Snowflake Delivered the Back-Half Acceleration in Q1 and 442bp of Operating Margin With It, But the Raise Was Only the Beat

Results vs. Consensus

Q1 FY2026 Scorecard

Product revenue and EPS consensus are the sell-side averages compiled ahead of the print; total revenue consensus is the average of two independent providers, which agreed within 0.4%. Non-GAAP EPS consensus was quoted at $0.21 by two providers and $0.22 by two others; we use $0.21 as the most widely cited. Snowflake does not guide dollar operating income, so the operating-margin comparison is against the full-year margin guide rather than a quarterly figure. All actuals trace to the Exhibit 99.1 income statement and GAAP-to-non-GAAP reconciliation, not to figures quoted verbally on the call.

Year-Over-Year Comparison

Sequential Comparison

Q4 FY25 comparatives are Snowflake's reported figures for the quarter ended January 31, 2025. Product revenue per day divides reported product revenue by the 89 days in the quarter ended April 30, 2025 and the 92 days in the quarter ended January 31, 2025; it is our calculation and is shown because the sequential optic is distorted by a three-day-shorter quarter.

Revenue

This is a clean beat with nothing to strip out. There is no acquisition contribution, no accounting change, and no meaningful currency distortion in a business that reports and largely bills in dollars. The only adjustment worth making runs the other way: the quarter had 89 days against 90 a year ago, and on the CEO's own arithmetic the leap-day lap costs roughly 140bp, taking underlying growth to 28%. We flagged that artifact in February when it was working against management's Q1 guide; it is only fair to net it out now that the quarter has printed.

The more important observation is structural. In February, Snowflake guided Q1 to roughly 21.3% growth and set a full year that required the following three quarters to average 24.3%. That was a guide with its evidence in the future. Q1 came in at 26.2%, Q2 is guided to 25.1% at the midpoint, and the residual asked of the second half is 24.3%. The number did not change; the amount of it that remains unproven did. Roughly half the year is now either delivered or guided at a rate above the full-year average, and the acceleration that was the central underwriting question in February simply did not need to happen.

Assessment: The revenue line resolves the largest open item from the initiation, and it resolves it by removing the question rather than by answering it. That is the better outcome for a consumption business, because a back-half step-up that must be delivered is a forecast while a front-half beat is a fact.

Margins

The operating margin decomposition is the single most important table in this print and it inverts the one we published in February. Non-GAAP gross margin fell 133bp year on year, to 72.2% from 73.5%. Non-GAAP operating expense fell 575bp as a share of revenue, to 63.4% from 69.1%, with sales and marketing contributing 337bp, research and development 160bp and general and administrative 78bp. Net the two and you land on the +442bp the operating line delivered, which is the figure the CFO cited on the call.

Compare that with Q4, where 240bp of sales and marketing leverage was almost entirely consumed by 211bp of gross margin compression and the operating line moved 25bp. Two things changed. The leverage got broader: it is no longer only sales and marketing, because research and development contributed for the first time in this cycle alongside general and administrative. And the compression got smaller: 133bp of total gross margin erosion against 211bp a quarter ago, with product gross margin essentially flat sequentially at 75.7% versus 75.8%.

That second point deserves emphasis because it is the condition we named in February. Non-GAAP product gross margin has now been sequentially stable for two consecutive quarters while holding above 75%, and the year-on-year compression has narrowed from 221bp in Q4 to 126bp in Q1. This is what a floor looks like while it is forming. It is not yet what a floor looks like once it has formed, and the full-year guide still says roughly 75%, which is below where Q1 landed.

Assessment: Bull-2 in our standing thesis, operating leverage, moves from at risk to on track. The mechanism was never in doubt; the result was, and the result arrived. Bear-1, gross margin compression, is challenged rather than resolved: two quarters of stability is the beginning of the evidence, not the end of it, and management continues to guide the metric down from where it is currently printing.

EPS and Below the Line

Non-GAAP diluted EPS of $0.24 beat the $0.21 consensus by 14.3% and grew 63.9% year on year on the unrounded reconciliation figures ($0.2353 against $0.1436) versus 67.3% growth in non-GAAP net income attributable. The 3.4-point shortfall between those two rates is share count: the non-GAAP diluted base rose 2.1% to 370.9M, almost entirely because the zero-coupon convertible notes now contribute roughly 14.2M if-converted shares that were absent a year ago. The quality of the EPS beat is therefore operational rather than financial. There is no tax-rate trick here; the income tax effect of the non-GAAP adjustments was $23.5M against $15.6M a year ago, moving in the direction that reduces rather than flatters the result.

The GAAP picture went the other way and the reason is discrete. GAAP operating loss widened to $447.3M from $348.6M, but $106.5M of that is the impairment on the San Mateo office facility Snowflake ceased using during the quarter. Strip it out and the GAAP operating margin improved 936bp year on year, to negative 32.7% from negative 42.1%. That is the honest read of the GAAP trend. The less flattering read is that the impairment writes down a right-of-use asset while leaving $414.2M of operating lease liabilities on the balance sheet, so the add-back removes an accounting charge without removing the cash obligation behind it.

Stock-based compensation on the cash-flow basis management uses was $379.5M, or 36.4% of revenue, down 364bp from 40.1%. That is already below the roughly 37% full-year target set in February, one quarter into the year. On the income-statement reconciliation basis, which additionally captures $19.5M of employer payroll taxes on employee stock transactions, the figure is $408.7M and 39.2% of revenue.

Assessment: The EPS beat is real and operationally sourced. The GAAP-to-non-GAAP gap of $517.4M this quarter remains the reason we value this business on cash rather than on adjusted earnings, and this is the quarter in which that choice of denominator became inconvenient.

Revenue Composition and Key Business Metrics

Snowflake reports as a single operating segment and does not disclose revenue by vertical, geography or product line. The composition that is disclosed is the split between consumption-based product revenue and the services attached to it, and the five key business metrics the company defines in every release. Those are the segment-equivalent disclosures and they are where the mix story lives.

Snowflake restates customer counts for acquisitions, consolidations and spin-offs and re-bases the Forbes Global 2000 cohort to the current-year list, so the Q1 FY25 counts shown are as originally reported and will not reconcile arithmetically to the company-computed growth rates in the YoY column. The implied restated bases are 477 and 725 respectively. The weekly AI or ML account figure is a call disclosure rather than a defined key business metric and carries no revenue attached to it.

