Snowflake Expands Cortex, Embeds GPT-5.2
Snowflake is expanding its Cortex Code platform to support data "anywhere," across clouds and formats, positioning it as an "AI agent layer" for enterprise data. The company also committed $200M to embed OpenAI's GPT-5.2 directly into its data cloud, giving its 12,600+ enterprise customers access to the model via Cortex AI.
- The partnership's "bring the model to the data" architecture allows enterprises to run AI on their data without moving it, addressing major data governance and compliance concerns for regulated industries like finance and healthcare. - The Cortex Code CLI is expanding to support external data engineering tools, starting with dbt and Apache Airflow, allowing developers to get AI assistance within their existing workflows. - The "AI agent layer" is powered by Cortex Agents, which can reason and act on data by using tools like Cortex Analyst to convert natural language to SQL and Cortex Search for retrieving information from documents. - For insurance, this technology can be applied to use cases like claims processing assistants that can manage and extract information from both structured forms and unstructured documents. - This deal provides OpenAI with a significant enterprise distribution channel independent of Microsoft, while accelerating Snowflake's transition from a data warehouse to a comprehensive AI platform. - The native integration of OpenAI models will be available across all three major cloud providers: Amazon Web Services, Microsoft Azure, and Google Cloud. - Early adopters leveraging this integration to build AI applications include the design platform Canva and the fitness wearable maker WHOOP. - The collaboration is reciprocal, with OpenAI using Snowflake for internal experiment tracking and analytics, and Snowflake using ChatGPT Enterprise to enhance its own workforce's productivity.
