All Data and AI Weekly #253 - 03 August 2026
All Data and AI Weekly #253 - 03 August 2026
AI, Data, Agentic AI, Cortex Code, NiFi, Iceberg, Polaris, Streamlit, Python, Java, SQL, MCP, LLM, RAG, Cortex AI, AIFN, Search, Unstructured Data
Tim's Projects
SNACK-AI CoCo RAG Contract Search - A full RAG-powered contract search application built with Snowflake Cortex Code and SNACK-AI patterns. https://github.com/tspannhw/SNACKAI-CoCo-RAGContractSearch
SNACK-AI CoCo Agent Testing - An agent testing framework built to validate and iterate on Snowflake Cortex Agents. https://github.com/tspannhw/SNACKAI-CoCo-AgentTesting
Snowflake AI Pulse - July 2026
The July AI Pulse covered major announcements including Dynamic Table Adaptive Refresh, Cortex Agent Sharing, Auto Agents, Per-User Cost Quotas, Managed Agents, Online ML with real-time feature serving, CoCo Desktop and Cloud Agents GA, and AI Function Studio. 84 minutes of demos and Q&A from the Snowflake Product and AI teams. https://www.snowflake.com/en/ai-pulse/july-2026/
Snowflake CoCo: Built to Scale Enterprise AI with Trust
CoCo Desktop is now GA on macOS and Windows. Cloud Agents are GA in Snowsight. New AI cost governance with per-user quotas. Agent Settings lets admins configure the CoCo experience and roll it out across teams. CoCo Mobile (iOS and Android) is in private preview. Skills and Plugins sharing is coming soon. Cost per prompt cut by ~28% with no quality degradation through REPL-based tool calling, compact output, on-demand skill loading, and smarter model routing. https://www.snowflake.com/en/blog/snowflake-coco-built-to-scale-enterprise-ai/
Run Apache Spark Pipelines on Snowflake with CoCo
One prompt migrates entire Spark codebases to Snowpark Connect. Customers report up to 5.1x faster performance and 42% lower costs. The spark-migration skill handles assessment, inventory, targeted fixes, reporting, and end-to-end validation automatically. Supports Python, Scala, Java, and notebook files. https://www.snowflake.com/en/blog/migrate-spark-to-snowflake/
From Iceberg to Intelligence: The AI-Ready Borderless Lakehouse with Snowflake and Google Cloud
Deep dive into the Snowflake + Google Cloud open lakehouse architecture. Covers bidirectional catalog federation between Snowflake Horizon and Google Lakehouse runtime catalog, vended credentials, MCP servers for AI agent access, Semantic View Autopilot, and how CoWork and CoCo operate over federated Iceberg tables with full governance. https://www.snowflake.com/en/blog/snowflake-google-cloud-open-lakehouse/
TPC-H: Snowflake Runs 100GB in Under 11 Seconds
Snowflake Interactive Warehouses process all 22 TPC-H queries on 100GB in less than 11 seconds. At 1TB (10x data), runtime only doubles. The post breaks down the 28 "choke points" of TPC-H, explains micro-partitioning, dynamic pruning, runtime Bloom filters, and how Interactive Warehouses at 0.6 credits/hour deliver blazing performance at low cost. https://www.snowflake.com/en/blog/engineering/snowflake-tpc-h-benchmark-performance0/
Cross-Region AI Inference, Data Residency and Sovereignty
An architecture deep-dive into how Snowflake Cortex AI handles inference across regions. Covers self-hosted models on Snowflake GPUs, cross-region mTLS encryption, frontier model access via CSP endpoints, data classification (Customer Data vs Usage Data vs AI Data), GPU tenant isolation, and Horizon Catalog governance that executes before any data reaches a model. https://www.snowflake.com/en/blog/engineering/cross-region-ai-inference-data-sovereignty/
Apache Polaris 1.6: Security Hardening and Live Migrations
The latest release of the open-source Iceberg REST Catalog brings JWT issuer validation, access token redaction, configurable AWS STS session names, stricter entity name validation, zero-downtime catalog migrations via register-table-with-overwrite, view registration, a production-ready event pipeline with ordered delivery and field-level redaction, Spark 4.0 client, Python 3.14 support, and CLI REPL mode. https://www.snowflake.com/en/blog/engineering/apache-polaris-1-6-release/
How Thrive Learning Improved Dynamic Tables Price-Performance by 2X
Thrive Learning cut Dynamic Table costs from hundreds of credits/day to ~7 credits/day using Custom Incrementalization (REFRESH_MODE = CUSTOM_INCREMENTAL). Auto-clustering dropped from ~150 to ~2 credits/day. Transformation compute dropped from ~150 to ~5 credits/day. The pipeline stayed in dbt lineage with no separate orchestration layer. https://www.snowflake.com/en/blog/thrive-dynamic-table-costs-custom-incremental/