Product revenue

The consumption engine did what the guide said it could not yet do. Management called out Snowpark and Dynamic Tables as outperforming expectations and named technology and retail as the strongest verticals. Neither product was sized. The one quantitative fact about new-product contribution that exists in the public record remains the disclosure from February that Snowpark represented roughly 3% of FY25 product revenue.

"Product revenue for Q1 was $997 million, up a strong 26% year-over-year. Excluding the impact of leap year, product revenue grew 28% year-over-year. Our growth rate was stable quarter-over-quarter, showing no deceleration."
— Sridhar Ramaswamy, CEO

Assessment: Stability at 26% is a better outcome than the reacceleration story would have been, because it requires nothing from the future. The disclosure gap is unchanged: a company whose product revenue growth depends on new workloads entering production still does not size any of them.

Professional services and other

The smaller line moved the wrong way on both counts. Revenue grew 15.7%, roughly ten points slower than product, and the non-GAAP gross margin flipped from positive 4.6% to negative 4.4%, a $3.8M swing on a $45.3M line. On a GAAP basis the gross loss widened to $18.2M from $13.7M. Neither management nor a single analyst mentioned it.

Assessment: Immaterial to earnings and not immaterial as a signal. Professional services is the implementation capacity that converts a signed contract into consumption, and a services business that is growing slower than the platform it enables while losing money at the gross line is either being deliberately subsidised to accelerate migrations or is under-priced by accident. Snowflake has not told us which, so we treat it as the former, but it costs the company nothing to say so.

Net revenue retention

124% is the sixth consecutive quarterly decline and it is precisely the number at or below which we said in February that the FY26 revenue bridge loses its foundation. That threshold has now been reached, because the shortfall in expansion has been offset by an unusually strong new-customer cohort, which is exactly the mechanism the CFO described when the metric was challenged in Q&A. The mechanism is real and it is also a change in the character of the growth: retention-driven growth is high-margin and self-sustaining, and acquisition-driven growth carries a sales and marketing cost that shows up before the consumption does.

"a number of our newer customers that are not in that cohort are contributing to our growth beyond what the NRR is. And there was one of our large customers that grew so much last year. In this year, they're still doing very well but they didn't grow as much this year."
— Mike Scarpelli, CFO

Assessment: The explanation is credible and the number is still at the floor. We now model net revenue retention exiting FY26 at 122%, two points below management's verbal range, because the trend has been monotonic for a year and a half and nothing on this call suggested an inflection.

Remaining performance obligations and the two large contracts

RPO of $6.7B grew 34%, up from 33% growth in Q4, and the reacceleration has a specific cause. Both of the large customers who exhausted their capacity commitments in Q4 and elected to buy on demand rather than re-up early signed contracts of more than $100M in Q1, and both are in financial services. That closes the loop on the single largest disclosure ambiguity from the February print, where consumption ran ahead of the guide while contractual coverage thinned.

"On our last call, I noted two large customers ran out of capacity in Q4 and elected to delay their larger renewals. As expected, both of these accounts signed $100 million-plus contracts in Q1. We view this variability in bookings as normal for our model."
— Mike Scarpelli, CFO

Assessment: This is the cleanest commitment-kept on the call. In February we wrote that the consumption was real and the contractual coverage behind it had thinned; the coverage came back, on time and at the size implied. Bear-3 in our thesis, visibility erosion, moves from emerging to contained.

Customer cohorts

Customers spending more than $1M on a trailing twelve-month basis reached 606, a sequential addition of 26, below the 38 added in Q4 but the strongest first-quarter figure the metric has produced. Forbes Global 2000 customers recovered to 754 from 745, which retraces the nine-account sequential decline we flagged in February as an unexplained list re-basing. The recovery is confirmation that it was a re-basing rather than churn, and management again said nothing about either direction of the move.

Assessment: The enterprise-penetration metrics are healthy and the disclosure discipline around them is not. A metric that falls nine accounts one quarter, recovers nine the next, and is never mentioned in either direction is a metric the company does not consider load-bearing. Investors who track it should price it accordingly.

Key Topics & Management Commentary

Overall Management Tone: Notably more forward-leaning than February, and for the first time in several quarters the confidence was attached to results already in hand rather than to a plan. Prepared remarks led with the growth rate and the margin expansion instead of with customer anecdotes, and the questions management handled least completely were the ones about the numbers they chose not to change: the unaltered full-year margin and cash-flow targets drew a one-clause answer about event costs, and the collapse in first-quarter cash conversion was pre-empted in the script and never revisited. Tone on the unresolved CFO succession was procedural to the point of terseness.

1. The Back-Half Acceleration Became a Front-Half Result

February's guide had a shape that made underwriting it difficult: decelerate to 21.3% in Q1, then accelerate to an average of 24.3% across the remaining three quarters, with the step-up placed specifically in the second half and none of its three named drivers sized. We wrote at the time that roughly 110bp of the Q1 optic was a leap-day artifact and that the genuine implied acceleration was closer to 200bp, which was a modest number dressed up by the calendar as a dramatic one.

The calendar argument turned out to be right and the acceleration turned out to be unnecessary. Q1 printed 26.2%, 500bp above the guided midpoint. Q2 is guided to 25.1%. The residual asked of the second half is 24.3%, which is the identical figure the February guide asked of the last three quarters, except that it now applies to two quarters instead of three and follows a first half running at 25.7% rather than a first quarter guided at 21.3%.

Assessment: The risk profile of the full-year number improved materially even though the number itself barely moved. A guide whose growth is front-loaded and delivered is worth a different multiple from a guide whose growth is back-loaded and promised, and this quarter converted one into the other.

2. Operating Leverage Reached the Operating Line

The 442bp of year-on-year non-GAAP operating margin expansion is the largest this company has produced and it is the first time in this cycle that the efficiency program showed up in a year-on-year comparison rather than only against a conservatively set quarterly guide. The CFO delivered the figure without elaboration; the CEO returned to it later in Q&A with a framing that connects the margin to Snowflake's own internal use of the product it sells.

