Built by Grok 4.6 HighSNOW | Earnings Review — FQ2 FY2027
Verdict: ACCELERATING. Product revenue $1,491.9 million (+36.8% YoY vs $1,090.5 million; company 37%) is the third straight step-up after FQ4 FY2026 +30.0% and FQ1 FY2027 +33.9%. Total revenue $1,546.8 million (+35.1%). Non-GAAP operating margin 15% (+400 bps YoY; +250 bps vs a 12.5% guide). Non-GAAP diluted EPS $0.62 (+77% YoY). The mix cost: non-GAAP product GM 75% (−100 bps YoY) and the FY hold cut to 74%.
Print (wide double-beat, not the designed 3% residual). Total vs FMP $1,483.2M = +$63.6M / +4.3%. NG EPS vs FMP $0.4468 = +$0.173 / +38.8%. Product cleared the company’s own FQ1-issued box of $1,415–$1,420 million by +$74.4M / +5.2% vs mid and +5.1% vs the high — a +680 bp overshoot vs the +30% box they set 90 days earlier. Consistent beater: revenue L12 12/12, NG EPS L12 11/12, L4 4/4 both. Dollar residual is expanding; EPS surprise re-widened.
Guidance (second consecutive in-year raise). FY2027 product raised +$230M / +500 bps to $6,070 million / +36% from $5,840 million / +31%. FQ3 product $1,588–$1,593 million is +37–+38% — the first next-quarter growth guide that does not step down versus the just-printed run-rate. nGAAP OM raised +100 bps to 14.5%; FQ3 OM 15.5%. Adj. FCF 23% reiterated. EPS is not guided. Post-print FMP FQ3 total $1,654.5M and FY total $6.321B already sit on the company product number.
Tone. More confident on durability, more honest on mix, same “observed consumption / no change in philosophy” slogan. Sridhar: AI products (CoCo, CoWork, AI functions, Cortex AI Gateway) are ~half of the +700 bp acceleration; the rest is core, migrations, notebooks, Streamlit. Official 8-K: CoCo 9,100 (+2,000 net), CoWork 5,800 (~+11% QoQ) — do not use the FMP name-swap.
Contradictions. The live management contradiction is Q1’s 75% FY product-GM hold (AWS offsets AI) flipping to a 74% FY mix cut. Robins then said “you saw GM go down to 74%” — the quarter still printed 75%. FMP garbled CoCo/CoWork, dropped a digit on 935 hires and 330 GAs — document errors, not CEO vs CFO.
Catalyst into 2 Dec. FQ3 must hold 37–38% without another GM cut. A 5% beat of the new box is ~$1.64B product / ~41% YoY. A meet at $1.591B after two +5% prints is the chronic-beater tape’s miss. Watch NRR, the $10M cohort (only +1 sequential after +8 in FQ1), and RPO after two sequential declines.
| Product revenue | $1,491.9M (+36.8% YoY) · +5.2% vs own mid | Total revenue | $1,546.8M (+35.1%) · +4.3% vs FMP |
| Non-GAAP diluted EPS | $0.62 (+77% YoY) · +38.8% vs FMP $0.45 | nGAAP OM / product GM | 15% (+400 bps YoY) · 75% (−100 bps) |
| NRR / RPO | 126% (held) · $9.0B (+30% YoY, 2nd seq. dip) | Customers / $1M / $10M | 14,554 (+21%) · 828 (+27%) · 65 (+1 seq.) |
| CoCo / CoWork (8-K) | 9,100 (+2,000) · 5,800 (~+11% QoQ) | FY27 product / OM / GM | $6.07B / +36% · 14.5% OM · 74% product GM |
| FQ3 product / OM | $1,588–$1,593M / +37–38% · 15.5% OM | L12 beat rate | Rev 12/12 · NG EPS 11/12 · product vs high 12/12 |
Revenue drivers sit above the consolidated total. YoY is same-fiscal-quarter, never mixed-quarter. Non-GAAP operating income is the operating-profit line Snowflake guides (Daloopa derived EBITDA is unpopulated).
