Snowflake Inc. (SNOW) Earnings

Management highlights

- **Agentic Enterprise Strategy**: Snowflake is positioning itself as the central foundation for the 'agentic enterprise,' providing governed data foundations, access to leading AI models, and an agentic control plane (Cortex Code and CoWork) to orchestrate actions. - **AI Product Adoption**: Cortex Code expanded to 5,800 accounts (up ~11% QoQ), while CoWork surpassed 9,100 accounts, adding over 2,000 net new accounts in Q2. These tools are being adopted across diverse personas including CFOs, CROs, and knowledge workers. - **Customer Growth & Expansion**: Net new customer additions increased 32% YoY, including 14 Global 2,000 customers. Total customers now stand at 14,600. Net Revenue Retention (NRR) remained strong at 126%, driven by migrations and AI use cases. - **Operational Efficiency**: Non-GAAP operating margin expanded over 400 basis points YoY to 15%. Headcount growth slowed significantly (334 employees added YTD, including 173 from Observe acquisition, vs. 35 in the prior year period). - **Internal AI Usage**: Management emphasized deep internal adoption of their own products; for example, finance planning was reduced from a 3-person team/50 spreadsheets to 1 analyst using models, and marketing agency spend was cut by $400k annually. - **Innovation Pace**: Over 33 product capabilities were launched to general availability in H1, a 35% increase compared to the same period last year.

Guidance

- **FY2027 Product Revenue**: Raised to $6.07 billion, implying 36% year-over-year growth (previously lower). This includes approximately 1 percentage point of growth attributed to the Observe acquisition. - **Q3 FY2027 Product Revenue**: Expected between $1.588 billion and $1.593 billion, representing 37-38% year-over-year growth. - **FY2027 Non-GAAP Operating Margin**: Increased from 13.5% to 14.5%. The revision reflects a mix shift toward faster-growing but currently lower-margin AI workloads, offset by operational discipline. - **Q3 FY2027 Non-GAAP Operating Margin**: Expected at 15.5%. - **FY2027 Non-GAAP Adjusted Free Cash Flow Margin**: Reiterated at 23%. - **Remaining Performance Obligations (RPO)**: Grew 30% YoY to $9 billion. Approximately 54% of RPO is expected to be recognized in the next 12 months, reflecting a 42% YoY growth in deferred revenue visibility.

Segment performance

Product revenue reached $1.49 billion, representing a 37% year-over-year growth rate and marking the second consecutive quarter of record sequential dollar growth. The company highlighted that AI-related revenue contributed significantly to this acceleration, with specific products like CoWork driving substantial consumption increases alongside core data platform strength. While the transcript does not break down absolute revenue by distinct product segment (e.g., Core Platform vs. AI Products) in separate line items, it notes that AI products contributed approximately half of the observed acceleration in growth.

Risks & headwinds

- **Margin Pressure from AI Mix**: The increase in AI workload revenue carries a lower contribution margin than the core platform, which has led to a downward revision in non-GAAP product gross margin guidance to 74%. - **Model Cost Volatility**: Reliance on third-party frontier models presents cost risks, although Snowflake mitigates this through model neutrality and routing options. - **Execution Risk in Agentic Transition**: Success depends on customers effectively building 'agentic' workflows; failure to achieve broad adoption or efficient consumption could impact the projected flywheel effect. - **Regulatory and Security Risks**: As with any data platform, maintaining enterprise-grade governance and security for AI agents is critical to retaining trust, particularly as usage expands into mission-critical operations.

Analyst Q&A

• Q: Analyst asked about the quality of revenue acceleration, specifically whether it stems from rational, efficient spending on AI use cases (like supply chain/finance) versus irrational consumption, and why Snowflake is the right platform for these complex business processes. /

  A: Sridhar Ramaswamy explained that acceleration comes from a broad swath of customers, not just AI-native firms. He emphasized that tools like Cortex Code enable easy cost optimization (debugging queries, idle warehouses), making efficiency a top skill. By demonstrating internal efficiency gains, Snowflake builds trust, proving that AI drives value without waste, thereby ensuring durable growth across diverse personas including CFOs and CEOs.

• Q: Analyst requested color on how much of the growth acceleration is driven directly by new AI products versus the flywheel effect on the core data platform. /

  A: Sridhar stated that AI products (including Cortex Code, CoWork, AI functions, and AI Gateway) contribute roughly half of the acceleration. However, he noted that core platform growth is also accelerating due to faster migrations (e.g., Teradata migrations completing in <3 quarters) and increased adoption of notebooks and apps. The synergy creates a compound effect where AI drives core consumption.

• Q: Analyst asked if model neutrality and choice are competitive advantages, and if shifting to open-source or different frontier models impacts gross margins. /

  A: Sridhar and Christian Kleinerman confirmed model neutrality is a key advantage, allowing customers to switch between frontier and open-source models to optimize cost and performance. They noted that open-source inference runs on Snowflake's infrastructure, offering future margin optimization opportunities. Brian Robins clarified that while AI mix lowers gross margins temporarily, the primary focus is on delivering the best business outcome; long-term operating leverage will be achieved through scale and efficiency, not by restricting model choice.

• Q: Analyst inquired about the biggest change driving the current 'AI moment' for Snowflake—whether it is tech improvements, better models, or customer readiness. /

  A: Sridhar described a 'flywheel' effect: CoWork allows quick value extraction from data, reducing backlogs for data teams. Simultaneously, Cortex Code makes Snowflake's sales and support teams 'AI-native,' enabling them to solve complex problems instantly. This combination drastically reduces 'time to 80% consumption' for new logos, accelerating activation and proving that AI drives higher lifetime consumption rather than just pulled-forward projects.

• Q: Analyst asked if Snowflake plans to train its own frontier models or if Arctic models will play a larger role in the strategy against competitors becoming databases. /

  A: Christian Kleinerman stated Snowflake has no intention of training frontier models to compete directly but continues developing specialized Arctic models for constrained tasks (document processing, embeddings) to ensure higher accuracy and efficiency. Regarding competition, he argued that application providers consolidating data into Snowflake via zero-copy partnerships is a stronger trend than them becoming databases themselves, reinforcing Snowflake's position as the central data hub.