"It was 4% in Q1 last year. It's 9% Q1 this year. And this is part of the benefit of practicing what we preach around AI. We spend a lot of time figuring out how engineers can be more productive with AI, how we can get small work done."
— Sridhar Ramaswamy, CEO

The composition matters as much as the size. In Q4 the leverage was almost entirely sales and marketing, with research and development actually rising 19bp as a share of revenue. This quarter research and development contributed 160bp and general and administrative 78bp alongside sales and marketing's 337bp. Broad-based leverage across all three expense lines while the sales organisation is hiring at its seasonal peak is a different and more durable signal than one line carrying the whole result.

Assessment: This is the quarter that converts the efficiency narrative into an operating fact. It is also the quarter in which management declined to raise the full-year margin guide, which we treat below as the more informative of the two decisions.

3. Product Gross Margin Held Its Line

Non-GAAP product gross margin of 75.7% is 15bp below Q4 and 126bp below the year-ago quarter, and the year-on-year compression has narrowed from 221bp in Q4 to 126bp now. The full-year guide remains approximately 75%, unchanged from February and below where the first quarter actually landed. Nobody asked management to reconcile the two, and management did not volunteer it.

The AI compute cost argument that has driven three years of compression was not discussed on this call at all. The closest thing to a margin comment on the product came from the head of product describing the Gen2 compute environment, which combines newer and more expensive cloud instances with software optimisation.

"What we've done is we've combined the latest hardware instances that we can get from the cloud providers, which are often faster but also more expensive with a good number of software and improvements that we have, and at the end of the day is part of our eternal ongoing promise to customers to always deliver the best price performance of the market."
— Christian Kleinerman, EVP of Product

Assessment: Two consecutive prints at or above 75% satisfy the first of the two conditions we set in February for an upgrade. The unchanged guide is a reminder that management still expects the metric to sit below the current run-rate for the year, and Gen2's economics, faster and more expensive silicon sold on a price-performance promise, describe a product decision that trades gross margin for competitive position.

4. The Raise Was the Beat

Full-year product revenue guidance moved to $4.325B from $4.28B, an increase of $45M. Product revenue beat the guide midpoint by $39.3M. The dollars allocated to the remaining three quarters therefore changed by $5.7M, or 0.13% of the full-year figure. Management raised the full year by the size of the beat and left the rest of the year exactly where it was.

This is worth stating plainly because the framing on the call was more expansive than the arithmetic.

"As you can see, we have started the year with strong revenue growth and overall very healthy results, and we are increasing our growth expectations for the year."
— Sridhar Ramaswamy, CEO

Growth expectations for the year did increase, from 23.6% to 24.9%. Growth expectations for the part of the year that has not happened yet did not.

Assessment: Flowing a beat through without adding to it is the conservative, entirely defensible convention this management team has used for years, and it is consistent with a guide-low-and-beat pattern. It is also the reason we do not treat this print as satisfying our second upgrade condition. A raise that confirms the second half would have added to the second half.

5. Free Cash Flow Went the Other Way

Adjusted free cash flow of $206.3M is 43.6% below the year-ago quarter and the margin fell to 19.8% from 44.1%, a 2,433bp decline. Operating cash flow fell to $228.4M from $355.5M. Capital expenditure nearly doubled to $45.0M from $23.9M. The balance sheet shows the mechanism directly: accounts receivable released $393.7M of cash against $579.3M a year ago, and deferred revenue consumed $271.4M. Total deferred revenue fell 10.5% sequentially and cash and investments fell $427.3M.

Management put the explanation in the prepared remarks rather than waiting for a question.

"As discussed on our last earnings call, we had several large customers purchase as they consumed in Q4. This booking behavior impacts the seasonality of our free cash flow. We expect this year to be more second half weighted."
— Mike Scarpelli, CFO

The explanation is coherent and it is also incomplete as an underwriting basis. The full-year adjusted free cash flow guide of 25% was left unchanged. On our estimate of the full-year revenue base, holding that guide requires roughly $938M of adjusted free cash flow across the remaining nine months, against $576M generated in the comparable nine months of FY25. That is a 63% increase in the back three quarters of the year, and it is being underwritten by the same seasonality argument that just produced a 2,433bp margin decline in the quarter it was applied to.

Assessment: This is the one genuinely new negative in the print, and it lands on the metric we identified in February as the honest valuation anchor, because adjusted free cash flow is the only figure in this company's disclosure that is simultaneously positive, cash, and net of the buyback needed to hold dilution flat. Bull-4 in our thesis, cash generation, moves from on track to at risk. It is a timing question rather than a quality question until proven otherwise, and Q2 is when the proof starts.

6. Bookings, New Logos and the Sales Compensation Change

Snowflake added 451 net new customers and passed 11,000 in total. The CFO attributed the strength to a dedicated new-logo acquisition team stood up during FY25 rather than to anything that changed this quarter, and said the model is being replicated from North America into EMEA. Separately, the bookings component added to sales compensation in FY26 was described as helpful but not decisive.

"We started last year with really breaking out an acquisition team that is just focused on new logos, and we're seeing the benefits of the groundwork that we put in place last year and we're pleased with the results."
— Mike Scarpelli, CFO

On the compensation change, the answer was more measured than the question invited: the two large deals were known to be coming, the bookings quota helped at the margin, and the principal driver of sales compensation remains consumption revenue. Management explicitly said it is too early to judge whether the change worked.

Assessment: Refusing to claim credit for a structural change whose effect cannot yet be isolated is the correct answer and an unusual one. The new-logo cohort is doing real work in the growth algebra now that net revenue retention is at 124%, which makes the durability of the acquisition motion a more load-bearing question than it was a year ago.

7. AI: 5,200 Weekly Accounts and Still No Revenue Figure

The AI disclosure improved on the only axis Snowflake discloses. Accounts using AI or machine learning weekly went from more than 4,000 to more than 5,200, a 30% sequential increase. Cortex Agents shipped, Llama 4 was available on day one, and the Microsoft partnership hosting OpenAI models on Azure was reiterated. There is still no revenue figure, no consumption figure and no attach rate, and this quarter management made the absence a stated position rather than an omission.

"As you know, we don't sell AI separately. It's not a SKU. Customers are not signing up for contracts on AI so it's on their existing spend."
— Sridhar Ramaswamy, CEO

The strategic logic is sound: bundling AI into platform consumption rather than pricing it separately is how you avoid a per-seat AI product that customers evaluate against a standalone budget line. The consequence for investors is that the AI contribution to the 26% growth rate is unmeasurable from outside, permanently, by design.

