Snowflake
Consumption-based (usage credits) SaaS; product revenue ~96% of total; multi-cloud, land-and-expand with high net revenue retention.
The thesis on this name
State of AI Compute
Cloud-native data platform (the 'AI Data Cloud') for storing, querying, sharing and now running AI/ML workloads on governed enterprise data, sold on a consumption (per-credit) model across AWS, Azure and GCP.
State of Enterprise AI SaaS
Cheapest of the hyper-growth consumption cohort (~7.9x NTM EV/rev vs DDOG ~11.4x) while RE-ACCELERATING: Q1 FY27 product rev +33.9% to $1.334B, FY27 product guide raised to $5.84B (~31%), Cortex Code to 7,100 accounts since Feb-2026 GA (Snowflake Q1 FY27, May 2026). Consumption pricing means AI workloads (training data prep, RAG, agent queries) directly grow the bill; the seat-cannibalization fear that crushed CRM simply doesn't apply. New-CEO operating discipline turned the 2026 derate into an entry.
State of Enterprise AI SaaS
Consumption model + AI tailwind real (Cortex/CoCo drove record sequential growth, product revenue reaccelerated to 34%, NRR back to 126%) — but ~9x sales on a still-unprofitable name leaves little cushion.
State of Enterprise AI SaaS
Cheapest hyper-growth consumption name (~7.9x NTM EV/rev) re-accelerating to +34%; highest reward-to-risk in the cohort. Top conviction long.
Earnings, margins, COGS & capex
Re-accelerating consumption growth (+34% product, fastest in years) driven by AI products (Cortex, Snowflake Intelligence) and core warehouse expansion; non-GAAP profitable and strongly FCF-generative, but still deeply GAAP-unprofitable due to ~$430M/qtr stock-based comp. Net-cash balance sheet. FY27 guidance raised on both revenue and margin after Q1.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~25¢ is cost of goods and ~63¢ operating expense, leaving ~12¢ of operating profit (~126¢ net).
Revenue trend
Margins
Stable/high
Expanding (+300bps YoY); FY27 guide raised to 13.5%
Negative, improving; gap = SBC (~$434M/qtr)
Strong, expanding
Up from 125% (Q4 FY26); ticked back above expansion threshold
COGS structure
Primarily third-party cloud infrastructure (AWS/Azure/GCP compute & storage) resold as usage credits, plus support and PS; gross margin ~75% non-GAAP reflects the hyperscaler resell spread plus optimization. The new $6B multi-year AWS agreement is both a cost commitment and a marketplace/co-sell channel (>$7B lifetime AWS Marketplace sales).
Capex
Minimal owned infrastructure capex — asset-light multi-cloud model; capacity is bought from hyperscalers (commitments like the $6B AWS deal) rather than built. Cash use skews to SBC, M&A (Natoma) and buybacks.
Latest earnings
Beat — non-GAAP EPS $0.39 vs ~$0.32 consensus (~22% beat); raised FY27 revenue and op-margin guidance; stock rallied ~36-37% post-print
FY2027 product revenue raised to $5.84B (+31% YoY) from $5.66B; non-GAAP op margin raised to 13.5% from 12.5%; adj FCF margin ~23%; Q2 product $1.415-1.42B (+30%)
- Product revenue
- $1.33B (+34% YoY)
- Net revenue retention
- 126%
- Total RPO
- $9.2B (+38% YoY)
- Customers >$1M TTM
- 779
- Customers >$10M TTM
- 64
- Total customers
- ~13,912 (+616 net adds)
- Non-GAAP EPS
- $0.39 (beat $0.32)
- Adjusted FCF margin
- ~19% (Q1)
Growth drivers
- AI products: Cortex (LLM functions), Cortex Code (launched Feb 2026, >7,100 accounts — fastest-adopted product ever), Snowflake Intelligence agentic analytics
- Core data-warehouse consumption expansion + new workloads (Iceberg open tables, Unistore/transactional, Snowpark)
- Land-and-expand: 126% NRR, 779 customers >$1M (vs 575 a yr ago), 64 customers >$10M
- Data sharing / Marketplace network effects and Openflow ingestion
- $6B AWS co-sell + Marketplace channel; Natoma (MCP/agent governance) acquisition
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-03-20. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’26
Bull & bear
A category-defining data platform re-accelerating to +34% product growth at $5B+ scale, with a credible AI consumption tailwind (Cortex), 126% NRR, strong FCF, a net-cash balance sheet, and a multiple (~10x fwd EV/sales) well below its ~20x history — a re-rating candidate if AI consumption compounds.
