.webp)
.webp)
Snowflake has rolled out new tools across its data and AI platform, including a coding agent, an automated semantic layer service, and updates to its Postgres offering that will run natively in the Snowflake environment.
The company also disclosed a multi-year partnership with OpenAI valued at USD $200 million. Snowflake said the agreement covers co-innovation and joint go-to-market activity. It also makes OpenAI models available to Snowflake customers through Snowflake Cortex AI.
The product updates centre on three areas. Snowflake has made Cortex Code generally available. It has also made Semantic View Autopilot generally available. It also announced enhancements to Snowflake Postgres, with general availability planned soon.
OpenAI deal
Snowflake said the OpenAI partnership will make OpenAI models available natively within Snowflake Cortex AI for its customer base of 12,600 organisations. The company said customers including Canva and Whoop already use OpenAI models with their enterprise data in Snowflake.
Snowflake also said OpenAI models will be accessible within Snowflake Intelligence. The company describes Snowflake Intelligence as an enterprise intelligence agent for access and analysis using natural language.
Snowflake said one of the OpenAI models expected to be accessible is GPT-5.2.
Cortex Code
Snowflake positioned Cortex Code as an AI coding agent designed for work tied to Snowflake data and operational context. The company said it targets end-to-end development tasks and it runs within Snowflake environments and local developer tools.
Snowflake said Cortex Code understands a user's Snowflake data, compute, governance, and operational semantics. It said this differs from generic coding assistants.
Users can access the tool through Cortex Code CLI, which Snowflake said is now generally available. Snowflake also referenced Cortex Code in Snowsight, which it said is generally available soon.
Snowflake also said Cortex Code will work with developer environments including VS Code and Cursor.
The company also announced an integration with v0 by Vercel. Snowflake said this is intended for "vibe coding" workflows. Snowflake said the integration is generally available soon.
Snowflake said apps created through the integration can be deployed inside Snowflake using Snowpark Container Services.
Semantic layer
Snowflake said Semantic View Autopilot automates the creation and governance of semantic views. The company said the service learns continuously from real user activity. It said the aim is to keep business logic accurate and up to date.
Snowflake linked the release to a broader challenge in enterprise AI deployments. The company said organisations often define business metrics manually and govern them inconsistently. It said that leaves AI systems without a shared understanding of business context.
Snowflake said Semantic View Autopilot automatically builds, optimises, and maintains governed semantic views. The company said this removes manual semantic modelling work.
The company also referenced its involvement in the Open Semantic Interchange. Snowflake said the initiative standardises a shared semantic layer across ecosystem participants. It said Semantic View Autopilot adds automation around creation and maintenance.
Postgres changes
Snowflake also announced enhancements to Snowflake Postgres, which it said will run natively in the Snowflake AI Data Cloud. Snowflake said the product will allow access, ingestion and migration of data from multiple locations into Snowflake.
Snowflake said many organisations still keep transactional databases and analytical databases on separate systems. It said this structure drives reliance on pipelines between systems and increases complexity.
Snowflake said Snowflake Postgres consolidates transactional, analytical, and AI use cases on one platform. It also said the product offers compatibility with open source Postgres and it can support migrations without code changes.
"For AI to truly deliver value, it must move beyond experimentation and become engrained within the systems teams rely on every day," said Christian Kleinerman, EVP of Product at Snowflake. "With our latest product advancements, we're reimagining how teams build and operate by embedding AI directly into the development lifecycle, making data AI-ready by design and helping enterprises deliver real business impact with AI. This marks a fundamental shift in how organisations build with data and AI, enabling users to build solutions that are reliable, governed, and ready to run at enterprise scale."
Snowflake said it will make Snowflake Postgres generally available soon as it continues to expand its AI Data Cloud product line-up.