Built by Grok 4.6 High

SNOW | Earnings Preview

FQ2 FY2027 (three months ended July 31) has not printed. Snowflake reports Wednesday, September 2, 2026, after the close; the call is 2:00 p.m. PT. Last-reported baseline is FQ1 FY2027 (May 27): product revenue $1,334.3 million (+33.9% YoY vs $996.8 million), total revenue $1,391.0 million (+33.5%), NRR 126% (+200 bps), non-GAAP operating margin 12% (+300 bps), non-GAAP EPS $0.39 (+62.5%). That print cleared the then-live product high of $1,267 million by +$67.3 million / +5.3% and raised FY2027. The raise is still in force. Do not treat this quarter as reported.

Growth trajectory — re-accelerated, then guided to slow. Product YoY trough-ed at +26% in FQ1 FY2026, re-accelerated to +34% in FQ1 FY2027, and the live Q2 guide itself decelerates 400 bps back to +30%. That is the sandbag after Cortex Code (CoCo) was finally put in the observed-behavior model — not a demand break the peer tape supports.

Four-quarter product run: +31.5% → +28.7% → +30.0% → +33.9%. The Q2 pin of +30% is a −390 bp deceleration vs FQ1 and −150 bps vs last year’s same-quarter actual. Company-level Street +29.5% is the official product mid plus a services plug — not a 3% beat.

Watch items into Wednesday

• Guidance / the real bar. Live Q2 product is $1,415–$1,420 million (+30% YoY), non-GAAP operating margin 12.5%. FMP $1,482.2 million / $0.45 sits on the product box once a normal ~$57 million services residual is added. A 12-for-12 beat of the company product high, with last-four average +3.9% vs midpoint, puts the historical-pattern whisper near $1,473 million product. Robins’s own “3% is a very solid beat” bar is ~$1,460 million. An in-line-to-guide print is a beat on the official scorecard and a miss versus both the whisper and the stated philosophy. The FY action (hold vs lift the $5.84 billion / +31% book) is the second event.
• Tariffs. Not a live item. Snowflake is a consumption-cloud / software name with no goods-cost stack. The analog risk is CoCo token-throttling and AI mix vs the 75% product-GM pin, not duty rates. Nothing in the May–August IR trail revises the “observed behavior” assumption.
• Product YoY vs the 30% pin. ≥32% (≈$1,440 million+) keeps the acceleration narrative intact. Printing the +30% mid is the first deceleration in five quarters. ≤29% puts the FY 31% book at risk.
• CoCo is now in the model. Q1’s +5.5% beat was the exception because CoCo had zero observed history. Robins layered one quarter of CoCo into the remainder of FY2027 and then capped beat convexity at 3%. A second +5% beat would mean another unmodeled AI ramp — or a wider cushion than advertised.

Peer tape into 9/2 is already printed and fast: DDOG +36%, NET +36%, NOW +24.5% (AI ACV $1B), TWLO +17% organic. There is no demand-side excuse if product comes in light of +30%. CRWD / CRM (8/26) and ESTC (8/27) can still change the setup; MDB prints the night before (9/1). Investor Day already dated GAAP profit to FQ4 FY2028 without lifting the Q2 or FY pins.

Issued May 27 on the FQ1 FY2027 call. Never revised. Snowflake guides product revenue + non-GAAP operating margin one quarter forward, plus a full-year product / product-GM / op-margin / adj.-FCF package. It does not guide EPS, total revenue, NRR, RPO, or bookings. Classification: CONSERVATIVE versus the +34% run-rate, versus last year’s FQ2 actual growth, and versus FMP total-revenue consensus. The FY raise (27% → 31%) was the aggressive action; the FQ2 print itself is still set to the 3% cushion.

Verbatim (CFO Brian Robins, FQ1 FY2027 call, May 27): “For FY ’27, we now expect product revenue of $5.84 billion, representing 31% year-over-year growth. In Q2, we expect product revenue between $1.415 and $1.42 billion, representing 30% year-over-year growth.” And, twice: “there’s no change to our guidance philosophy, where 3% we view as a really strong beat.” (Daloopa transcript 27028186)

Same-quarter last year. Last year’s Q2 guide was $1,035–$1,040 million (+25%) and printed $1,090.5 million (+5.1% vs mid / +31.5% YoY). This year’s growth-rate pin is ~500 bps more aggressive and essentially matches last year’s actual. They are no longer pretending Q2 is a mid-20s quarter. They are still guiding a 400 bp deceleration off FQ1 and telling the Street not to extrapolate the +5.5% CoCo analog.

If the 3% philosophy holds: midpoint $1,417.5 million × 1.03 ≈ $1,460 million product, which would be +33.9% YoY — a flat hold of the FQ1 growth rate, not a deceleration. That is the number the beat-history says to model. Street total $1,482 million would then be a ~2.5% total-revenue beat once services are added.