Assessment: Bull-3 in our thesis stays at risk for the same reason it was at risk in February, with one change in character. In February the absence of AI revenue disclosure looked like a company waiting until it had a number worth showing. After this call it looks like a company that has decided the number will never exist as a separate disclosure. That is a defensible product strategy and a permanent gap in the investment case.

8. Iceberg, Open Formats and the Compute Question

Snowflake extended data sharing, security and performance optimisation to Apache Iceberg tables during the quarter, which materially widens what customers can do with data that Snowflake does not store. The head of product framed open formats as a customer-choice feature rather than a defensive one.

"the other investment that we made is around Iceberg, which also creates opportunity for customers who have an open architecture and be able to mix and match technologies as they see fit."
— Christian Kleinerman, EVP of Product

We wrote in February that the storage-leakage thesis was dead, since storage is roughly 11% of revenue, and that the live question was whether compute over externally held Iceberg tables earns the same gross margin as compute over native Snowflake storage. That question went unasked for a second consecutive quarter. Extending core platform capabilities to Iceberg makes it more urgent, not less, because it increases the share of Snowflake compute that could run against data the customer is free to point another engine at.

Assessment: Strategically correct and analytically unresolved. Two quarters of stable product gross margin while Iceberg capability expands is mild evidence that the compute economics are not materially different, which is the only evidence available.

9. Macro, Tariffs and the Absence of an Optimization Wave

Neither prepared remarks nor the numbers showed macro pressure, and the CFO drew an explicit contrast with 2022 and 2023 when asked. The distinction he made was about customer composition rather than about the economy: the digitally native, well-funded startups that drove the post-COVID optimization cycle are no longer the marginal spender.

"I am not seeing any big optimizations plan within our customers like what we saw coming out of COVID with those. But I will remind you, our customers are constantly optimizing. That may be a little bit, but they're always looking to do things more efficiently and that will continue."
— Mike Scarpelli, CFO

The CEO added a point that reframes optimization as a sales discipline rather than a demand risk, noting that inefficient customer spend produces a contraction later regardless, so the sales organisation now tidies up implementations proactively.

Management also said the current tariff news flow had shown no visible effect, and pointed to new-customer additions and RPO growth as the evidence.

Assessment: This is the most convincing macro answer we have heard from this company, because it is grounded in a structural change in the customer base rather than in an assertion about demand. The caveat is timing: the quarter ended April 30 and the tariff announcements were still fresh, so the test is Q2, not Q1.

10. Capital Allocation: A Buyback Executed at the Low

Snowflake repurchased 3.2 million shares for $491M at a weighted average price of $152.63, with $1.5B remaining on the authorisation through March 2027. The context matters: the stock closed at $137.00 thirty days before the print, and the average price paid sits 14.8% below the $179.12 close the shares carried into the release. In February the company had repurchased nothing in the fourth quarter and offered no rationale, and we flagged the silence as a missed opportunity to tell the market what price it considered value. The Q1 execution answers the question by action.

For the first time, the GAAP weighted-average share count declined year on year, to 332.7M from 333.6M. The non-GAAP diluted count still rose 2.1%, to 370.9M, on the if-converted treatment of the convertible notes.

"We will continue to evaluate share buyback on a quarterly basis, and we have no plans right now. We've been more opportunistic in terms of the buyback, but we do fully anticipate between now and 2027, we will utilize that."
— Mike Scarpelli, CFO

Assessment: Buying $491M of stock during the April drawdown at a 15% discount to the pre-print level is the best capital allocation decision this company has made in the period we have covered it. It does not resolve Bear-2, because $491M of repurchase against $379.5M of quarterly stock-based compensation is an anti-dilution program rather than a return of capital, but it is the difference between a mechanical buyback and a priced one.

11. The CFO Succession Is Still Open

Three months after announcing his retirement with no successor and no effective date, the incumbent CFO provided a status line and nothing more.

"Before opening up the line for questions, I just want to update you on my transition as mentioned on the Q4 conference call last quarter. We are in the process of interviewing many great candidates, and we will make an announcement in the future when we have more firm details to share."
— Mike Scarpelli, CFO

No analyst asked a follow-up. The CEO made one passing reference to a newly arrived executive looking at the federal opportunity; the transcript labels that person the new CFO, but no successor had been named and the same call introduces a new chief revenue officer by the same first name, so the label is almost certainly a transcription error.

Assessment: Bear-4 stays emerging. The guidance philosophy this market prices into Snowflake, guide conservatively and beat, is a personal convention rather than an institutional one, and the person who authored it is now guiding a year he will not finish. That risk has not changed since February and the market has stopped asking about it, which is not the same as the risk having been resolved.

12. Public Sector as a New Surface

Snowflake launched Snowflake Public Sector Inc. and disclosed a Department of Defense Impact Level provisional authorization, which is the gating credential for classified and controlled-unclassified workloads.

"With the launch of Snowflake Public Sector Inc., and our recent Department of Defense Impact Level provisional authorization, we're now equipped to deliver mission-critical data and AI solutions to the national security community, including the United States Department of Defense, its military branches and industry partners."
— Sridhar Ramaswamy, CEO

Asked whether federal procurement would resume after the current spending review, the CEO described active conversations across departments, cited low operational overhead as the competitive argument, and said Snowflake would have more to say in coming quarters.

Assessment: A genuine addressable-market extension with a long sales cycle and no revenue attached to it yet. Worth tracking as an FY27 item rather than an FY26 one, and worth noting that it is the only new growth surface management introduced this quarter that has a certification gate rather than a product gate in front of it.

Guidance & Outlook

The revenue guide is a raise and the margin guides are not, and the gap between those two decisions is the most informative thing in this section. Snowflake delivered a first quarter that cleared the full-year product gross margin guide by roughly 65bp, cleared the full-year operating margin guide by roughly 80bp, and fell 520bp short of the full-year adjusted free cash flow margin guide. It then left all three unchanged.

