re-run of it; the published fair values stand until a revision, which is a separate decision.
The point is the size of it. A 4.5% beat, the largest in the model's short life, moves the base case by about a dollar and a half — and the shares moved $7.88 after hours. The model's Act 2 case, built on the restructuring working, is $46.56; the bear case, built on the seven-quarter deceleration simply continuing, is $29.26. What was bought after the close was neither of those. It was Flex and the 42% net ARR growth the CFO cited on the call, neither of which is a line in any statement GitLab files.
What is coming
The third quarter is guided below the second. $281–283 million against the $286.254 million just printed is a guided sequential decline of 1.1% to 1.8%, and +15.0% to +15.8% on the year-ago quarter. The implied fourth quarter is $295.6–301.6 million, +13.5% to +15.8%. Both are the full-year guide less what is now reported, and the subtraction is ours.
Three things settle between now and December. GitLab has committed to quantifying the Flex revenue-recognition impact each quarter for the rest of the year, which is when up to $13 million stops being a ceiling and becomes a number. Paid CRR against the $100 million objective is the only read on whether the consumption business is a product line or a metric. And JiHu expense of about $50 million for the year against $13 million last year, on a China entity the company still wants to deconsolidate and "cannot predict if or when that will happen", is a fourfold step-up inside a margin guide that already asks for less than this quarter delivered.
One item is not an earnings question at all: the same filing that carried the results disclosed that the Chief Accounting Officer resigns on 16 September, explicitly not over any disagreement on financial reporting, and the CFO absorbs the role. It is a year since the last CFO left.
What we learned
- The full-year raise is this quarter, not the second half. Guidance went up $16.0 million at the midpoint against a $13.3 million beat; the implied second half rose $2.7 million, from $577.8 million to $580.6 million, or 0.5%.
- The margin guided down in the quarter of the cuts came in at 14.9%. Non-GAAP operating income was $42.566 million against a guided $30–32 million — then the third quarter is guided back to 12.4–13.2%.
- Up to $13 million of this year's revenue will be deferred by Flex, and none of it is in the raised guide. The same contract is excluded from current RPO entirely, worth a three-point headwind to the 20% growth reported.
- Subscription cost of revenue grew 75.9% against 21.5% subscription revenue. Subscription gross margin fell 89.77% to 85.19%; the CEO attributes it to SaaS mix, not inference, since most customers pay their token costs outside the GitLab contract.
- The seven-quarter slide in the two base metrics stopped. Customers above $5,000 of ARR grew 8% against 7% in April, and net retention held at 117% — though the whole-percent disclosure cannot show the "sequential acceleration" management claimed.
GitLab reported the three months ended 31 July 2026 after the US close on 1 September 2026. Revenue and its two lines, the cost, operating expense, restructuring, share-based compensation, cash flow and balance sheet figures, the non-GAAP reconciliation, the customer counts, net retention, remaining performance obligations, and third-quarter and fiscal 2027 guidance are GitLab's own, from that release; the full figure set is on our GitLab Q2 FY2027 earnings page. Calculated billings growth, net ARR growth, Paid Consumption Run Rate, Flex commitments and the Flex revenue-recognition and current-RPO effects come from the investor letter GitLab published the same day and from the earnings call; Paid CRR and net ARR are company-defined measures that reconcile to no reported line. Ours rather than GitLab's: the implied second half and fourth quarter, every margin computed from a guided range, the year-over-year and sequential moves, and the base-case sensitivity, which is arithmetic on the assumptions in our GitLab model of 31 August 2026 and not a company forecast. The $0.18 and $273.1 million consensus is press-reported from third-party estimate feeds on a non-GAAP basis and is not a series this site verifies. Prices are the 1 September 2026 close of $45.09 and a $52.97 after-hours quote at 7:59 p.m. Eastern that day.