Remaining-year arithmetic (not company guidance). Q1 actual $1,334.3 million + Q2 guide mid $1,417.5 million = $2,751.8 million in H1. Against the FY pin of $5.84 billion, that leaves $3,088 million for H2 vs last year H2 $1,158.4 million + $1,226.6 million = $2,385.0 million → H2 +29.5%. A 3% Q2 beat with no FY raise would drop implied H2 to ~+27.7% — a deceleration they have not been willing to advertise. Base case into the call is another FY product raise (last year’s Q2 raise was +$70 million). A raise to ~$5.90–$5.95 billion would keep H2 near +30% and is the number that matters more than the Q2 print.

Annual trajectory (fiscal years ended January 31). Product growth trough-ed at +29% in FY2025–FY2026 after a +38% FY2024, and is being guided back up 140 bps to +31% in FY2027. That is a re-acceleration year on the official book, paid for by CoCo + core, not a peak-and-fade.

YoY only. Snowflake is a consumption / product-revenue platform — the driver stack sits above consolidated revenue (product, then a thin professional-services stub). Internal estimates are n/a.

(a) Current quarter — FQ2 FY2027 (upcoming)

FMP does not publish a product-revenue estimate. A $70 million FMP services residual would be a mix illusion — Q2 last year professional services were $54.5 million and Q1 this year $56.6 million. Street total $1,482.2 million is essentially the official product mid plus a normal services plug.

Last-reported mix (FQ1 FY2027) — the baseline this print will be read against

Q1 product printed +33.9% against a +27% guide — the same ~400 bp overshoot now being guided back out of Q2. CoCo was “the largest driver to the increase in our forecast” because they only forecast observed behavior.

How to read the print

• Bullish: product ≥ $1,460 million (+3% vs mid, +33.9% YoY) and nGAAP OM ≥ 13.0%. That keeps +34% Q1 from looking like a one-quarter CoCo spike and funds a FY raise.
• In-line / soft: product $1,418–$1,445 million (clears the official range, misses the 3% philosophy). Street $1,482 total can print as a “beat” on FMP’s consolidated number while product is only +0–2% vs guide.
• Bearish: product ≤ $1,417.5 million or YoY ≤ +30%, NRR back to 125% or below, or FY product held at $5.84 billion after a 3%+ Q2 beat (that would dump the beat into a slower H2).

(b) FQ+1 — FQ3 FY2027 (expected; will be issued Sep 2)

The FQ3 FY2027 guide does not exist. Italic = expected from the still-in-force FY product envelope, last year’s Q2→Q3 sequential sandbag, and FMP FQ3 consensus — not issued.

Method: FY product $5.84 billion − Q1 actual $1,334.3 million − Q2 guide mid $1,417.5 million leaves $3,088 million for H2. Last year’s H2 split was 48.6% in Q3 → ~$1,500 million Q3 product if they do not raise FY. Applying ~+29–30% to last year’s Q3 $1,158.4 million also lands at $1,490–$1,506 million.

Hitting FY 13.5% with Q1 already at 12% and Q2 guided 12.5% (Summit) requires an H2 average near 14.8%. A Q3 guide below 13% against a 13.5% FY number would be another sandbag, not a cut. A 3%+ Q2 product beat that is not followed by a FY raise would force Q3+Q4 down (bearish). A beat recycled into a FY raise to ~$5.90–$5.95 billion would push the expected Q3 product guide toward $1,520–$1,540 million.

(c) FY+1 — FY2028 (unguided) and the standing FY2027 book

Snowflake does not issue FY+1 (FY2028) guidance. It typically only opens the next fiscal year on the Q4 call. Investor Day (June 2) dated GAAP profit to FQ4 FY2028 without lifting FY2027 product.

FY2028 vs FY2027 street: revenue +25.8%, EPS +41.1% — a deceleration from FY2027’s +30.2% / +54.4% versus FY2026 actuals. Street is 30 bps below implied FY product. Prior FY2027 guide, issued on the FQ4 FY26 call (Feb 25): product $5.66 billion / +27%; nGAAP OM 12.5%. The May 27 raise was +$180 million / +400 bps of growth and +100 bps on operating margin.

Observe / Natoma inside the old guide: ~1 ppt of the +31% is Observe (organic ~+30%). The 13.5% / 23% margin and FCF guides still embed ~150 bps of Observe dilution. Natoma (20 employees) is not in the numbers as a closed deal.

Tone into the quarter

Management walks into September 2 having raised the FY2027 book on May 27 and having not updated those pins since. No 8-K, no pre-announce. Conviction on CoCo + core as a durable FY 31% book is up versus the conservative February initiation; conviction that Q2 will crush the pin is what they just took away.