- Growth re-accelerated to +34% product YoY (from +26-30%), the opposite of a maturing-software story
- AI products adopting faster than anything in company history (Cortex Code >7,100 accounts) — optionality not in the base
- RPO $9.2B (+38%) and 779 $1M+ customers signal durable, expanding demand
- Asset-light, ~23% FCF-margin model with net cash and buyback capacity
- Trades ~10x fwd EV/revenue vs ~20x historical avg and below peer P/S — multiple has room to expand on AI re-rating
- $6B AWS co-sell + Marketplace channel and dbt/Fivetran/Salesforce ecosystem deepen distribution
Still GAAP-loss-making with heavy dilution, sells a consumption model exposed to cost-optimization, and faces the three deepest-pocketed companies in tech (Amazon, Microsoft, Google) plus Databricks bundling 'native' AI analytics — all at a premium valuation that already prices in the AI re-acceleration.
- GAAP operating loss -23% with ~$434M/qtr SBC — 'non-GAAP profitable' masks real dilution
- Consumption revenue can decelerate fast when customers optimize or budgets tighten
- Hyperscalers (Redshift/BigQuery/Fabric) can bundle and undercut; Databricks targets the AI/ML and governance layer
- Open table formats (Iceberg) erode storage lock-in that underpinned the moat
- ~10x fwd EV/sales is cheap vs its own history but still a premium absolute multiple requiring sustained 30%+ growth
- Competitor surveys flag potential share loss as enterprises reassess Snowflake commitments
What it is worth
Relative multiple (EV/Sales, P/S) cross-checked with reverse-DCF sanity on implied growth/margin
~$150-200
consumption decelerates on optimization/competition, AI monetization underwhelms, multiple compresses toward ~6-7x fwd sales (Macquarie-style ~$200 low; downside to ~$150 on a growth scare).
~$285-295
growth normalizes to high-20s/low-30s%, FCF margin ~23-25%, multiple ~11-12x fwd sales — roughly consensus avg target.
~$340-370
AI consumption (Cortex/agents) compounds, growth holds >30% with FCF margin expansion, multiple re-rates toward ~14-15x fwd sales (UBS-style ~$370 target).
At ~$238/sh (~$79-83B mkt cap, ~$77B EV after net cash), ~10.6x P/S and ~10x fwd EV/revenue vs a ~20x 3-yr historical average and ~14.8x peer P/S. Reverse-DCF: at ~10x fwd EV/sales the price implies sustaining roughly high-20s%-to-30% revenue CAGR for several years while expanding FCF margin toward the ~20-25%+ zone — broadly consistent with the raised FY27 guide (+31% product, 23% FCF) but with little margin for a deceleration. GAAP losses + SBC dilution are the key downside risk to per-share value. Consensus ~Buy/Strong-Buy, avg target ~$288-292 (range $200 Macquarie to $370 UBS, Jun 2026).
SWOT
Strengths
- Re-accelerating top line (+34% product) at ~$5B+ scale — rare to re-accelerate at this size
- High gross margin (~75%) and strong adjusted FCF (~23% guide) with a net-cash balance sheet
- Best-in-class consumption metrics: 126% NRR, RPO $9.2B (+38%), 779 $1M+ customers
- Brand/standard status as the cloud data platform; deep multi-cloud + data-sharing/Marketplace network effects
- Fast AI-product adoption (Cortex Code fastest product ever) extending the platform beyond storage/query
Weaknesses
- Deeply GAAP-unprofitable (-23% op margin) due to heavy stock-based comp (~$434M/qtr), diluting shareholders
- Consumption model makes revenue sensitive to customer cost-optimization and macro IT budgets
- Lower gross margin than pure-SaaS peers because it resells hyperscaler compute
- Customer concentration in large accounts; expansion can stall in downturns
- Premium valuation leaves little room for a growth stumble
Opportunities
- Monetizing the AI/agent layer (Cortex, Snowflake Intelligence, MCP via Natoma) to lift consumption per customer
- Open-format (Apache Iceberg) and Openflow to capture lakehouse/open-table workloads from Databricks
- International expansion and verticalized data products / Marketplace data exchange
- Unistore/transactional + Snowpark widening addressable workloads beyond analytics
- Large >$10M-customer cohort still early — ample expansion headroom
Threats
- Hyperscalers as both supplier and rival — AWS Redshift, Google BigQuery, Microsoft Fabric bundling AI/analytics natively
- Databricks lakehouse + Unity Catalog encroaching on AI/ML and governance workloads
- AI 'native vs bolted-on' narrative used by competitors against Cortex
- Pricing pressure / customers optimizing spend as the consumption model rewards efficiency
- Open table formats (Iceberg) reducing lock-in and commoditizing storage
Moats, dependencies & bottlenecks
Moats
Pipelines, governance, and accumulated data make migration costly — but open Iceberg tables partly erode lock-in.
