Snowflake Openflow Revolutionizes Data Movement for AI and Interoperability

The promise of enterprise AI is as alluring as it is elusive. Many early adopters of generative AI have reported seeing material improvements around efficiency and an average 41% return on their investments already, according to those surveyed for our report The Radical ROI of Gen AI. But getting to the point of ROI is rife with challenges — around data preparation, data management and data quality. Persistent issues around responsiveness, scalability and the handling of multimodal data, of course, hamper productivity — but even more importantly, they put practical AI out of reach.

Legacy integration pipelines — rigid and difficult to adapt, extend and scale — require a staggering number of tools to manage and outsize effort to maintain. Data ends up moving through a labyrinth of fragmented pipelines, requiring elaborate engineering to integrate data from all its various sources and in all its various forms. And yet, still, these pipelines may fall short.

What modern data engineers need is a radical fix, a platform that meets all their data integration needs — adaptable and extendable, yet simpler to manage and control.

What they need is finally here: Introducing Snowflake Openflow, an open, extensible, managed multimodal data integration service that makes data movement between data sources and destinations effortless. Supporting all data types — including structured and unstructured, batch and streaming — Openflow holds the keys to enterprise AI, enabling seamless ETL and revolutionizing data movement altogether in Snowflake. 

All data integration is unified in one platform, with limitless extensibility and interoperability to connect to any data source. Openflow facilitates any data architecture, with enterprise-grade reliability and governance, allowing companies to scale their integration needs with confidence. We are thrilled to announce that Snowflake Openflow is now generally available in all AWS commercial regions with Bring Your Own Cloud (BYOC) deployment.

A comprehensive framework around Snowflake Openflow

To bring this vision to life, Openflow is built with a powerful foundation and key components for modern data integration, including:

• Open source Apache NiFi: At its core, Openflow is powered by the OSS Apache NiFi data integration framework, which is trusted by thousands of enterprise customers. Openflow supercharges Apache NiFi with enterprise-ready governance, security, deployment flexibility and observability.   

• Deployment and runtime: Customers will run the Openflow connectors and pipelines in an Openflow deployment. Customers will be able to execute in their own VPC via Bring Your Own Cloud (BYOC), generally available in AWS, or in Snowflake Snowpark Container Services (SPCS) (private preview), with the integration service managed by Snowflake. Customers can have multiple deployments tied to the control plane, each serving different use cases. Within a deployment, runtimes can scale horizontally, support multi-node and multi-cluster deployments, and establish disaster recovery resilience.

• Openflow service and API: The Openflow service and API run in Snowflake as managed services, allowing customers to provision and manage data pipelines easily with an intuitive UI. Through this, customers get a unified user experience with deep observability, including real-time monitoring and alerts, DAG visualization and refresh history information.

• Openflow connectors: Connectors in Openflow are curated, managed, secure and governed. Read the Snowflake documentation to find connectors that are generally available and in preview. Additionally, Openflow connectors are easily extensible and customizable, and Openflow power users can build their own flows with processor building blocks.

Unlocking countless doors to integrate data freely

With this managed service, Openflow brings ease of use, convenience and security to the integration-building process. This unlocks integration use cases, including: 

• Turbocharging ETL for AI: Openflow connects enterprise data to AI with near real-time, bidirectional data flows, no matter where a model or agent lives. With certain competitors, unstructured data connection may be out of reach, which makes delivering AI agents and retrieval-augmented generation (RAG) impossible. Openflow supports diverse data structures and requirements and maintains consistent user permissions as data moves from source to target. Complying with source system access control list (ACL), Openflow carries consistent user permissions with the source. It also comes with built-in capabilities to parse and preprocess the unstructured files before writing to Snowflake, leveraging native Snowflake Cortex AI capabilities all in one platform. 

• Hybrid deployment across the data estate: Openflow provides the flexibility to deploy pipelines via Snowflake’s managed infrastructure or BYOC, giving customers full control over where pipelines run while reducing operational burden as both options are managed services. Customers are no longer forced to sacrifice ease of use or control. 

• High throughput, low latency streaming data: Openflow directly connects to streaming sources, including Apache Kafka, Amazon Kinesis and Kafka Sink, enabling streaming data to flow into Snowflake and back to streaming systems. Thanks to the new Snowpipe Streaming integration with Openflow (in preview), streaming ingestion now delivers 10GB/s throughput with five-seconds-to-query latency and inline transformation. 

• Facilitating open architecture: Legacy data platforms may offer integration services, but they are often a black box and require vendor lock-in. It’s difficult to extend data integration pipelines or work with other data platforms. Openflow, on the other hand, is designed with interconnectivity at its heart, facilitating open and interoperable architecture. Users can move data to their choice of data lakes and lakehouses, and easily adapt to new industry standards, such as Apache Iceberg. 

Openflow liberates data movement with hundreds of both out-of-box connectors and easily extensible customizations, simplifying integration of diverse data into one platform. The broad range of data sources include Box, Google Ads, Microsoft Dataverse, Microsoft SharePoint, Oracle, Proofpoint, Salesforce, ServiceNow, Slack, Workday, Zendesk and more. In particular, Snowflake is partnering with Oracle for a high-performance, scalable and cost-effective solution for replicating Change Data Capture (CDC) from Oracle databases to Snowflake.  

Openflow is built with interoperability and scalability at its heart, ready to grow with any use case, especially those that support AI. Last but not least, Openflow comes with simplicity, control, observability and enterprise-grade security and governance features.

“Our goal is to deliver seamless, data-driven experiences to our customers, and Snowflake Openflow is a foundational pillar to make that real," said David Libesman, SVP & GM of Data & Analytics at Workwave. "Particularly, we have hundreds of SQL Server databases. Openflow will make it effortless to replicate data across multiple clouds and regions. But we are also excited about what it can deliver for AI and how it will prepare us to be ready for the future. It's the ETL solution that will power the next generation of our products and services."

The endless possibilities are truly electrifying as our innovative partners are ready to harness the power of Openflow. We are proud to share the excitement from our broad ecosystem with more than 30 launch partners around the globe. Together, we are ready to deliver the tremendous opportunity and value Openflow can bring to our customers. 

Get started with Snowflake Openflow

Snowflake has collaborated with implementation partners around the globe to get customers up and running with Snowflake Openflow.

Learn more about Snowflake Openflow, or try it today. If you’d like to see more demos and learn more about Openflow, join us at Data Engineering Connect on July 29, 2025. Sign up here. 

Forward Looking Statements

This article contains forward-looking statements, including about our future product offerings, and are not commitments to deliver any product offerings. Actual results and offerings may differ and are subject to known and unknown risk and uncertainties. See our latest 10-Q for more information.