The best way to get people using AI is to let them build with it. 🛠️ Snowflake's Arnnon Geshuri joins DCN to share why hands-on experimentation beats compliance training. Employees closest to the friction become builders and believers. Watch here 👉 bit.ly/4slgg0f

Iceberg metadata should be portable across storage environments. Snowflake Horizon Catalog now translates internal s3compat:// locations into standard s3:// paths, so compatible external engines can read and write Snowflake-managed Iceberg tables on S3-compatible storage. Learn more: bit.ly/4z2cTxV

Same prefix on every string clustering key? The first five bytes won’t distinguish much. Snowflake Optima Clustering – now GA – uses up to 1 KB of key values. One internal table saw 99%+ fewer partitions scanned after switching. See how: bit.ly/4emDPjw

Your next data or AI project has stakeholders in sales, marketing or finance. How do you turn what they need into something they can use? Explore practical sessions on connected data, governed agents and workflow automation at Snowflake Expedition. See our builder picks and build your agenda: bit.ly/4si2n2N

The relationships in your data provide the strongest signal to predict what happens next. @nvidia's newly open-sourced Kumo runs with Snowflake ML, giving enterprise agents and applications the ability to forecast outcomes, score risk, and recommend actions directly from business data, without building a custom machine learning pipeline for every request. Learn more 👉 bit.ly/4zQahDr

Close faster, trust every number and keep AI costs in check. 📊 Explore how finance teams are approaching those priorities at Snowflake Expedition, with perspectives from Snowflake Finance, @stripe, @DoorDash and @NorthernTrust. Join us on November 5, Day 3 of Snowflake Expedition, for finance-focused sessions on putting data and AI to work, leading change and managing costs.👇 bit.ly/4ywjmzW

Is your data foundation built for AI – or just built for governance? The difference is bigger than you think👇 bit.ly/47y0sOh

Detecting fraud in real-time requires production-grade ML.. Snowflake’s Online Feature Store is now generally available, serving fresh ML signals to models in milliseconds. See the fraud detection architecture: bit.ly/4scrXGJ

More time with customers. Less time on repetitive work. Better insights into where to focus. Explore how AI, grounded in your data, can help your sales and marketing teams get there. Join us on November 4, Day 2 of Snowflake Expedition, for sales- and marketing-focused sessions. Virtual and free.👇 bit.ly/3UdiamW

When AI decisions sit inside a live workflow, speed and structure both matter. ⚡️ We just introduced snowflake-decision-fast, a lower-latency, lower-cost model that returns quick answers and probabilities your code can act on directly, from a single API call to millions of rows in SQL. Read the deep dive 👉 bit.ly/485tGUL

Before your AI agent can explain retention, it needs context: what the data means and where it came from. That's Horizon Context. 🧠 Now in public preview with metadata connectors for 18 sources, steward permissions and description propagation. Details: bit.ly/4yVvR9h

The #SnowflakeStartupChallenge is back. 🔔 Last week, we celebrated our 2026 winner, @lgnd_ai, with a bell-ringing ceremony at the @NYSE. They went from early-stage startup to the NYSE trading floor, and the next winner could be you. Ready to apply? The Startup Challenge is open now. 🚀 bit.ly/4i8S5gO

What if pipeline failures came with a clear diagnosis and suggested fix already waiting for review? 🔧 Our data team's AI agent opens a ready-to-review PR in under 16 minutes on average. A person still decides what ships. This is what Autonomous Data Engineering looks like👇 bit.ly/4iqexn2

Migrating Informatica PowerCenter means preserving the workflow DAG, not just converting code. Snowflake AIM maps tasks to Snowflake Tasks, dependencies to AFTER, sessions to EXECUTE DBT PROJECT and worklets to stored procedures. See the conversion path: bit.ly/3TYqaIo

Trusted AI answers start long before anyone builds an agent. ☑️ Klarent defined its business data once on Snowflake, then built agents with Snowflake CoCo so teams asking questions in Snowflake CoWork get the same trusted answer. Get the full story 👉 bit.ly/4s8qnph

Canva protects 265M+ people every month, and every alert has to be worth an analyst’s time. 🎯 @Canva’s security team processes hundreds of terabytes of logs every day and surfaces near-real-time alerts with Snowflake Interactive Analytics. Learn how: bit.ly/3WaUwYY

.@xai's Grok-4.7 and @Zai_org's GLM-5.3 are now in public preview on Snowflake Cortex AI across • Snowflake CoCo • Snowflake CoWork • Cortex AI Functions • Cortex Inference • Cortex Agents More model choice for coding, knowledge work and long-running agent tasks, right next to your governed data

Open table access shouldn’t require rebuilding data policies in every engine. Snowflake Horizon Catalog now supports reads and writes to policy-protected tables through the Iceberg REST Catalog API. It uses the next-generation policy evaluation engine with native support for the Iceberg Scan Plan API to enforce Snowflake row access and masking policies across compatible external engines. Use a single policy plane to govern access to Snowflake-managed and externally managed Iceberg tables uniformly across compatible engines: bit.ly/4hNPVEC

Why did IDC MarketScape name Snowflake a Leader in data intelligence? Three strengths were noted: 1️⃣ Native, unified governance within a broader data platform 2️⃣ Enterprise-grade resiliency and security 3️⃣ Deep partner ecosystem and industry alignment See the breakdown: bit.ly/4yv4sdx

Different dbt model groups need different schedules, but they don’t need duplicate project objects. Run slices of one dbt Project on Snowflake concurrently with --select. Use WRITEBACK = FALSE to prevent artifact collisions. Snowflake Data Superhero Satoshi Sagara's walkthrough: bit.ly/4dAzJnK