Blogs›Snowflake Feature Trails
See how major capabilities shipped, upgraded, and evolved across Snowflake's engineering blog.
Publishing pulse
2026–2026 · peak 2026
17 posts mapped

This post announces the integration of Claude Opus 5, a state-of-the-art large language model, into Snowflake Cortex AI. This integration enhances Cortex AI's capabilities by providing access to a more powerful and versatile AI model for complex tasks such as advanced reasoning, summarization, and content generation directly within Snowflake. This builds upon the existing foundation of Snowflake Cortex AI, offering users a more sophisticated AI toolkit for data analysis and application development. This post specifically details how agentic intelligence is leveraged for contract review on Snowflake, showcasing a concrete application of these advanced AI models.
Timeline

This post details the technical underpinnings and scaling strategies behind Snowflake CoCo, the framework enabling Apache Spark to run directly on Snowflake. It emphasizes how CoCo is engineered to handle enterprise-grade AI workloads by providing robust scaling, efficient resource utilization, and built-in trust mechanisms, building upon the initial introduction of Apache Spark on Snowflake.
Timeline

This post details the importance and technical implementation of bidirectional Iceberg REST, a standard that enables seamless data exchange and interoperability between different systems and Snowflake. It highlights how this standard facilitates better data management, integration, and unlocks new possibilities for data collaboration and analysis by allowing external systems to interact with Iceberg tables in a standardized, efficient manner.
Timeline

This post introduces custom incrementalization for dynamic tables, a new capability that allows users to define custom logic for how dynamic tables are incrementally updated, potentially improving performance and control over data freshness.
Timeline

This post announces the General Availability (GA) of Adaptive Compute across select AWS, Azure, and Google Cloud regions. Adaptive Compute is a capability that allows for dynamic adjustment of compute resources, likely to optimize performance and cost based on workload demands. This GA milestone signifies its readiness for production use in these cloud environments.
Timeline