Snowflake Puts AI-Ready Enterprise Data at Your Fingertips

We know developers don’t want to wait for an access request to be approved or a data set to update before they start building the next great app or AI agent. We also know that even the best model can’t make up for data that’s incomplete, fragmented and out of date. 

To be ready for AI, data must be continuously available, accessible, usable and governable — it’s what makes the difference between AI that fails and AI that succeeds in the real world. That’s why Snowflake is committed to making sure our customers always have AI-ready data to power their production AI deployments and intelligent applications. Our newest platform enhancements span transactional processing, analytics, automated optimization, interoperability and enterprise-grade governance capabilities and resilience so you can:

• Make all of your enterprise data AI-ready, including transactional and real-time data

• Keep AI systems fast and responsive with built-in continuous performance optimization

• Reduce data movement and fragmentation across operational, analytical and AI workloads

• Securely use sensitive data in AI initiatives with governance controls that are embedded in your platform, not add-ons

• Scale AI confidently with enterprise-grade interoperability and resilience

Let’s take a closer look at how Snowflake is closing the loop between data, AI and action.

AI-ready data: The secret to AI success

To deliver true enterprise intelligence, AI requires readily available operational data with relevant context. That means modernizing both your data strategy and the systems where that data lives so that AI runs on fresh, connected data. 

Snowflake is addressing this challenge on multiple fronts. For organizations looking to modernize legacy platforms, SnowConvert AI offers AI-powered code conversion (now generally available) to accelerate the path to AI-ready data. Improvements to the underlying AI agents enable higher accuracy and lower latency, while a new built-in unit testing approach uses synthetic data to verify execution across both source and Snowflake target databases. Expanded migration coverage for Sybase stored procedures, user-defined functions (UDFs) and SSIS-to-dbt projects preserves business-critical logic while reducing manual migration effort. AI-powered conversion to Snowflake managed Apache Iceberg™ tables (now generally available) offers another source table conversion option for Teradata that addresses the need for open table format interoperability while preserving Snowflake managed performance, reliability and governance controls.

For those looking to stop bolting on external databases or caching layers when building apps, there’s the one-two punch of Snowflake Postgres and Interactive Analytics. These complementary offerings have a shared goal: helping developers build high-performance applications faster and easier by running all data workloads on Snowflake.

Snowflake Postgres, generally available soon, is built for transactional use cases. It brings Postgres to the Snowflake AI Data Cloud, delivering a production-ready foundation for applications and AI agents, simplifying your architectural footprint through infrastructure consolidation. Snowflake Postgres connects transactional data to analytics without brittle pipelines and supports enterprise-grade trust with battle-tested security, resilience and scale. Developers get 100% Postgres compatibility and high performance, while businesses gain a unified platform to power the next generation of real-time apps and AI agents.

Snowflake Postgres is well-suited for use cases such as:

• Modern app and AI development: Power a new class of context-aware, intelligent applications that require high-throughput transactions and large-scale analytics simultaneously.

• Real-time analytics: Make fresh operational data ready for analysis in the AI Data Cloud without the hidden tax of ETL.

• Operational store: Modernize and consolidate backends for web and enterprise applications by lifting and shifting existing Postgres apps without rewriting code. 

Interactive Analytics is built for read-heavy analytical use cases, allowing thousands of users to hit an analytics dashboard or API simultaneously with sub-second responses. That’s a valuable competitive advantage when it leads to confident decisions, faster (and better-informed) customer service or quick pivots to match market trends. Use Interactive Analytics in high-volume scenarios such as:

• User-facing analytics: Power dashboards or reporting modules in a SaaS app where users expect lightning-fast response times.
  High-throughput APIs: Serve as the analytics backend for data-rich APIs that handle large amounts of concurrent read requests.

• Real-time monitoring: Provide the low-latency serving layer for telemetry, observability, IoT consoles and other situations where data must be visualized instantly as it arrives. 

Performance at the speed of AI

Effective AI must operate in real time — not just in batch or offline — and it needs fast, predictable data access. It also must scale to meet demand without constant tuning or hands-on management. 