Demand
| Metric | FQ3'25 | FQ4'25 | FQ1'26 | FQ2'26 | FQ3'26 | FQ4'26 | FQ1'27 | FQ2'27 |
|---|---|---|---|---|---|---|---|---|
| Product ($M) | 900.3 | 943.3 | 996.8 | 1,090.5 | 1,158.4 | 1,226.6 | 1,334.3 | 1,491.9 |
| Product YoY | +28.9% | +27.8% | +26.2% trough | +31.5% | +28.7% | +30.0% | +33.9% | +36.8% |
| Total revenue ($M) | 942.1 | 986.8 | 1,042.1 | 1,145.0 | 1,212.9 | 1,284.0 | 1,391.0 | 1,546.8 |
| Total YoY | +28.3% | +27.4% | +25.7% | +31.8% | +28.7% | +30.1% | +33.5% | +35.1% |
| Prof. services ($M) | 41.8 | 43.5 | 45.3 | 54.5 | 54.5 | 57.4 | 56.6 | 54.9 |
| NRR | 127% | 126% | 124% | 125% | 125% | 125% | 126% | 126% |
| RPO ($B) | 5.7 | 6.9 | 6.7 | 6.9 | 7.88 | 9.77 | 9.21 | 9.0 |
| RPO YoY | +54% | +33% | +34% | +33% | +38% | +42% | +38% | +30% |
| Next-12-mo RPO mix | 50% | 48% | 50% | 50% | 48% | 46% | 50% | 54% |
| Customers | 10,618 | 11,159 | 11,578 | 12,062 | 12,621 | 13,328 | 13,912 | 14,554 |
| $1M+ TTM | 542 | 580 | 606 | 654 | 688 | 733 | 779 | 828 |
| $10M+ TTM | 37 | 39 | 40 | 45 | 49 | 56 | 64 | 65 |
| G2K | 754 | 745 | 754 | 751 | 766 | 740 | 813 | 829 |
Driver read. Product YoY bottomed at +26.2% in FQ1 FY2026, bounced, paused in FQ3 FY2026 (+28.7%), then ran 30.0% → 33.9% → 36.8%. That is +1,060 bps off the trough and +530 bps vs year-ago FQ2. Services $54.9 million are +0.8% YoY — the entire dollar beat is product. NRR has printed 126 / 126 after the FY2026 trough; that is stability, not a new regime. The $10M cohort is the tell that stopped compounding sequentially: 40 → 45 → 49 → 56 → 64 → 65 (eight crossings in FQ1, one in FQ2). RPO’s +30% YoY is real but sequential RPO has given back $0.77B of the $9.77B FQ4 peak; cRPO mix actually rose to 54%, so this is Q4-weighted bookings, not a demand break.
Profitability
| Metric | FQ3'25 | FQ4'25 | FQ1'26 | FQ2'26 | FQ3'26 | FQ4'26 | FQ1'27 | FQ2'27 |
|---|---|---|---|---|---|---|---|---|
| nGAAP product GM | 76% | 76% | 76% | 76% | 76% | 75% | 75% | 75% |
| nGAAP op. income ($M) | 58.9 | 92.8 | 91.7 | 127.6 | 131.3 | 139.2 | 165.8 | 237.0 |
| nGAAP OM | 6% | 9% | 9% | 11% | 11% | 11% | 12% | 15% |
| nGAAP diluted EPS | 0.20 | 0.30 | 0.24 | 0.35 | 0.35 | 0.32 | 0.39 | 0.62 |
| GAAP diluted EPS | −0.98 | −0.99 | −1.29 | −0.89 | −0.87 | −0.90 | −0.86 | −0.55 |
| nGAAP adj. FCF ($M) | 86.8 | 423.1 | 206.3 | 67.8 | 136.4 | 782.2 | 265.5 | 92.3 |
| nGAAP FCF margin | 9% | 43% | 20% | 6% | 11% | 61% | 19% | 6% |
| Employees | — | 7,834 | 8,240 | 8,769 | 9,023 | 9,060 | 9,250 | 9,394 |
Two margin stories have split. Operating leverage is the print: nGAAP OM 5% → 9% → 11% → 12% → 15% over eight quarters, and Q3 is guided to 15.5%. Headcount +7.1% against product +36.8% is the mechanism (YTD +334 employees, 173 from Observe, vs official year-ago 935). Gross margin is the cost of the AI mix: product GM rolled off a two-year 76% hold the moment CoCo showed up, and the FY hold was just cut to 74%. Adj. FCF 6% is the seasonal Q2 trough — same as last year; do not annualize.
