Snowflake published results for the second quarter of fiscal 2027, which ended on 31 July. Product revenue rose 37% to $1.49 billion.
The number itself is good. What makes it interesting is its direction over time.
Errors or inaccuracies are possible. Spotted something that looks wrong? Write to me and I will correct it.
The Figures
| Second quarter | |
|---|---|
| Product revenue | $1.49 billion, +37% |
| Total revenue | $1.55 billion, +35% |
| Net revenue retention rate | 126% |
| Customers above $1 million a year | 828, +27% |
| Forbes Global 2000 customers | 829 |
| Remaining performance obligations | $9.00 billion, +30% |
| Adjusted free cash flow | $92.3 million, 6.0% of revenue |
The Thing That Is Hard to Do
Chief financial officer Brian Robins put what I think is the heart of the report like this:
"Q2 marks our third consecutive quarter of product revenue growth acceleration... Importantly, we delivered this accelerating growth while expanding operating margin."
Those two halves of the sentence do not usually hold together.
A software company that accelerates almost always does so with money - more salespeople, more marketing, more incentives. So acceleration normally arrives with an eroded margin, and the only question is by how much.
Here the opposite happened: full-year adjusted operating margin guidance was raised from 13.5% to 14.5% in the same breath as the growth guidance.
Why the Retention Rate Explains It
126% means a given cohort of customers spent 26% more this year than last - without a single new customer.
And that ties directly to Snowflake's model, which charges by consumption rather than by fixed subscription. When a customer runs more queries, the bill grows on its own. There is no need to sell to them again.
That explains how it is possible to accelerate and expand margin simultaneously: a large part of the growth comes from rising consumption among existing customers, and growth of that kind costs almost nothing in sales expense.
The Guidance
| Fiscal 2027 | Previously | Now |
|---|---|---|
| Product revenue | $5.840 billion, +31% | $6.070 billion, +36% |
| Adjusted operating margin | 13.5% | 14.5% |
| Adjusted free cash flow | 23.0% of revenue |
And What This Week's Context Adds
This was the week Broadcom, Dell, HPE and NetApp also reported - all growing between 30% and 86%, and all selling hardware. Several software companies reported alongside them, and the split is interesting.
| Company | Quarterly revenue growth |
|---|---|
| Broadcom | +86% |
| Dell | +58% |
| HPE | +34% |
| NetApp | +30% |
| Snowflake | +35% |
| MongoDB | +30% |
| GitLab | +21% |
| C3.ai | revenue of $52.4 million, in a turnaround |
Snowflake is the only software company in this group growing at the pace of the hardware makers.
The plausible explanation is its position in the chain: it does not sell an AI application to an organisation, but the layer where the data sits. A consumption-based model benefits from every additional use of data - including use originating from models.
And that stands in contrast to C3.ai, which sells enterprise AI applications directly, and reports revenue of only $52.4 million and an adjusted operating loss of $36.2 million - inside a restructuring its chief executive explicitly calls a "turnaround".
So even within AI software there is a large difference between sitting close to the infrastructure and sitting close to the application.
הזווית שלי
דעה אישית של אילן אברמוב - לא ייעוץ ולא המלצה
What I take from this report is a rule I try to apply to every software company: check whether the growth was bought or accrued.
Bought growth looks like this - revenue rises, sales expense rises faster, and the margin erodes. It is entirely real, but it depends on the spending continuing. Accrued growth looks different: the same customers spend more, and so it arrives with a widening rather than a narrowing margin.
126% retention alongside an expanding margin is the signature of the second kind.
And what I suggest noting: Snowflake is really a bet on data volume, not on AI. It benefits from AI because models read and write a great deal of data, not because it sells a model. That distinction sounds semantic and is not - it says the business depends on the quantity of usage rather than on any particular application succeeding.
And what I will watch is exactly what the CFO said: whether it becomes a fourth consecutive quarter of acceleration. Three quarters are a pattern. Four, particularly against a comparison base that has already risen, becomes hard to attribute to timing.
(It is important to stress: this is my personal opinion only, and nothing herein constitutes a recommendation to take any action.)