Conviction read: The Q2 product pin is a floor they want treated as the forecast, not a sandbag they want the Street to fade by another 400 bps of growth. +30% after a +34% print is the number they will be judged on. The FY 31% book is a beat-and-raise that still leaves 2H consumption open.

Assumptions still live

• Forecast = observed consumption only. “As always, our forecast is based on existing consumption patterns.” This is the religion.
• A 3% beat is “very solid.” Repeated to Singh and Lenschow. Q1’s ~5.5% was the exception because CoCo had zero observed history in the Feb book.
• CoCo is now in the FY model. GA Feb 5; “largest driver to the increase in our forecast.” One quarter of observed behavior, layered through the remainder of the year. A CoCo deceleration vs the Q1 run-rate is the live risk.
• Core data platform also accelerated, and that is in the FY book too. Dual-engine is the 31% story.
• Observe ~1 ppt of FY product growth; ~150 bps OM / FCF headwind. Unchanged from February. Not a 2 Sep swing.
• AI products have lower GM; 75% product GM is held via AWS bandwidth. New 5-year $6 billion AWS contract is “fully incorporated.”
• Customers will not wholesale-throttle CoCo. Value (10× faster projects) plus new cost-governance (account / agent / user token limits). The Jul 28 Horizon cost-management PR is the product proof they are building the throttle before customers demand it.
• Bookings remain Q4-weighted. RPO $9.21 billion is already off the $9.77 billion Q4 spike; that is the model, not a demand miss.
• Organic hiring stays tight. 17 organic adds in Q1; AI is substituting for headcount. This is how 13.5% OM and the FQ4 FY28 GAAP-profit date both work.
• No GTM reset under new CRO J.B. Investor Day: “No changes to sales compensation plans.”
• Natoma (20 employees) is intended, not closed. Extends the agent control plane into SaaS actions. Live as a product narrative; not in the Q2 pin.

Tone arc — adjectives up on the year, down on the Q2 beat

Net shift Q4 → Q1: one-notch up on the annual number (product +400 bps, OM +100 bps, AI now a named P&L engine), one-notch down on implied Q2 beat convexity. The bull case they want believed is “CoCo + core is a durable 31% compounder and the agentic control plane is real.” The thing they do not want believed is “every quarter beats by 5%.”

Post-guidance updates (May 27 → Aug 17)

No formal guidance revision. No pre-announce. No 8-K changing the May 27 ranges. Daloopa has no calendar-2026Q2 / 2026Q3 documents. The book above is still the book.

No management commentary at a sell-side conference between 27 May and 17 Aug is on the IR archive. The next scheduled IR event is the 2 Sep call itself.

What Wednesday has to do

1. Product YoY vs the ~30% pin. ≥34% again = they sandbagged the thing they said they wouldn’t. 29–31% = they meant it. <28% = the Q1 inflection was the peak, and the FY 31% range is live risk.
2. Do they hold, raise, or “roll the beat” the FY book? A raise after a merely in-line Q2 is the bull case their Q1 language set up. A hold is the base case they telegraphed. A cut would be a break with the May raise and the June GAAP-profit date.
3. CoCo / CoWork language vs the P&L. “Meaningful, layered into the year” is consistent. Any claim that agentic is now the majority of growth, without product accelerating through 30%, is a narrative/P&L split. Ask for CoCo account count (was >7,100).
4. NRR and $1 million / $10 million logos. 126% / 779 / 64 is the expansion proof. A NRR fade with product still at 30% says new logos, not expansion, are carrying growth.
5. RPO YoY. $9.21 billion / +37.5% is the post-Q4-spike base. Sequential softness is the model. YoY through +38% is the tell.
6. Gross margin and the 75% FY pin. A Q2 mix giveback they cannot offset with AWS bandwidth is the first real hit to 13.5%.
7. Guidance-philosophy sentence. If Robins drops “3% is a strong beat,” they are inviting another 5% print. If he repeats it, an in-line Q2 is a hit.
8. J.B. / Benoit. Any “transition” hedge on bookings or product velocity is new. ID said there isn’t one.

Ranked by information value on September 2. “Consensus” = FMP / last primary management comment. Bloomberg and Visible Alpha are not connected.