FMP scores total revenue and non-GAAP diluted EPS. The company guides product revenue and nGAAP operating margin. Those are different bars.
| Metric | FY25Q3 | FY25Q4 | FY26Q1 | FY26Q2 | FY26Q3 | FY26Q4 | FY27Q1 | FY27Q2 |
|---|---|---|---|---|---|---|---|---|
| Revenue vs Street | B +4.8% | B +3.0% | B +3.6% | B +5.2% | B +2.4% | B +2.2% | B +5.1% | B +4.3% |
| NG EPS vs Street | B +33% | B +70% | B +13% | B +31% | B +12% | B +17% | B +22% | B +39% |
| Product vs co. high | AH +5.3% | AH +3.5% | AH +3.8% | AH +4.9% | AH +2.5% | AH +2.2% | AH +5.3% | AH +5.1% |
| NG OM vs co. guide | +300 bps | +500 bps | +400 bps | +300 bps | +200 bps | +400 bps | +300 bps | +250 bps |
B = beat vs FMP Street. AH = above company product-guide high (Q(N) guide applies to Q(N+1) actuals). No Street miss in this L8Q window. Two residual clusters: ~2.2–2.5% when the consumption model lands, ~5% when a new product/AI step-up is not in the snapshot. This print is the wide cluster.
This quarter vs consensus / own box
| Metric | Bar | Actual | Variance | Result |
|---|---|---|---|---|
| Total revenue | $1,483.2M (FMP) | $1,546.8M | +$63.6M / +4.3% | BEAT |
| Product revenue | Co. $1,415–$1,420M (mid $1,417.5M) | $1,491.9M | +$74.4M / +5.2% vs mid; +5.1% vs high | Above high |
| nGAAP diluted EPS | $0.4468 (FMP) | $0.62 | +$0.173 / +38.8% | BEAT |
| nGAAP OM | Co. 12.5% | 15% | +250 bps vs guide; +400 bps YoY | Above guide |
| nGAAP product GM | FY was 75% | 75% | −100 bps YoY vs 76% | Mix cost of the beat |
| GAAP diluted EPS | n/a (FMP prints NG) | −$0.55 | Loss narrowed $0.34 YoY | Not the Street line |
Street was already above the company’s product midpoint on total revenue. Clearing Street is a residual overshoot, not just beating a sandbag. Robins’s own “3% is a really strong beat” bar was ~$1,460M product; they printed $1,492M.
Beat rates
| Window | Rev vs Street | NG EPS vs Street | Product above co. high | Avg rev surprise | Avg product vs mid |
|---|---|---|---|---|---|
| L12Q | 12/12 (100%) | 11/12 (91.7%) | 12/12 | +3.53% / +$36.9M | +4.13% |
| L4Q | 4/4 | 4/4 | 4/4 | +3.49% / +$46.6M | +3.99% |
The only Street miss in 24 prints is FY2025Q1 NG EPS. Current streak: 9 consecutive double-beats. Classification: consistent beater. Revenue percent magnitude is stable in a 2–5% band; dollar residual is expanding. EPS percent compressed for two years, then re-widened this print.
Attribution of this quarter’s beat. (1) Product consumption ran $74M / +5.2% above the $1,417.5M mid, taking growth from guided +30% to printed +37%. (2) Management splits that +700 bp acceleration ~50/50 first-party AI vs core + faster migrations + notebooks/Streamlit. (3) +250 bps of NG OM vs a 12.5% guide, from revenue overshoot plus headcount discipline, converting a +4.3% sales beat into a +38.8% NG EPS beat. (4) No one-time P&L item — GAAP EPS is below NG. (5) Observe remains ~1 ppt of FY product growth — not the residual. (6) Product GM is the quality cost.
stable/earnings lastUpdated 2026-09-21. Guide offset: Q(N) product/OM guide applies to Q(N+1).Snowflake guides product revenue (not total) and nGAAP operating margin one quarter forward, plus a full-year product / product-GM / OM / adj.-FCF / diluted-share package. It does not guide EPS, NRR, or bookings. Quarterly Daloopa guidance series are tagged to the call quarter.