Implied quarter-over-quarter ramp. The full-year product revenue target of $4,325M less the $996.8M delivered and the $1,037.5M guided at the Q2 midpoint leaves $2,290.7M for the second half, or 24.3% growth against the $1,843.6M Snowflake produced in the second half of FY25. The first half, actual plus guide, runs at 25.7%. The year is therefore now shaped as hold-then-drift rather than the decelerate-then-accelerate structure the February guide described.

Implied margin path. Q1 printed 8.8% and Q2 is guided to 8%, so an unchanged 8% full-year figure requires no expansion at all in the second half. On our estimate of the full-year revenue base the arithmetic implies a second-half non-GAAP operating margin closer to 7.7%, below both the quarter just reported and the quarter just guided. The CFO's stated reason covers Q2 but not the second half.

"Well, what I would say is in Q2 is when we have a big user event, and that's a very expensive event, Summit that we operate, and that typically has an impact on our operating margin in Q2. And that's factored in. And we'll just continue to revise our forecast for the year on a quarterly basis going forward."
— Mike Scarpelli, CFO

Implied cash-flow path. This is where the unchanged guide asks the most. Holding a 25% full-year adjusted free cash flow margin against our estimated revenue base requires roughly $938M across Q2 through Q4, against $576M in the same three quarters of FY25, an increase of about 63%. The stated basis is the reversal of the Q4 consume-as-you-go buying pattern, which pulled collections forward. That is a real mechanism and it is a large number to hang on it.

Street at: the Q2 product revenue guide of $1,035M to $1,040M sits about 1.6% above the $1,021M consensus at the midpoint, and the full-year raise takes the target roughly 1% above where the Street had been carrying it. Both are modest, and both are consistent with the beat-and-raise cadence this management team has run for years.

Guidance style: unchanged and conservative. Q4 FY25 delivered a 9.4% operating margin against a 4% quarterly guide. Q1 FY26 delivered 8.8% against a full-year 8%. Snowflake sets numbers it intends to clear, which is why the revenue raise should be read as a floor and the margin non-raise should not be read as a ceiling. The reason we still treat the unchanged margin guides as informative is that they are the company's own statement about where reinvestment goes when the quarter comes in ahead.

"our guidance is based upon the observed behavior we see within our customers, coupled with we spend a lot of time than we have for the last five quarters now, in really identifying new workloads going to production. We have a pretty good visibility of those, and we're very close with our customers, and we know what we're doing."
— Mike Scarpelli, CFO

Analyst Q&A Highlights

Intra-Quarter Consumption and What May Looked Like

The first question of the call went straight at the only piece of information the print does not contain: how consumption behaved through the quarter and into the current one. Management declined the first half and answered the second half indirectly, pointing at the guidance itself as the disclosure. The exchange establishes the ground rule for everything that follows, which is that this company will discuss the shape of consumption only through the numbers it has already guided.

Q: "Sridhar, I want to talk about some of the trend lines of the business, particularly around consumption as it progressed through the quarter. How was consumption sort of exiting the quarter and through the month of May?"
— Sanjit Singh for Keith Weiss, Morgan Stanley

A: "As you know, we don't comment on consumption within a quarter. Overall, Q1 consumption was very strong coming out of the holiday period, and you see that in our results. Of course, Q1 had one less day compared to Q1 last year. But overall, we feel very good about our consumption."
— Sridhar Ramaswamy, CEO

Assessment: A refusal, followed immediately by the CFO noting that the Q2 guide reflects customer behaviour observed through the day of the call. That second half is the substantive answer and it is the reason the Q2 guide carries more information than a normal quarterly outlook: it was set with three weeks of the quarter already visible.

Why Net Revenue Retention Is Not Higher

The sharpest question of the call attacked the retention metric from an unusual angle. Rather than treating 124% as a decline to be explained, it argued that the combination of a record pace of product launches, a higher-quality enterprise customer mix and an AI tailwind that did not exist in the prior cycle should be pushing the number up, not down. The answer conceded the point implicitly by relocating the growth: the expansion is coming from customers outside the measurement cohort.

Q: "If I could take the liberty of poking at the NER, so 124% is good. Not many companies even reach that number at your scale if they reach your scale. But you have had 137 new products launched in the most recent quarter, and Scarpelli will be quick to point out that the NER metric is more of a 24-month trailing metric, which I completely appreciate. But why wouldn't and why shouldn't NER be better, given the tail effect of new product introduction, you're landing customers at a record pace."
— Kasthuri Rangan, Goldman Sachs

A: "What I would say on that is a number of our newer customers that are not in that cohort are contributing to our growth beyond what the NRR is. And there was one of our large customers that grew so much last year. In this year, they're still doing very well but they didn't grow as much this year. And that's really the dynamics. It's those newer customers coupled with that. But once again, over time, NER and revenue growth rate will converge as we become a more mature company."
— Mike Scarpelli, CFO

Assessment: The most consequential exchange on the call. Management is telling you that the growth composition has shifted from expansion within the installed base toward acquisition of new logos, and that the two metrics will converge downward over time. Both statements are honest and both reduce the quality of the growth relative to the 128% cohort-expansion engine of a year ago.

Why the Full-Year Margin and Cash-Flow Targets Were Left Unchanged

One questioner isolated the exact inconsistency in the guidance: a first quarter that cleared both the operating margin and the framework behind the cash-flow target, followed by no change to either full-year figure. The answer covered the second quarter and not the second half, and no follow-up pressed the gap.

Q: "But one key operating margin was strong. Nonetheless, the fiscal year op margin was left unchanged, as was the fiscal year free cash flow margin target. So I guess, what were the puts and takes there as to why leave those targets unchanged?"
— Patrick Colville, Scotiabank

A: "Well, what I would say is in Q2 is when we have a big user event, and that's a very expensive event, Summit that we operate, and that typically has an impact on our operating margin in Q2. And that's factored in. And we'll just continue to revise our forecast for the year on a quarterly basis going forward."
— Mike Scarpelli, CFO

Assessment: A one-variable answer to a two-variable question. Summit explains Q2's 8% guide against Q1's 8.8% print; it does not explain why the second half is implicitly guided below both. The free cash flow half of the question was not addressed at all, which is notable given that adjusted free cash flow margin fell 2,433bp in the quarter being discussed.