Snowflake Optima gives developers (and their AI projects) a boost by continuously and automatically analyzing workload history and delivering intelligent optimizations that accelerate performance. Optima Metadata (generally available) generates workload-specific metadata to improve query performance; using this tool alongside Optima Indexing means developers can now automatically optimize storage, indexing and metadata performance. In addition, Query Acceleration Service is now enabled by default for Gen2 and multi-cluster warehouses, speeding up critical queries, including those on Apache Iceberg tables, and giving users a better view of query usage and performance. 

Meanwhile, Snowflake’s ongoing enhancements to Snowpipe Streaming, Dynamic Tables and Gen2 warehouses keep data continuously ingested, transformed and available in near real time. Coupled with Snowflake Optima and the sub-second analytical performance of Interactive Analytics, they give developers reliable performance to support the needs of real-world generative AI and agentic innovations.

Governed, open and interoperable data for trusted AI

Your AI is only as trustworthy as the data behind it. You should know where the data originally came from, be able to securely tap into the wealth of insights in sensitive data and be able to access data stored in various formats and locations. As more people use AI tools for business-critical functions, every hiccup or hallucination could not only tarnish your reputation but lead to more serious errors: According to a recent KPMG study, 66% of respondents rely on AI output without evaluating accuracy, and 56% are making mistakes in their work due to AI.

Developers shouldn’t have to spend extra cycles thinking about governance and security — it should just work. With Snowflake, governance, security and resilience are embedded directly into the platform, so the apps and AI you build on Snowflake have a solid governance foundation from the start. Capabilities such as AI Redact (now generally available), new EU-specific sensitive data classifiers and new tag-based row access and projection policies allow you to analyze and use sensitive data sets safely, without impacting privacy or rendering the data unusable. 

To make it simpler to assess and manage data quality, we have enhanced our data quality notifications (public preview) and anomaly detection (generally available) features. External lineage capabilities (public preview) allow you to pull in data from external sources and get a complete, end-to-end view of data lineage so you can rest assured that your AI models are trained on reliable inputs. And once your data is in good shape, Snowflake Backups can create immutable backups with retention lock to help protect data from loss, tampering or cyberattack and help you restore access and recover valuable data in the case of outages.

Snowflake is also expanding open lakehouse interoperability, enabling AI to securely access governed data regardless of where it’s stored across clouds, catalogs and formats. Support for Apache Iceberg V3, including the ability to process semi-structured data with VARIANT data types (public preview soon), and Horizon Catalog interoperability capabilities (external read support generally available; external write support in public preview) make it easier to access Snowflake managed tables from any engine via open APIs for maximum flexibility. Horizon Catalog also simplifies the process of extending unified governance across the Apache Iceberg ecosystem by allowing you to enforce row and column masking policies consistently across Apache Spark and Snowflake for Iceberg tables (generally available). 

A newly launched integration with Microsoft OneLake, now generally available, enables mutual customers with secured bidirectional read access for Apache Iceberg data managed by Snowflake or Microsoft Fabric. This means you can seamlessly access all your data across both platforms without complexity or data duplication. Hear from our product executives on how this is enabling real openness and interoperability, and watch this vide to dive deeper into what's new. And to facilitate data sharing across teams, clouds and regions, open table format sharing extends Snowflake’s zero-ETL sharing model to additional formats, including Apache Iceberg and Delta Lake. Developers can collaborate across open formats while maintaining control over both access and costs.

AI-ready enterprise data by design

Snowflake is raising the bar when it comes to setting an easy, connected and trusted data foundation that makes enterprise data AI-ready by design. Developers can build intelligent apps and production-ready AI agents on one central, integrated platform that has governance, interoperability and high performance already built in from the ground up. 

Learn more about how to make your data AI-ready with the power of the Snowflake platform at snowflake.com/en/make-data-ai-ready, then check out how the AI Data Cloud can help you build better with new features that deliver enterprise intelligence and agentic AI at scale and modernize your developer workflow. 

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.