Headline: Second consecutive in-year raise. FQ3 growth guide sits on top of the +36.8% print, not 400 bps under it. Last quarter they guided +30% after a +34% print. They stopped doing that.
New guide (issued 2 Sep 2026)
| Metric | New guide | Implied YoY | vs prior (FQ1 call) |
|---|---|---|---|
| FQ3 product | $1,588–$1,593M; mid $1,590.5M | +37–+38% vs $1,158.4M | First Q3 guide. Old-FY remainder reconstr. ~$1,464M |
| FQ3 nGAAP OM | 15.5% | +450 bps vs FQ3'26 11% | First Q3 OM guide; +50 bps vs FQ2 print |
| FY2027 product | $6,070M / +36% | vs FY26 $4,472.3M | +$230M / +500 bps vs $5,840M / +31% |
| FY nGAAP product GM | 74% | vs printed 75% | −100 bps vs 75% |
| FY nGAAP OM | 14.5% | +450 bps vs FY26 10% | +100 bps vs 13.5% |
| FY nGAAP adj. FCF | 23% | seasonal | Reiterated |
| FY nGAAP diluted shares | 380M | — | +4M vs FQ1 8-K 376M |
| Observe | ~1 ppt of FY product growth | Unchanged | Unchanged |
| EPS / total revenue | Not guided | FMP FQ3 $1,654.5M / $0.60; FY $6,321M / $2.22 | FMP ~0.5% above implied total |
FY product waterfall — three in-year steps
Source: Daloopa yearly product guidance (ids 169177693, 179159190). FMP FY2027E total $6,320.7M less ~$220M services. New guide is ~$31M / 0.5% light of FMP-implied product.
H2 math at the new year number. YTD product $2,826.2M. Remaining FQ3+FQ4 = $3,243.8M vs year-ago $2,385.0M → +36.0% H2. Subtract FQ3 mid $1,590.5M → implied FQ4 $1,653.3M / +34.8%. A 3% FQ3 beat ($1,638M) would leave FQ4 at $1,606M / +30.9% if the year is a ceiling. History says the year is a floor.
OM is still the conservative sandwich. H1 nGAAP OM already ~13.6%. FQ3 guided 15.5%. A 14.5% full-year needs H2 only ~15.3% — Q4 can decelerate from Q3 and they still make the year.
Classification
| Item | Stance | Why |
|---|---|---|
| FQ3 product growth rate | In-line to slightly aggressive vs run-rate | First next-quarter growth guide that does not step down |
| FQ3 product dollars | Still conservative if 3% philosophy holds | 3% beat = $1,638M / +41% YoY |
| FY2027 product | Aggressive vs prior; slight sandbag vs FMP | +$230M / +500 bps vs FQ1; ~0.5% light of FMP-implied |
| FY nGAAP OM | Raised, still conservative | 14.5% after a 15% print and a 15.5% Q3 guide |
| FY nGAAP product GM | The cut is real, not sandbag | 74% vs 75% printed this quarter |
Total revenue YoY reaccelerated for a third straight quarter to +35.1% (+162 bps of Accel; +937 bps above the FQ1 FY2026 trough). Product — 96% of the print — at +36.8%. nGAAP EPS +77% to $0.62.
Inflection map.
- FQ3 FY2025–FQ1 FY2026: orderly deceleration into the trough. Total +28.3% → +25.7%. Product to +26.2%.
- FQ2 FY2026: first reacceleration (+604 bps) — not yet durable. Total +31.8%. Gave back 304 bps the next quarter.
- FQ3 FY2026: one-quarter pause to +28.7%. Why FQ4 is the true start of the current streak, not FQ2.
- FQ4 FY2026: streak begins. Product +30%. Robins: “stable growth in our core business and a step-up in growth contribution from AI workloads.”
- FQ1 FY2027: +336 bps to +33.5%; product +33.9%. CoCo was “the largest driver to the increase in our [FY] forecast.”
