Snowflake says Cortex Code adoption jumped

Snowflake said its Cortex Code coding agent reached use by half of customers within nine weeks of launch. (snowflake.com) Chief Executive Officer Sridhar Ramaswamy disclosed the figure in a post about Cortex Code, which Snowflake formally unveiled on February 3, 2026. Snowflake describes the product as an artificial intelligence coding agent that works with enterprise data context inside local development tools and Snowflake’s own environment. (snowflake.com) Snowflake launched Cortex Code as part of its Cortex artificial intelligence suite, aimed at helping teams build data pipelines, analytics, machine learning systems, and other software by writing prompts in plain language. The company expanded the command-line version on February 23, 2026 to support dbt and Apache Airflow, two widely used data engineering tools outside Snowflake itself. (snowflake.com 1) (snowflake.com 2) On March 26, 2026, Snowflake said Cortex Code in Snowsight became generally available, the command-line tool added native Windows support, and new “Agent Teams” features let the software split larger jobs into parallel tasks. In the same update, Snowflake said every Snowsight user could access Cortex Code inside Workspaces, Notebooks, and other workflows tied to the company’s governed data environment. (snowflake.com) The adoption claim lands as Snowflake is pushing investors to see artificial intelligence features as a driver of core usage, not just a demo layer on top of storage and computing. In its fiscal fourth-quarter results, Snowflake said more than 9,100 accounts were using Snowflake artificial intelligence features as of January 31, 2026, and about 2,500 accounts were using Snowflake Intelligence after three months. (snowflake.com) That matters in Snowflake’s business model because customers pay based largely on consumption. Snowflake’s pricing page says on-demand customers pay for usage month to month, and its documentation says Cortex Code command-line usage can also be billed on token consumption for pay-as-you-go accounts. (snowflake.com) (docs.snowflake.com) Snowflake is also widening the product beyond its own warehouse, which addresses a common complaint about developer tools tied too tightly to one platform. The February 23 update said Cortex Code was moving toward support for “any data source across systems,” starting with dbt and Apache Airflow. (snowflake.com) The company has been building the surrounding pieces for that pitch for months, including semantic views that help artificial intelligence systems understand business definitions and governed data structures. Snowflake highlighted those semantic-view capabilities during its June 2025 Summit announcements and has since positioned Cortex Code as a way to use that context directly in development work. (docs.snowflake.com) (snowflake.com) Snowflake’s message is that coding help tied to live enterprise data can lift customer activity on the platform. The next test is whether those early usage numbers hold as Cortex Code moves from launch novelty into everyday engineering work. (snowflake.com)