Scorecard on the 2nd

Where management and the Street actually disagree (investing-principles: the only contrarian setup is a credible team repeating something bullish the Street will not underwrite):

1. A 3% beat is the success case, not a miss. Twelve straight product prints above the high end, L12 average +3.93% vs mid. Robins is asking the Street to treat +3% as “really strong.” FMP is already on the box, so the leftover edge is the stated residual, not another CoCo-sized air-pocket.
2. CoCo + core is a durable 31% compounder, not a launch-quarter spike. Ramaswamy has been repeating “agentic control plane” for two calls; the P&L is still core consumption plus a new, lower-GM AI overlay. The Street paid up for an AI re-rate after FQ1 (+36.5% gap); an FQ2 that is “just” 30% product with AI still “layered in” is the gap.
3. The 34% product run is not the run-rate they will underwrite. Robins is asking the Street to accept 30% as honest now that CoCo is in the model. The print resolves whether the Street was right to fade 34% into 30%, or whether management is still sandbagging a consumption business that has not rolled over.

Window: 27 May 2026 (Q1 print) through 17 Aug 2026. Company and brand news excluding the Q1 print itself and the 3 Aug call notice. The window is constructive and unusually clean: same-day as the print, management dropped a five-year $6 billion AWS commit and intent to acquire Natoma; everything since has been on-thesis execution of the agentic control plane with no guidance revision, no outage, no C-suite surprise.

What this does / does not change

• Does not change Q2 FY27 guide math. Company already guided $1.415–$1.420 billion with product +30%. Nothing above is a Q2 shipment beat in hand. Gateway is still preview.
• Does confirm the Q1 agentic-control-plane thesis is being executed. Summit packaged CoWork / CoCo / Horizon / Iceberg v3; Jul 28 started to ship Natoma as Cortex AI Gateway. Both are consistent with “leaders remain leaders” in governed enterprise data.
• Does not invert sentiment. Street enthusiasm after the Q1 beat-and-raise is already robust (FMP grades: Buy 42 / Hold 9 / Sell 1). This is confirmatory, not a worse→better inversion. The risk is the opposite: the May 27 print bought a +36.5% gap (Daloopa $175.26 → $239.20 on May 28) that the news flow has only reinforced.
• Governance is clean. No C-suite turnover, no outage, no activist letter, no contested AGM item. Benoit Dageville stepping away from day-to-day mid-June (remains on the board) and new CRO Jonathan Boulier executed on schedule.

SNOW is a Consistent beater. Versus FMP street it is 12 / 12 on total revenue and 11 / 12 on non-GAAP EPS over the last 12 printed quarters, and 4 / 4 on both lines over the last 4. Versus the high end of its own product box (guide issued in Q(N), scored on Q(N+1)) it is 12 / 12. The single blemish is FY2025Q1 EPS: $0.14 vs FMP $0.1785 (−21.6% / MISS) in a quarter that beat revenue by +5.2%. Revenue beat magnitude is stable-to-improving (L4 average +3.7% / +$45M vs first-4 +2.9% / +$22M). EPS beat magnitude is deteriorating as Street has marked up (L4 average +20.7% vs first-4 +70.4%). The residual on product vs the company box is not luck — it is policy. L12 average vs the product midpoint is +3.93%.

Versus own product guide (the more informative bar). Offset: guide issued in Q(N) applies to Q(N+1). Last four:

The two “normal” Robins quarters (FY26 Q3–Q4) printed +2.4% to +2.7% — exactly the 3% slogan. The two outliers (FY26 Q2 +5.1%, FY27 Q1 +5.5%) were the quarters management later attributed to lumpy migrations and to CoCo not being in the model. Five-for-five recent OM beats, average +320 bps.

What they are telling the Street not to do into FQ2. Robins, answering Barclays on the FQ1 call: no change in philosophy; 3% is a very solid beat; CoCo is now in the model; core acceleration is also in the FY. Combined with Sridhar’s FQ3 “we calibrate back to 3%” rule, that is management trying to talk the Street out of extrapolating the +5.5% CoCo analog. The FY product guide was raised from $5.66 billion / +27% to $5.84 billion / +31% — they rolled the CoCo run-rate into the year rather than leaving it as quarterly surprise.

Bar into Wednesday

Two whispers, one print. The historical-pattern whisper (~$1,473 million product) is what three years of 100% product beats trained the tape to expect. The new-philosophy whisper is Street $1,482 total plus a token residual. FQ2 is the test of whether anyone believed him. Directionally, another beat vs the product high is still the higher-probability outcome. The magnitude the last print trained the tape on (+5.1% / +36.5% gap) is the wrong analog.

Last-print +1-day reaction (AMC convention, Daloopa OHLCV): +36.5% ($175.26 on 2026-05-27 → $239.20 on 2026-05-28) on more than 2× print-day volume. Do not call the Wednesday −1.3% the earnings reaction — that session closed before the AMC release. OptionsLam weekly implied move into this print is ±14.85% (expires Sep 4); monthly ±16.29% (expires Sep 18). The options market is pricing a much smaller move than the last realized close-to-close.

Do not treat FY2027Q2 as printed.