- FQ2 FY2027 (this print): third straight; 8-quarter high. Product +36.8% / +295 bps. Sridhar: “After exiting 30% year over year growth, we have now added 7 points of acceleration in just 2 quarters.” Split now ~50/50 AI vs core — no longer CoCo-only.
EPS trajectory is healthy in dollars, noisy at the second derivative. nGAAP EPS went $0.20 → $0.62 over eight quarters. The Accel swings are small-base math ($0.14 and $0.30 year-ago prints), not operating chaos. GAAP still a loss but narrowing (−$0.55 vs −$0.89). Operating leverage is doing more work than revenue mix is undoing.
Annual still looks like a stall: FY2025–FY2026 product growth ~29% for two years. FQ2 FY2027 at +37% is the first clean break above that FY2026 run-rate, which is why the year was walked from 27% → 31% → 36%.
The 2 Sep print retired the “guided deceleration to +30%” setup. The live debate into 2 Dec 2026 is whether FQ3 can hold 37–38% without a second product-GM cut. Post-print FMP is already on the company number.
| # | Catalyst | Timing | Street bar | Implication |
|---|---|---|---|---|
| 1 | FQ3 product + first look at FQ4 | 2 Dec 2026 | FMP total $1,654.5M / $0.60. Implied product ~$1,599M — at/just above the $1,593M high, not a 5% beat | A $1,590M product print is the 37–38% hold and will screen as in-line. A 5% beat is ~$1,670M / ~41% — the path to a second FY raise. Printing 37% “as guided” after two 5% beats is the chronic-beater tape’s miss |
| 2 | FY product hold $6.07B / 36% or raise again | Scored 2 Dec | FMP FY total $6.321B / $2.22 is company product plus ~$250M services | Holding $6.07B after a 37% Q3 is the base tape. A Q3 beat that does not come with a FY raise is a Q4-fade admission (implied Q4 +34.8%) |
| 3 | AI mix vs product GM | Every print | Only published bar is FY nGAAP product GM 74% | Quantified “AI is X% of product” that does not force another GM cut is the bull surprise. Another 100 bp cut (73%) is the bear. Robins refused the BofA uplift-vs-pull-forward question |
| 4 | NRR and the $10M cohort | FQ3 print | Street treats 126% as the floor that justifies 36% FY | NRR ≥127% or another high-single-digit $10M crossing is the unmodelled positive. A second +1 $10M quarter says the up-market motion is saturating — Databricks just put &;100 $10M customers on the tape |
| 5 | RPO / Q4-weighted bookings | FQ3 is the third sequential test | Holding ~$9.0B is the bar | A third sequential down RPO print is the “bookings were a one-off” tape even if product holds 37% |
| 6 | nGAAP OM vs the 74% GM cut | FQ3 guided 15.5%; FY 14.5% | FMP FQ3 EPS $0.60 is the 15.5% OM quarter | They are paying for AI mix with opex, not COGS. A 15.5% Q3 OM with product GM still 75% confirms the offset |
| 7 | Databricks / share | Continuous | Databricks last print (13 Aug): &;$7B RTR, &;80% YoY | A 37% SNOW print next to an 80% Databricks print is still the “losing the AI layer” tape even if SNOW beats its own guide |
| 8 | CoCo / CoWork intensity | FQ3 Q&A | Official: CoCo 9,100 / CoWork 5,800 | Users/account, $/account, or the “time-to-80%-of-purchased-consumption” KPI as a published number. Silence is the miss |
| 9 | MongoDB analog (already printed) | 1 Sep | Atlas +29%; Q3 Atlas ~26% | FQ3 comparison is SNOW guided 37–38% vs MDB Atlas guided ~26%. If SNOW holds 37% after Atlas decelerates, consumption software is not one tape |
Bull print on 2 Dec. Product ≥$1,670M and a FY raise above $6.07B / 36%, with NRR ≥126% and the $10M cohort resuming high-single-digit crossings.
Base / in-line. Product inside $1,588–$1,610M, total around FMP $1,654M, OM at 15.5%, FY held. Against two 5% guide beats, it will be sold as a miss by anyone who faded the new 37% guide the same way they faded the old 30% guide.