Cross-org secure data sharing and the Marketplace get more valuable as more orgs join.
De-facto cloud-data-platform standard with ~35% segment share; pressured by hyperscaler bundling.
Runs across AWS/Azure/GCP — a differentiator vs single-cloud Redshift/BigQuery/Fabric, but depends on those same suppliers.
dbt, Fivetran, Salesforce, ServiceNow integrations + Cortex/agent layer raise stickiness.
Dependencies
Supplier + channel + competitor Largest infra host; new $6B multi-year AWS deal and >$7B Marketplace sales — but Redshift competes directly.
Supplier + competitor Hosts Snowflake workloads; Microsoft Fabric (OneLake + Power BI + Copilot) competes for the same budget.
Supplier + competitor GCP host; BigQuery (deep Gemini/Vertex integration) is a leading rival.
Technology partner GPU/AI stack underpinning Cortex AI workloads via the hyperscalers.
Consumption model is directly exposed to customer spend optimization and budget cycles.
Advantages
- Multi-cloud, single-platform architecture with separation of storage and compute
- Re-accelerating growth at scale with strong NRR and a large, expanding $1M+ customer base
- Net-cash balance sheet and ~23% adjusted FCF margin — self-funding the AI build-out and buybacks
- Fast-adopting AI/agent product layer (Cortex, Snowflake Intelligence, Natoma MCP) extending consumption
Weaknesses
- Persistent GAAP losses and heavy dilution from SBC
- Lower gross margin than pure-SaaS peers (resells hyperscaler compute)
- Premium valuation vulnerable to any growth deceleration
- Three of its key suppliers (Amazon, Microsoft, Google) are also direct competitors
Bottlenecks
- Reliance on hyperscalers (AWS/Azure/GCP) for the underlying compute it resells — caps gross margin and ties cost to suppliers who also compete
- Stock-based compensation magnitude — the gap between non-GAAP profit and GAAP losses/dilution
- Consumption-model revenue visibility — usage can swing with optimization and macro
Top signals & trends
Top signals
Counter to the 'maturing' thesis; AI consumption + core both contributing.
Management confidence; stock +36% on the print.
Expansion within base re-strengthening.
Real-economics dilution persists despite non-GAAP profit.
Hyperscaler + Databricks bundling pressure.
Channel expansion + agent-governance capability for the AI roadmap.
Trends
Shifts the question from 'where to store data' to 'where to govern/activate it' — Snowflake's Cortex/Intelligence push targets this.
Lets Snowflake capture lakehouse workloads but erodes storage lock-in.
Native AI-analytics bundling pressures Snowflake's neutral-platform value prop.
DBMS market ~$120B (2024), cloud ~64% of spend, projected ~$161B by 2026 — large expanding TAM.
Customers optimize spend; rewards efficiency but pressures near-term revenue.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Primary cloud-infra host; $6B multi-year agreement; resold compute is the bulk of COGS.
Cloud-infra host for Azure-region workloads.
Cloud-infra host for GCP-region workloads.
GPU/AI compute underpinning Cortex AI features via hyperscalers.
779 >$1M TTM) Includes large financial, retail, tech and healthcare enterprises across Fortune 500 / Global 2000.
Adobe, AT&T, Pfizer-type enterprises (illustrative large accounts) Representative of the large-account cohort driving consumption; specific TTM spend not individually disclosed.
Lakehouse + Unity Catalog + Data Intelligence Platform; primary AI/ML and governance rival; IPO-watch name.
OneLake + Power BI + Copilot bundle; ~12% segment share and a supplier to Snowflake.
Serverless analytics leader, deep Gemini/Vertex AI integration; ~28% segment share.
AWS-native warehouse; ~20% share; also Snowflake's largest infra supplier/channel.
Incumbent enterprise data warehousing displaced by cloud platforms; residual competition.