What Underpins Confidence in the Rest of the Year

A direct request for the assumptions behind the full-year ramp produced the clearest statement of Snowflake's forecasting method available in the public record: guidance is set from observed consumption behaviour plus a workload-level pipeline of production deployments that the company has been tracking systematically for five quarters. Migration tooling was named as the specific accelerant.

Q: "Guidance implies a robust ramp through the remainder of the year. And I think we all see the pace of innovation. Excited for what's to come at Summit. But what, if anything, would you call out that underpins your confidence things we might not be thinking about or any key assumptions worth calling out?"
— Brad Zelnick, Deutsche Bank

A: "I would just say, as I said, our guidance is based upon the observed behavior we see within our customers, coupled with we spend a lot of time than we have for the last five quarters now, in really identifying new workloads going to production. We have a pretty good visibility of those, and we're very close with our customers, and we know what we're doing. Migrations are moving nicely. We announced where we've made Snow Convert available to all of our customers and partners, and we're seeing an uptick in the amount of usage around that, and that's what gives us the confidence in the guidance that we gave."
— Mike Scarpelli, CFO

Assessment: A better answer than the equivalent one in February, and still unsized. The workload-pipeline method is credible and is the same method that produced a guide beaten by 3.8% this quarter. What an investor cannot do with it is independently verify anything, which is why the front-loading of this year's growth matters more than the explanation of it.

The Two $100 Million Contracts and the New-Logo Cohort

The question that closed February's largest loose end asked both what drove the new-customer strength and where the two delayed mega-contracts landed. The answer disclosed the vertical for the first time and attributed the logo momentum to a structural change made a year earlier rather than to anything that happened this quarter.

Q: "Mike, you mentioned earlier on the call you had the strongest new logo quarter, which was fantastic. Just a question around that. Are you seeing -- is this a result of better, stronger execution and strategic focus? Are you seeing a more favorable win rate in competitive environment? And just as a follow-up, too, to that is of those two $100 million deals, can you just tell us which verticals they were in?"
— Joel Fishbein, Truist

A: "So on the $100 million deals, they were both in the financial services vertical. And what I would say on the number of new customers, this is not a result of something we put in place this quarter. We started last year with really breaking out an acquisition team that is just focused on new logos, and we're seeing the benefits of the groundwork that we put in place last year and we're pleased with the results."
— Mike Scarpelli, CFO

Assessment: Both halves are favourable and the second half is the more durable one. Two nine-figure financial services renewals landing in the quarter after they slipped resolves a specific concern; a purpose-built acquisition team producing 451 net adds in its second year addresses the structural question of where growth comes from once net revenue retention converges toward the corporate growth rate.

Macro Contrast Against the Post-COVID Optimization Cycle

The macro question was posed as a comparison rather than as a forecast, which produced a more useful answer than the equivalent exchange in February. The distinction drawn was about who the customers now are, not about what the economy is doing.

Q: "Mike, in the comments from you and Sridhar, there was really no mention of macro per se and no evidence in the numbers that you guys really saw much pressure. You certainly did back in 2022, 2023. And I'm just curious, Mike, how you would draw a contrast. Is it that a lot of the post-COVID optimization efforts are now largely behind you, Snowflake just in a better place in terms of the product portfolio?"
— Karl Keirstead, UBS

A: "I would say coming out of COVID, I think it was very different. In that environment, we had a lot of digitally-native well-funded start-ups that were spending crazy and weren't really focused on costs as much. Our customer base has really evolved into the -- some of the largest companies in the world that are much more mature, that are much more cost-focused. And I am not seeing any big optimizations plan within our customers like what we saw coming out of COVID with those."
— Mike Scarpelli, CFO

Assessment: The strongest macro answer this management has given, because a change in customer composition is verifiable and an assertion about demand is not. The follow-up disclosed that AI-native companies are each less than 1% of revenue, which removes the most obvious source of low-quality upside from the quarter and makes the 26% growth rate more attributable to the enterprise base.

Hyperscaler Data Fabric and the Coexistence Argument

The competitive question of the call was about whether the hyperscalers converging on open, agent-ready data fabric architectures compresses Snowflake's position. The answer declined the framing entirely and argued for coexistence, citing interoperability work in both directions with the platform that had made the announcement in question a day earlier.

Q: "When you look at what's happening with the hyperscalers, Microsoft certainly talked about fabric yesterday, on and build, certainly seem to be making a lot of progress in the idea of data fabric at scale, truly open to enable AI agentic architecture. It's a thing that's not lost on the big guys. Where does that lead Snowflake? And what is going to be the one or two or three things that the Snowflake's platform would do better so you can build your vision towards being a $10-plus billion revenue company?"
— Kasthuri Rangan, Goldman Sachs

A: "I think it was six, seven months ago that we announced, for example, that from Snowflake, you could read tables that are in 1 lake, and we are also actively talking to them about 1 lake being the data layer for Snowflake that's at the bottom. We also collaborate with them at the top, where we have things like Cortex analysts and Cortex agents be available as components in Office Copilot, for example." … "Yes, there is competition, but I think there are more cases than not where we are very, very effectively working together, and it's on the uptick, especially with Azure."
— Sridhar Ramaswamy, CEO

Assessment: A confident answer that also, read carefully, describes Snowflake becoming a compute and intelligence layer over storage it does not own. Coexistence is genuinely the near-term commercial reality and it is the same architecture that makes the Iceberg compute-margin question the right one to ask.

The Capital Expenditure Step-Up

The one question about the balance sheet got a specific, checkable answer and a forward commitment. Capital expenditure nearly doubled year on year, and the cause is office build-out rather than infrastructure.

Q: "And then just a quick one, getting into the weeds a little bit. CapEx was up a bunch to a fairly high number. Is there onetime items there or is this the new level?"
— Brad Reback, Stifel

A: "No. The CapEx was really associated with our new headquarters in San Mateo. As I spoke about previously, we signed a new lease in a Menlo Park office and there's a fair bit of CapEx that went into that as well as in Bellevue. I talked about that before, too. That really -- that just opened this week, that office, and there was a fair bit of CapEx that went into that as well, too. I'm not expecting any major office buildouts the next couple of years actually now."
— Mike Scarpelli, CFO

Assessment: A clean, falsifiable commitment on a small line item, and a useful one because it converts roughly $20M of the quarter's free cash flow shortfall into a non-recurring item. It sits alongside the $106.5M impairment on the San Mateo facility Snowflake exited during the same quarter, so the real estate footprint is being simultaneously written down and rebuilt.