Bear print. Product <$1,588M or YoY <36%, NRR back to 125%, $10M still +1, or a GM print that forces them off 74%.
Call 2026-09-02. Operator hard-limited to one question. 12 analysts. Tally: 9 well-answered (3 of those soft) · 3 deflected.
| # | Analyst | Topic | Badge |
|---|---|---|---|
| 1 | Sanjit Singh — MS | Quality of acceleration; CoCo right-to-win; consumption hygiene | Well answered |
| 2 | Kirk Materne — Evercore | Split of acceleration: AI vs core flywheel | Well answered — the print-defining line |
| 3 | Karl Keirstead — UBS | Model neutrality / mix / GM vs Frontier Labs | Well answered |
| 4 | Raimo Lenschow — Barclays | Geographic mix (Europe / Asia) | Deflected |
| 5 | Ryan MacWilliams — Wells | Why AI is accelerating SNOW now | Well answered |
| 6 | Matthew Hedberg — RBC | CoCo depth; CoWork as sandbox → turnkey | Well answered (soft on sandbox) |
| 7 | Koji Ikeda — BofA | AI-adopter vs non-adopter consumption uplift | Deflected |
| 8 | Brent Thill — Jefferies | CoWork land size / consumption intensity | Deflected |
| 9 | Dan for Brad Zelnick — DB | Arctic / own models as a third leg | Well answered |
| 10 | Aleksandr Zukin — Wolfe | Budget ambit / Frontier / outcome-based | Well answered (soft on $) |
| 11 | Tyler Radke — Citi | SaaS+LLM becoming databases | Well answered |
| 12 | James Dinkel — JPM | Parse FY raise: core vs AI vs last quarter’s CoCo-heavy raise | Well answered (soft on $ split) |
Q2 (Materne) — the number they did give. Sridhar: AI products “contributed approximately half of the acceleration.” Remainder: notebooks, Streamlit/React, migrations running faster (a Teradata migration in <3 quarters that used to take 2–3 years). Street can now model the 700 bp overshoot vs the old +30% box as ~half AI attach, ~half core.
Q3 (Keirstead) — the P&L tell. Robins: “You saw our non-GAAP product gross margin go down to 74%. Because we have increased our guidance so much.” Product-first, adoption-second, margins third. “Give our customers the best answer… then we will continue to work on margins as we go forward.” Quarterly GM still printed 75%; the 74% is the FY guide.
Three deflections the Street still needs.
| Analyst | Asked | Got | Why it matters |
|---|---|---|---|
| Lenschow | Europe / Asia vs US index | “All regions are performing.” | 10-Q already shows EMEA $258.7M / +39.6% and APJ $94.9M / +40.3% vs Americas ~+33.7%. He had a good story and chose not to tell it |
| Ikeda | AI-adopter vs non-adopter uplift; pull-forward vs LTV | “We are not ready to share the exact uplift numbers.” | Highest-value unanswered question. Without it the Street cannot separate “Cocoa is 50% of the acceleration” from “Cocoa users pulled 2027 work into 2026” |
| Thill | Bigger CoWork lands — users, consumption out of the gate | CoCo GTM process. “Easiest sales.” | No ACV, no users/account, no $/account |
Watch on the 2 Dec call: (1) will they share the uplift they “are not ready” for; (2) any CoCo/CoWork intensity metric; (3) whether they finally say the geo numbers already in the 10-Q.
Checked and not contradictions: “second consecutive quarter of record sequential dollar growth” vs “third straight quarter of acceleration” (different metrics). “No change in forecast methodology” vs FY 27% → 31% → 36% (raises on observed consumption, which is the stated method). Observe ~1 ppt held. FCF 23% held. GAAP profitability still Q4 FY2028. Sequential RPO dip vs YoY +30% and Q4-weighted bookings language.
