Snowflake

Executive Summary#

Snowflake is a cloud-native data platform (the "Data Cloud") that lets enterprises store, query, share, and govern massive datasets across Amazon, Microsoft, and Alphabet clouds. Its consumption-based model means revenue scales with usage. The AI thesis: generative AI is only as good as the proprietary data it can access, and that data increasingly lives in platforms like Snowflake, making it a strategic AI data layer (via Cortex AI, Snowpark, and partnerships). For PCA SOF, Snowflake is a satellite-sized bet on the principle that the data layer captures durable value in the AI stack, complementing the compute (NVIDIA) and the apps. The debate is competition (Databricks, hyperscaler-native tools like Microsoft Fabric/BigQuery) and whether consumption growth re-accelerates with AI.

Investment Thesis#

Data has gravity and switching costs: once an enterprise's data, pipelines, and governance live in Snowflake, moving is painful. AI raises the value of that data (RAG, fine-tuning, analytics) and should drive consumption. Snowflake monetises per-query, so AI workloads = more consumption. The thesis: a sticky, mission-critical data platform with a long runway as enterprises consolidate data for AI, if it can defend against Databricks and hyperscaler-native rivals and reignite net revenue retention.

Why PCA SOF Owns This Company#

• Role: AI Data Layer, the data foundation for enterprise AI.
• Theme: AI Data Layer → Cloud Computing → Artificial Intelligence.
• Layer: Layer 5 of The AI Value Chain.
• Portfolio logic: owns the data leg of the AI stack; complements Datadog (observes the same estate) and runs on the hyperscalers (so it's also a customer of Amazon/Microsoft/Alphabet). Sell trigger: durable share loss to Databricks/Fabric, consumption stalling, or net retention falling below growth thresholds.

Company Overview#

US cloud data-platform company; consumption-priced; runs on top of AWS/Azure/GCP (multi-cloud). Founder-driven culture; led by CEO Sridhar Ramaswamy (ex-Google ads, ex-Neeva, an AI-forward leader).

Business Segments#

Single platform: data warehousing/lakehouse, data sharing/marketplace, Snowpark (data engineering/ML), and Cortex AI (LLM functions on governed data). Sold via consumption.

Revenue Breakdown#

(Directional) Overwhelmingly product (consumption) revenue; net revenue retention historically very high (130%+), moderating as the base scales, the key metric to watch.

Geographic Breakdown#

US-majority with growing international; enterprise-heavy customer base.

Customer Base#

Large enterprises across finance, retail, healthcare, tech; expanding "$1m+" customer cohort. Runs on and partners with the hyperscalers, both partner and competitor.

Supplier Relationships#

Compute/storage rented from Amazon AWS, Microsoft Azure, Alphabet GCP (its cost of goods + its competitors). AI model partnerships (incl. NVIDIA, Anthropic, OpenAI integrations).

Strategic Importance#

Represents the fund's conviction that the data layer is a durable value-capture point in AI, a different bet from compute or apps.

Competitive Advantages#

• Switching costs / data gravity: entrenched data + governance.
• Multi-cloud neutrality: not locked to one hyperscaler.
• Ease of use + ecosystem: marketplace, data sharing network effects.
• Cortex AI: bringing LLMs to governed enterprise data.

Competitive Threats#

• Databricks: the primary head-to-head rival (lakehouse + AI).
• Hyperscaler-native: Microsoft Fabric, Alphabet BigQuery, Amazon Redshift (the clouds it runs on). → Competitor, Snowflake vs Hyperscaler Data
• Open table formats (Iceberg) commoditising storage.

Industry Position#

A leader in cloud data warehousing/the data cloud; in a fierce duopoly-plus battle with Databricks and the hyperscalers.

Key Products#

Snowflake Data Cloud, Snowpark, Cortex AI, Native Apps, Data Marketplace, Iceberg support, Horizon governance.

Management Team#

CEO Sridhar Ramaswamy, AI-credible, refocused the company on AI + execution. Founder/chair Frank Slootman legacy of operating rigor.

Capital Allocation#

R&D-heavy; buybacks to offset dilution; SBC is high (a watch item); strategic AI acquisitions/partnerships.

Historical Growth#

Hyper-growth IPO darling; growth decelerated as the base scaled + optimisation headwinds; AI is the hoped-for re-acceleration.

Historical Earnings#

GAAP-unprofitable (high SBC) but strongly FCF-positive; the path to GAAP profitability + NRR stabilisation is the story. → Snowflake Earnings Analysis

Earnings Quality#

FCF is real; GAAP muddied by stock comp. Watch FCF margin + NRR + RPO.

Margin Analysis#

High gross margins (~70%+ product); improving operating leverage on a non-GAAP/FCF basis; GAAP pressured by SBC.

Return Metrics#

Improving FCF returns; GAAP returns weak due to SBC, a key debate.

Balance Sheet Strength#

Strong net cash, no meaningful debt.

Cash Flow Analysis#

Strong and growing FCF margin, the bull's anchor against GAAP losses.

Valuation Discussion#

A high-multiple software name. What you must believe: AI re-accelerates consumption, NRR stabilises, share holds vs Databricks/hyperscalers, and SBC moderates. Higher-variance satellite. → Valuation Framework

Major Risks#

• Competition (Databricks + hyperscaler-native) → Competitor, Snowflake vs Hyperscaler Data.
• Consumption deceleration / NRR decline.
• SBC/dilution.
• Iceberg/open formats commoditising the core.
• Valuation sensitivity to growth + rates.

Major Opportunities#

• AI/Cortex driving new consumption.
• Data sharing/marketplace network effects.
• Enterprise consolidation of data for AI.
• New workloads (Snowpark, apps, unstructured data).

Important Acquisitions#

AI/data tuck-ins (e.g., Neeva, bringing Ramaswamy + search/AI talent; Streamlit; TruEra). Bolt-ons to build the AI/app layer.

Important Divestments#

None material.

Industry Trends#

Data consolidation for AI, lakehouse convergence, open table formats, RAG/governed-data AI, consumption pricing.

Macroeconomic Sensitivities#

• Enterprise IT/cloud budgets (optimisation cycles).
• Rates: high-multiple growth → Interest Rate Sensitivity.
• AI Capex Cycle Risk (indirect, enterprise AI adoption).

Future Outlook#

Base: steady high-growth + improving FCF as AI consumption builds. Bull: Cortex/AI re-accelerates NRR and Snowflake owns the enterprise AI data layer. Bear: Databricks + hyperscalers compress growth + Iceberg commoditises storage.

Why It Matters To PCA SOF#

Snowflake is the data foundation of enterprise AI, the layer feeding the models that run on NVIDIA GPUs in the clouds of Amazon/Microsoft/Alphabet (its suppliers and rivals). It complements Datadog (observability) and the security pair (CrowdStrike/Zscaler) as the fund's "AI estate" software cluster. → The AI Value Chain, AI Ecosystem Map.

Linked Notes#

• Related Holdings: Datadog · NVIDIA · Microsoft · Amazon · Alphabet · CrowdStrike · ServiceNow · IQVIA
• Themes: AI Data Layer · Cloud Computing · Artificial Intelligence
• Maps: The AI Value Chain · Competitor, Snowflake vs Hyperscaler Data · AI Ecosystem Map · Knowledge Graph
• Risks: Interest Rate Sensitivity · AI Monetisation Risk
• Earnings: Snowflake Earnings Analysis