What They're NOT Saying

1. Any revenue figure for AI, and now a stated reason why there will not be one. The disclosure remains a weekly active account count, 5,200 and rising 30% sequentially, with no revenue, no consumption and no attach rate. The new development is that management framed the absence as a product decision rather than a timing one. A permanent non-disclosure is a legitimate strategy and it removes any prospect of independently sizing the AI contribution to growth.
2. Why the full-year gross margin and operating margin guides sit below the quarter just printed. Q1 delivered 75.7% product gross margin against a roughly 75% full-year guide, and 8.8% operating margin against an 8% full-year guide. Both were maintained. Only the Q2 half of the operating margin question was answered.
3. What produces $938M of adjusted free cash flow in the remaining nine months. The 25% full-year margin guide was reaffirmed after a quarter that delivered 19.8%. The stated mechanism, reversal of Q4 consume-as-you-go buying, was offered pre-emptively and never quantified, and no analyst asked for the bridge.
4. The size of Snowpark, Dynamic Tables or any other named growth driver. Both were called out as outperforming expectations. Neither was sized. The last quantified new-product disclosure remains Snowpark at roughly 3% of FY25 product revenue, given in February.
5. Whether compute over Iceberg tables carries native gross margin. Snowflake extended data sharing, security and performance optimisation to Iceberg this quarter, which increases the share of compute that could run against externally governed data. The question connecting that to the gross margin guide has now gone unasked for two consecutive calls.
6. Who the next CFO is, or when. A status line in the prepared remarks, no follow-up from a single analyst, and a passing reference by the CEO to a newly arrived executive already looking at the federal opportunity that does not obviously fit a search still described as an interview process.
7. Why professional services gross margin went negative. The non-GAAP margin on the line flipped from positive 4.6% to negative 4.4% and the GAAP gross loss widened to $18.2M. Not mentioned by management, not raised in Q&A.
8. Any path or timeline to GAAP profitability. Absent from prepared remarks and from Q&A for a second consecutive quarter. GAAP operating loss widened $98.7M year on year, and stockholders' equity fell 19.7% sequentially to $2,408.0M.
9. What the second half assumes about macro. The macro answer was framed entirely retrospectively, against 2022 and 2023. The quarter ended April 30, days after the tariff announcements, and management said the effects were not yet visible. Nobody asked what the guide assumes if they become visible.
10. Why Forbes Global 2000 customers recovered nine accounts. The count fell 754 to 745 in Q4 without comment and returned to 754 in Q1 without comment. The round trip confirms the re-basing read but the company has now declined to explain the metric in both directions.

Market Reaction

• Pre-print setup: SNOW closed at $179.12 on May 21, up 16.0% year to date, up 30.7% over the trailing thirty days and up 10.1% over the trailing twelve months, against a 52-week closing range of $108.56 to $192.78. The S&P 500 was down 0.6% year to date over the same period. The thirty-day figure is the important one: the stock had already rallied off a $137.00 close on April 21, so this was not a beaten-down name going into the print.
• After-hours move: shares rose approximately 6% to around $190 in the extended session on May 21, with the initial move attributed to the product revenue beat, the above-consensus Q2 guide and the raised full-year target. That is less than half the move the stock ultimately made, which means the balance was added during and after the 5:00 PM ET call rather than on the release itself.
• Next-day session: the stock opened at $192.00 on May 22, a 7.2% upside gap, traded a $191.99 to $204.00 range, and closed at $203.18, up 13.4% or $24.06 on the day. The close sat $0.82 below the session high.
• Volume: 26.2M shares against a 30-day average of 3.8M, roughly 6.9 times normal.
• Index context: the S&P 500 was unchanged on the session, so essentially the entire move was idiosyncratic.

The shape is the story, and it is the exact inverse of February. Three months ago the stock gapped 12.5% higher and closed 0.06% above its session low, surrendering two thirds of the opening move on nearly four times normal volume. This time it gapped 7.2% higher, ran through the session, and closed within $0.82 of the high on nearly seven times normal volume. A stock that opens up and finishes on the high has met a buyer for every seller the print produced, and the volume says the buying was institutional rather than a thin drift.

What the tape priced. The overnight move priced the two headline facts, a 3.8% beat against the product revenue ceiling and a Q2 guide 1.6% above consensus. The daylight session priced the second layer, and this quarter the second layer was favourable rather than adverse: 442bp of operating margin expansion, product gross margin holding, the two delayed mega-contracts landing, and RPO growth reaccelerating. Those are the facts that require reading past the summary tables into the reconciliation, which is roughly the amount of work a market does between the after-hours print and the following afternoon. In February that work produced selling. In May it produced buying.

The setup cuts against the bulls from here. Entering the print the shares were up 30.7% in thirty days and 16.0% year to date; leaving it they are up 31.6% year to date at a fresh 52-week closing high, having moved through the middle of the pre-print sell-side target cluster in a single session. A stock that has already re-rated 13.4% on the operating improvement is no longer being paid to deliver it. That is the mechanical reason the rating does not change even though the quarter was better than the one that preceded it.

Street Perspective

Debate: Is the FY26 Raise a Real Raise?

Bull view: the beat-and-raise cadence is intact and this management has never guided a number it did not clear. Flowing the beat through in the first quarter of the year is exactly what a conservative CFO does when there are three quarters left to see, and the Q2 guide coming in 1.6% above consensus is the more forward-looking signal than the full-year point estimate.

Bear view: the raise is $45M against a $39.3M beat. Dollars allocated to the unfinished part of the year moved $5.7M. Management had a quarter that cleared every full-year margin target it had set and chose to change none of them. A company genuinely raising its expectations for the year raises the part of the year that has not happened.

Our take: the bears have the arithmetic and the bulls have the pattern, and the pattern has been right for six years. What settles it for us is that the composition of the year improved far more than the total did. Front-loading 25.7% growth into a first half that is now half delivered and half guided against a second half asked for 24.3% is a materially better risk position than the February structure, and that improvement is worth more than the $5.7M the guide added. It is not worth a rating change on its own.

Debate: Has Product Gross Margin Actually Found a Floor?