No rates, CPI, consumer, or industrial-cycle color. The macro that is in the call is enterprise IT demand, regional spend, and AI-capex mix. Robins: “strong demand environment.” Sridhar: acceleration is from a “very broad swath of customers,” not concentrated in AI-native companies, which remain “a small part of our overall revenue stream.” That is the anti-2021-froth check.
| Relationship | Name | Datapoint | Implication | Confidence |
|---|---|---|---|---|
| Competitor (legacy) | Teradata (TDC) | Network-equipment OEM doing a Teradata migration in &;3 quarters that used to take 2–3 years. Ramaswamy has put AI-compressed migrations on six straight calls | Worse for TDC. Cleanest competitive read-through of the print. Cycle-time collapse is the bear case | Direct / high |
| Supplier | AMZN AWS / MSFT Azure / GOOGL GCP | Accrued third-party cloud infra $168.8M vs $101.8M (+65.8% YoY) against product +36.8% | Better near-term for hyperscalers — AI mix is a cloud-consumption multiplier. Medium-term SNOW is pushing open-model self-hosting to claw GM back | Direct / high |
| Frontier labs | OpenAI / Anthropic | Cost “absolutely become a concern.” Both labs standing up own capacity and landing on other clouds. Customers regret single-model commits | Better for SNOW model-neutrality / Cortex AI Gateway. Worse for exclusive lab commits. Frontier price cuts are a GM tailwind for SNOW | Direct / high |
| Competitor (implied) | Databricks | Not named this print. FQ1: CoCo talks to Glue, Airflow, “and in fact even Databricks” | Mixed. Do not treat this print as a Databricks-miss confirmation | Implied / moderate |
| SaaS → database | CRM / NOW / WDAY (inferred) | Radke asked; Sridhar: owning the UX is critical; Kleinerman: CIOs “not going to copy my data into 3 different platforms” | Worse for SaaS vendors trying to become the SoR via an LLM UI. Hybrid tables / Postgres = long-dated wedge vs standalone OLTP | Direct on strategy |
| Customers named | BLK, XYZ, RDDT, SNY, Indeed, Sayari, 1Password, a top Australian bank | CoCo / CoWork / sharing / Frontier (outcome-based: “we get paid only when we deliver”) | Enterprise AI demand is broad G2K, not AI-native. EMEA +39.6% and APJ +40.3% say this is not a US-only CIO fad | Direct / high |
| Action-layer partners | CRM (Slack) / TEAM (Jira) via Natoma | Cortex AI Gateway: summarize Slack, open Jira tickets, without leaving CoCo/CoWork | Coexistence: SNOW capturing the control plane. Natoma closed 3 Jun (683,356 shares) | Direct / moderate |
| Observability | Observe (acquired) | Still ~1 ppt of FY product growth, 173 of the YTD headcount adds | Neutral-to-slightly worse for DDOG only if Observe is winning the same workloads — management did not claim that | Direct on contribution |
Top 3. (1) TDC is the cleanest loser. (2) Hyperscalers collect the COGS — cloud infra +66% vs product +37%; FY GM 74% is explicit. (3) Named F100 demand is broad G2K, not AI-native — BlackRock, Block, Reddit, Indeed, Sanofi; NRR 126%; G2K 829; international outgrowing Americas.
- FQ3 is a hold-the-rate test, not another step-up. FMP is already on $1,654.5M. A meet at the $1,590.5M product mid after two +5% prints is the setup the Street will sell.
- The 3% product-beat policy did not hold this quarter even with CoCo “in the model.” If Q3 prints another +5% vs the 37–38% box, the model is still late; if it prints +2–3%, they finally caught the run-rate.
- GM mix. 75% printed, 74% guided. If CoCo/CoWork mix keeps rising and open-source / auto-routing lag, 74% becomes 73% and the OM raise is the thing that breaks, not the revenue raise.
- $10M cohort stalled (+1 sequential). Compatible with still-healthy YoY (+44%) and a saturating up-market motion. A second +1 quarter vs Databricks’ >100 $10M customers is the quality tell.
- They will not size the multiplier. Until FQ3, the model has a 50/50 acceleration split and a 100 bp mix tax — and nothing on unit economics of the AI attach.
HOLD. Own the acceleration; do not underwrite another 5% residual against a guide that is already at the print. The 2 Dec number that matters is product vs $1,588–$1,593M and whether FY $6.07B / 36% is a floor or a ceiling.