Most AI can retrieve or generate. Fewer can reason across multiple steps and data sources. Snowflake AI Research shows how ArcticSwarm tackles that and is transforming hybrid deep research for enterprise intelligence: bit.ly/4y1NvZa

Governance isn't slowing innovation, it’s enabling it. Snowflake CMO Denise Persson unveils the 5th Modern Marketing Data Stack report. Get the blueprint for the agentic enterprise. 👉🏻 bit.ly/4uXrWGr

Need to scale model inference over millions of unstructured datapoints? Snowflake now supports job-based batch inference powered by Ray: one API call, distributed GPU execution, and results written back to Snowflake: bit.ly/4nGN15S

Bad data is expensive because you usually catch it too late. Snowflake now supports enforced CHECK Constraints + Error Tables, so validation happens at write time and rejected rows are captured instead of landing in the table. Check out this blog to see how to get started in 60 seconds and use Cortex Code to analyze rejected rows: bit.ly/433ke1L

Running Apache Kafka® at scale shouldn't mean your connector is doing work it was never supposed to own. ⚡ The new Snowflake Kafka Connector V4 is a ground-up rewrite that moves heavy lifting like buffer management and schema validation into the Snowflake AI Data Cloud. By shifting this logic server-side, we have boosted throughput to 10 GB/s and cut costs by 50 percent. The connector has one straightforward job now: deliver the rows while Snowflake handles the rest. Learn more 👉🏻 bit.ly/4tGSyLZ

Detection is only half the job. Remediation is where the real work starts. With Cortex Code integrated with Snowflake Trust Center, we're helping teams investigate violations in chat, get AI-guided remediation, and move faster from finding to fix in Snowsight—with human oversight built in. Check out this step-by-step guide on how to get started: bit.ly/4mTpH4v

Enterprise AI should do more than generate answers. Snowflake Intelligence is now your personalized work agent that not only understands what is happening, but can also take action. 
 Here's a look at what's new that’s changing how you work and redefining enterprise AI👇

Many AI coding assistants can generate code. Fewer understand your actual data stack. The latest's updates in Cortex Code is changing that, expanding beyond Snowflake with support for key data platforms—plus new skills, SDK support, IDE integrations, and new Snowsight experiences. Deep dive into everything we've just shipped 🧵 bit.ly/4mItHol

Put more powerful AI to work on your enterprise data—securely, quickly, and at scale. Start building with Claude Opus 4.7—available now in public preview in the U.S. and EU—in Cortex Code, Cortex AI Functions, and Cortex REST API on Snowflake Cortex AI. See what’s possible: bit.ly/3QdwRUU

❄️ The April Snow Report is LIVE ❄️ The Snowflake AI Data Cloud is accelerating this spring, with expanded access to Cortex Code and major interoperability upgrades for Iceberg Tables. It is now even easier to build and govern enterprise AI at scale. 🚀 What’s new Cortex Code, everywhere Our agentic assistant now reaches more developers than ever. It now supports Windows in the CLI and offers deeper integration across Snowflake Workspaces and Notebooks. Iceberg interoperability, unlocked External Engine Reads for Snowflake managed Iceberg Tables are now GA, powered by Horizon APIs. This enables seamless access with Spark and Trino. Next up External Engine Writes are coming soon in Public Preview. 🗓️ Don’t miss what’s ahead • Apr 15: Hands on lab. Build data engineering pipelines in Snowflake Notebooks • Apr 16: Snowflake Connect: Analytics • Apr 21 to 30: Accelerate 2026 virtual series covering Retail, Financial Services and Healthcare and more • Jun 1 to 4: Snowflake Summit in San Francisco. Secure early bird pricing ➡️ Watch the full report to see it all in action and learn why 92% of early AI adopters are already seeing positive ROI. bit.ly/4crgADf

Snowflake and @googlecloud are rewriting the rules for high-performance AI. 🤝 We have deepened our technical integration to bring Google Cloud’s new Axion processors directly into Snowflake. By incorporating this custom silicon into our latest warehouses, your most complex data workloads benefit from: ❄️Faster throughput ❄️Lower latency ❄️Enhanced memory Don't let hidden hardware bottlenecks slow down your most complex data tasks. 👉 bit.ly/4sUrWXk

Find your most important ML features with a single prompt. Cortex Code can: - Audit your model - Flag weak signals - Suggest improvements Faster insights. Less guesswork. Learn more about agentic ML in Snowflake: bit.ly/41DV74R

New agent skills standardize how you build on data. From Streamlit apps to data movement, cost tracking, ML workflows, and AI Functions, Cortex Code now handles more of the heavy lifting so you can focus on building: bit.ly/4rRCqFx

Accelerate the entire machine learning lifecycle with agentic ML from Cortex Code Build, train, and deploy production-ready ML pipelines from natural language, directly in Snowflake. Automate feature engineering, training, and inference with simple prompts on the same platform as your governed data: bit.ly/3Q0xP6C

Cortex Code is getting more powerful across your entire workflow. What’s new: ✨ Now GA in Snowsight ✨ Native Windows support for CLI ✨ Agent Teams for parallel work ✨ New agent skills for building on data Dive in 👇

Imagine reclaiming 500 hours of work in just 20 days. 🤯 Evolv CTO Trent Foley explains how they’re using Cortex Code to build at lightning speed by giving AI agents full context within the Snowflake AI Data Cloud. 👉🏻 bit.ly/4cxpdxO

Cortex Code CLI brings Snowflake expertise into your terminal or IDE. Ask for a dataset summary, break down trends with Python, and generate visualizations or a @Streamlit app. No schema deep dives. No boilerplate. Learn more: bit.ly/4bkcXzF