SNOW · Forward model · Platform consumption (product revenue) · Bear case
What has to happen in Platform consumption (product revenue)
Model as of
This page changes Platform consumption (product revenue) inside the complete SNOW model. Every other vertical, corporate programme and valuation assumption stays in the selected portfolio case.
Shares this vertical and portfolio case. Slider and horizon edits stay in your browser.
Platform consumption (product revenue)
Basis quarter$1.49B
Final quarter$3.99B
Implied CAGR+22%
Final revenue mix99%
Snowflake bills consumption of compute, storage and data transfer, not seats, so product revenue is the disclosed customer count multiplied by a derived product revenue per customer. Expansion inside the installed base is the dominant half: net revenue retention was 126% in the basis quarter and 779 customers now spend more than $1 million on a trailing twelve-month basis, up 29% year over year, so the spend is concentrated in a widening large-account tier rather than spread thinly. Product revenue is 96% of total revenue.
Last four quarters
2026 Q3
$1.16B
Reported
2026 Q4
$1.23B
Reported
2027 Q1
$1.33B
Reported
2027 Q2
$1.49B
Reported
Compute consumptionStorageData transferFirst-party AI products (Cortex, Snowflake Intelligence, Cortex Code) - bundled inside product revenue, not separately disclosed
Subscribers
15K
48.5% of a 30K addressable base
14,554 total customers, the figure Ramaswamy gave on the Q3 call. The release itself publishes only net adds, not the count.
Addressable subscribers
30K
the S-curve ceiling
Assumed. Snowflake discloses no addressable market; 30,000 is about 2.2x the base, set off the eight-quarter add rate.
Net adds
642/qtr
ramping toward 1K/qtr, throttled as the base approaches the TAM
642 a quarter, the observed change in the disclosed count. The company's own 'net new customers' metric is 692; the count absorbs churn.
Net-add ceiling
1K/qtr
what supply can deliver at full rate
Add capacity thins toward 1,000 as the large-enterprise pool empties; 813 of the Forbes Global 2000 are already on.
ARPU
$34,168.41/mo
drifting +4.2% per quarter, floor $0.00
$34,168 a month, or $102,505 a quarter: product revenue of $1,491.9M over 14,554 customers. Derived.
Non-subscriber revenue
$0/qtr
growing 0.0% per quarter
Zero. Everything outside product revenue sits in the professional services vertical, which is the only other line reported.
Platform consumption (product revenue)
Latest: $3.99B (2032Q2E)
| Period | Value |
|---|---|
| 2025Q1 | $790M |
| 2025Q2 | $829M |
| 2025Q3 | $900M |
| 2025Q4 | $943M |
| 2026Q1 | $997M |
| 2026Q2 | $1.09B |
| 2026Q3 | $1.16B |
| 2026Q4 | $1.23B |
| 2027Q1 | $1.33B |
| 2027Q2 | $1.49B |
| 2027Q3E | $1.57B |
| 2027Q4E | $1.66B |
| 2028Q1E | $1.75B |
| 2028Q2E | $1.85B |
| 2028Q3E | $1.95B |
| 2028Q4E | $2.05B |
| 2029Q1E | $2.16B |
| 2029Q2E | $2.27B |
| 2029Q3E | $2.39B |
| 2029Q4E | $2.51B |
| 2030Q1E | $2.63B |
| 2030Q2E | $2.76B |
| 2030Q3E | $2.90B |
| 2030Q4E | $3.04B |
| 2031Q1E | $3.18B |
| 2031Q2E | $3.33B |
| 2031Q3E | $3.49B |
| 2031Q4E | $3.65B |
| 2032Q1E | $3.82B |
| 2032Q2E | $3.99B |
Assumptions & reasoning
- Product revenue per customer is derived, not disclosed. Snowflake publishes the total customer count and the product revenue every quarter; the ratio is arithmetic. It was $95,912 per customer per quarter in the basis quarter, up 11.4% year over year, and the driver carries it as $31,970.67 a month because the engine bills ARPU monthly.
- The engine damps net adds by the headroom left against the ceiling, so the 1,157 net-add capacity in the field is the disclosed 616 net adds grossed up by the 53.6% of the 30,000-customer pool still unfilled. The published net-adds trace prints 616 in the first projected quarter, which is what Snowflake reported.
- Total customer counts are restated for acquisitions, consolidations and spin-offs. The FY2026 10-K states 13,328 total customers at 31 January 2026; the FY2027 Q1 10-Q restates that same date to 13,296, and the 616 net adds disclosed in the release are computed against the restated figure. The history above uses each count as first reported.
- The AI products - Cortex, Snowflake Intelligence and Cortex Code - sit inside product revenue with no separate disclosure. The CFO said Cortex Code 'had the largest driver to the increase in our forecast' but gave no dollar figure, so this model creates no AI revenue line.
- Tested for seasonality by ratio to a centred four-quarter moving average: the fiscal-quarter factors come out 0.982, 1.007, 1.004 and 0.993, a 2.5-point signal, while the same method on the longer thirteen-quarter total-revenue series gives a 1.5-point signal against a 1.1 to 2.3 point window-to-window spread within each quarter. The spread swamps the signal, so this line is modelled as aseasonal.
- Observe closed on 2 February 2026 for $595.8 million, inside the basis quarter, and Snowflake discloses no revenue contribution from it. Part of the customer and revenue step-up is therefore inorganic and cannot be sized from disclosure.