Snowflake announced a complete AI development platform at BUILD London 2026, rolling out Cortex AI Functions, general availability for Notebooks, an Online Feature Store, and Agent Evaluations tooling. The suite enables enterprises to build, test, and deploy AI agents directly within Snowflake's data cloud without moving data to external platforms.
The London timing reflects European enterprises' accelerating shift toward production AI, particularly in financial services. HSBC partnered with Mistral AI for regulated AI workloads, while BNP Paribas deployed Mistral's models for compliance and risk analysis. Wells Fargo launched Google Agentspace for internal knowledge systems, and Citigroup built its Stylus platform for AI-powered financial analysis.
Cloud vendors are embedding AI capabilities directly into infrastructure stacks. AWS released Bedrock AgentCore for autonomous agent orchestration. Google Cloud expanded Vertex AI with managed agentic workflows. Azure AI added model inference and prompt management tools. Each platform aims to capture enterprise customers before they commit to competing ecosystems.
Snowflake's Online Feature Store addresses a critical gap: serving real-time data to AI models with sub-100ms latency. Traditional data warehouses can't support instant credit decisions or fraud detection, forcing companies to build separate infrastructure. Cortex AI Functions pre-package common AI tasks like sentiment analysis and entity extraction, reducing development time from weeks to hours.
Agent Evaluations provide automated testing for AI reliability, a key concern for regulated industries. European financial regulators require explainable AI decisions and audit trails. The tooling lets compliance teams validate agent outputs before production deployment.
European adoption lags US markets by 12-18 months, but regulated sectors show stronger enterprise readiness. Financial institutions demand on-premise security, GDPR compliance, and model governance—exactly what cloud platforms now offer. The competitive dynamic favors vendors who can prove security and regulatory compliance over raw model performance.
Snowflake competes on data proximity: AI runs where enterprise data already lives. AWS and Google leverage broader cloud services. Azure benefits from Microsoft 365 integration. The winner will likely be determined by which platform reduces enterprise AI friction fastest while meeting European regulatory requirements.