Bull view: two consecutive quarters of sequential stability above 75%, year-on-year compression narrowing from 221bp to 126bp, and the sequential move a rounding error at 15bp. The AI compute cost curve is falling industry-wide and Snowflake is offsetting what remains with software optimisation, which is what Gen2 is. The metric that was the central bear point six months ago was not raised by a single analyst on this call.

Bear view: two quarters is not a floor, it is two quarters, and management guides the full year to roughly 75% while printing 75.7%, which is the company telling you it expects the metric lower than where it currently sits. Gen2 explicitly buys faster and more expensive instances and passes the price performance to customers, which is a gross margin decision dressed as a product decision. Extending Iceberg support widens the share of compute running against data Snowflake does not store, and the economics of that compute have never been disclosed.

Our take: the bulls have the better of this for the first time since we initiated. The compression has decelerated for two consecutive quarters and the guide is a conservative convention rather than a forecast at this company. We move Bear-1 from materialising to emerging and we will move it to contained on a third print at or above 75%. The Iceberg compute question remains the one unasked question that could reopen it.

Debate: Does the Free Cash Flow Collapse Matter?

Bull view: it is pure timing. Several large customers bought as they consumed in Q4, which pulled collections into that quarter and out of this one, and the accounts receivable and deferred revenue movements show exactly that mechanism. The full-year guide is unchanged because management can see the collections calendar. Capital expenditure nearly doubling on one-time office build-out accounts for another $20M and is explicitly not recurring.

Bear view: a 2,433bp margin decline is a large amount of timing, and the unchanged 25% guide now requires roughly 63% more adjusted free cash flow in the remaining nine months than the same period produced last year. Deferred revenue fell 10.5% sequentially and cash and investments fell $427M in a quarter that also spent $491M on buybacks. Not one analyst asked for the bridge, which means the market has accepted a seasonality explanation for the metric that carries the valuation.

Our take: the bulls are probably right on mechanism and the bears are right that it is unverified. We split the difference by modelling 24% rather than the guided 25% and treating Q2 as the test. If adjusted free cash flow margin does not recover above 25% in Q2, the seasonality argument becomes a cash conversion argument, and this is the metric on which we value the company.

Debate: Is 16x Forward Product Revenue Still Reasonable?

Bull view: a platform compounding product revenue at 25% with 75.7% product gross margins that has just demonstrated 442bp of annual operating leverage, with $4.87B of cash and investments, no net debt under if-converted treatment, and a demonstrated willingness to buy its own stock in a drawdown, deserves to trade above the multiple it carried when the operating leverage was hypothetical. The February de-rating priced a company that could not convert growth into margin. That company no longer exists.

Bear view: the multiple went from 13.9x to 16.1x on enterprise value to forward product revenue in three months while the full-year revenue guide moved 1%. On enterprise value to forward adjusted free cash flow the stock is at roughly 64 times, against 53 times in February, and the cash flow denominator just deteriorated. The GAAP company still loses $447M a quarter and stock-based compensation runs at 36.4% of revenue. The re-rating has out-run the fundamental change by a wide margin.

Our take: both are true and the bears have the arithmetic on the pace. The fundamental improvement in this quarter is real and it is worth a higher multiple than 13.9x; it is not obviously worth 16.1x plus a 64 times cash flow multiple on a quarter in which cash flow fell 43.6%. The market re-rated the operating margin in a single session and has not yet priced the cash conversion question, because nobody on the call raised it. That asymmetry is the reason we raise the fair value range and keep the rating.

Model Update Needed

Valuation framework. At the $203.18 close and the guided 372M non-GAAP diluted share count, the equity is worth approximately $75.6B. Cash, cash equivalents and short- and long-term investments were $4,866.8M at April 30. Under the if-converted treatment already embedded in the guided share count, the zero-coupon notes are equity rather than debt, so enterprise value is approximately $70.7B. That is roughly 16.1x our FY26E product revenue of $4,390M and approximately 64x our FY26E adjusted free cash flow of about $1.10B, against 13.9x and roughly 53x on the same conventions at the February close of $173.69.

Fair value range: $185 to $225. The low end corresponds to about 14.6x enterprise value to FY26E product revenue, which is where we think the stock finds support if adjusted free cash flow margin fails to recover in Q2 or net revenue retention breaks below 124%. The high end corresponds to about 18.0x, achievable if product gross margin prints at or above 75% for a third consecutive quarter, cash conversion normalises, and the Q2 print raises the full year by more than the size of the Q2 beat. The midpoint of $205 sits roughly 1% above the current price, which is the arithmetic behind maintaining the rating rather than acting on the operating improvement.

What would change the number. Upside: an adjusted free cash flow margin above 25% in Q2, a Q2 raise to the full year that exceeds the Q2 beat, a third consecutive product gross margin print at or above 75%, a named CFO with a comparable guidance philosophy, or long-term financial targets introduced at the June Investor Day that exceed the current framework. Downside: net revenue retention below 124%, adjusted free cash flow margin below 22% in Q2, product gross margin below 74.5%, RPO growth decelerating again, or any trim to the FY26 revenue guide.

Thesis Scorecard Post-Earnings

The pillars below are the same eight established at initiation in February, graded against what this quarter's print and call revealed. The status tags are the ones carried in our thesis of record.

Overall: the thesis strengthened. Two of the four bull pillars improved, two of the four bear points were challenged, and the single largest underwriting risk we identified at initiation, an unverifiable back-half acceleration, was substantially dissolved by a first quarter that simply delivered. Set against that, one previously safe pillar deteriorated sharply: cash conversion, which is the metric we chose as the honest valuation denominator. Snowflake is closer to being a compounding company than it was in February and it is not yet demonstrably one.

Action: Maintain Hold. Of the two conditions we set in February, the gross margin condition is met and the guide-raise condition is not, because the raise passed the beat through and left the second half untouched. The stock has meanwhile paid for the operating improvement in a single 13.4% session, moving from 13.9x to 16.1x enterprise value to forward product revenue. Build on weakness toward the low end of the $185 to $225 range. Upgrade to Outperform on a Q2 print that recovers adjusted free cash flow margin above 25% together with a full-year raise larger than the quarter's beat. Downgrade to Underperform if net revenue retention breaks below 124% or the FY26 revenue guide is trimmed at any point in the year.