Snowflake is announcing major updates to Snowflake CoCo, the company’s AI-powered code agent that helps developers, analysts, and businesses build applications, automate workflows, and operationalize AI faster. It is also launching Snowflake Datastream, a new fully managed streaming service for Apache Kafka that will make it easier to combine real-time data and AI in the same platform.
On Snowflake Summit 2026 Snowflake announced several new features for Snowflake CoCo, formerly known as Cortex Code. The solution acts as an AI-based code agent that helps users develop applications, automate processes, and create AI-powered services through simple text instructions.
At the same time, the company is introducing Snowflake Datastream, a new streaming service with built-in support for Apache Kafka. The service makes it possible to stream data directly to Snowflake without the need for separate brokers, connectors, or additional streaming infrastructure.
AI development where users are already working
A central part of the initiative is to make CoCo available in the tools that developers and businesses already use in their daily work.
CoCo is now available as a native desktop application and also integrates with tools like Microsoft Excel, VS Code, Claude Code, Slack, and mobile devices. The goal is to enable users to build, monitor, and manage AI projects regardless of their work environment.
According to Snowflake, users can now also have CoCo perform tasks autonomously. This means that recurring workflows, validations and monitoring can be automated without continuous manual control. The solution also benefits from Snowflake’s existing security and authorization model, including role-based access control and audit trails.
Will open AI development to more users
Christian Kleinerman, Executive Vice President of Product at Snowflake, believes that CoCo can change how organizations build AI solutions.
– CoCo makes work significantly faster and easier for experienced developers, but also opens the door for new types of users, such as analysts and data-savvy business users, to create workflows, automations and AI applications themselves, says Christian Kleinerman.
He also emphasizes that the combination of CoCo and Datastream will reduce the complexity of real-time data and thus make it easier to build AI solutions that work with current information.
Customers are already using the solution in production
Several large organizations are already using CoCo to streamline data management and AI development.
Reference customers include Fanatics, Thomson Reuters and WHOOP. According to Snowflake, the platform is used to reduce development times, simplify data management and enable more teams to contribute to AI initiatives in production.
Maddy Want, VP of Data at Fanatics, describes how developers who previously needed several days to troubleshoot data processes can now solve the same problems in a few hours.
Also Thomson Reuters highlights that the organization uses Snowflake as a central data platform with over 37,500 managed data tables and more than 350 data sources, where CoCo will contribute to faster development and modernization of legacy systems.
Snowflake Datastream will simplify real-time data
One of the biggest news is the launch of Snowflake Datastream.
Many organizations today use separate streaming platforms alongside their data platforms, which often require additional infrastructure, integrations and specialist expertise. Snowflake now wants to bring these capabilities directly into its own platform.
Datastream is built directly on top of the Snowflake platform and is compatible with Apache Kafka. This means that companies can connect existing streaming applications without having to rebuild their solutions. At the same time, security, data governance, access controls and traceability are integrated in the same way as for other data in Snowflake.
Real-time data and AI in the same platform
By combining Datastream with CoCo, Snowflake wants to create a common platform for both real-time data and AI development.
Datastream delivers continuously updated data while CoCo acts as the AI-powered development agent on top of the platform. Together, the solutions will enable building AI applications and automated workflows based on current data streams using simple text instructions.
The initiative is part of Snowflake's broader strategy to establish an agent-based AI platform where developers, analysts and business users can collaborate on data, applications and artificial intelligence within the same controlled environment.
Swedish
Danish
Norwegian
Finnish
English
German
Dutch