Consuming Delta Shares in Horizon Catalog (GA)
Delta Sharing consumption via Catalog Linked Database is now generally available. Horizon Catalog automatically converts Delta tables to Iceberg format. Two SQL statements to get started. Full governance (masking, row access policies, data controls) applies to consumed Delta shares. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-21-delta-sharing-horizon-catalog-ga
AI_TRANSCRIBE Now Supports M4A and MOV Files
Snowflake Cortex AI_TRANSCRIBE adds support for M4A audio files (including Apple Lossless ALAC) and MOV video files. Transcribe recordings from Apple devices, screen recordings, and professional video production assets directly in Snowflake with a single SQL query. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-31-ai-transcribe-m4a-mov-support
Enterprise AI Security: Agentic MCP Governance
How Snowflake secures the agentic enterprise with MCP governance, covering trust boundaries, tool authorization, and policy enforcement for AI agents. https://www.snowflake.com/en/blog/enterprise-ai-security-agentic-mcp-governance/
Securing the Agentic Enterprise
Snowflake's approach to securing AI agents at enterprise scale with unified monitoring and cost management. https://www.snowflake.com/en/blog/securing-the-agentic-enterprise/
Snowflake Advances Trusted Agentic Enterprise with Unified Monitoring and Cost Management
Press release on new capabilities for governing AI agents, including unified observability, cost controls, and trust boundaries. https://www.snowflake.com/en/news/press-releases/snowflake-advances-the-trusted-agentic-enterprise-era-with-unified-monitoring-and-cost-management/
Multi-Engine Lakehouse Interoperability Series
A four-part engineering series on how Snowflake achieves lakehouse interoperability across engines:
Part 1: Data Interoperability Guide https://www.snowflake.com/en/blog/engineering/lakehouse-data-interoperability-guide/
Part 2: Business Logic Interoperability https://www.snowflake.com/en/blog/engineering/lakehouse-business-logic-interoperability/
Part 3: Data Governance Interoperability https://www.snowflake.com/en/blog/engineering/lakehouse-data-governance-interoperability/
Part 4: Multi-Engine Lakehouse Interoperability https://www.snowflake.com/en/blog/engineering/multi-engine-lakehouse-interoperability/
Snowflake Postgres: ML Online Feature Store
Using Snowflake Postgres for real-time ML online feature serving with low-latency lookups. https://www.snowflake.com/en/blog/snowflake-postgres-ml-online-feature-store/
Simpler, Safer URLs: Snowflake Per-Account Hostnames
End-to-end production implementation of per-account hostnames for cleaner, more secure Snowflake URLs. https://www.snowflake.com/en/blog/engineering/snowflake-per-account-urls/
Dynamic Tables Custom Incrementalization (GA)
Custom Incremental refresh mode is now generally available, giving full control over merge logic while Snowflake manages scheduling, retries, and lag. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-27-dynamic-tables-custom-incremental-ga
Data Quality Monitoring Dashboard (Public Preview)
A built-in dashboard for monitoring data quality metrics across your Snowflake account. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-29-data-quality-monitoring-dashboard-pupr
Data Clean Rooms Update
New capabilities for Snowflake Data Clean Rooms collaboration. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-30-dcr
Cortex Code AI Suggestions (GA)
AI-powered code suggestions in Cortex Code are now generally available. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-28-cortex-code-ai-suggestions-ga
CoCo Desktop v1.20.2
Latest CoCo Desktop release with improvements and fixes. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-28-coco-desktop-v1-20-2
Query History: Agent Type Column
New AGENT_TYPE column in query history for tracking AI agent activity. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-27-query-history-agent-type-column
MCP External OAuth Support
MCP servers now support External OAuth for authentication. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-22-mcp-external-oauth
Data Movement Policies
New policies for controlling how data moves within and across accounts. https://docs.snowflake.com/en/release-notes/2026/other/2026-08-17-data-movement-policies
PrivateLink Endpoint Multiple Hostnames (Azure)
Azure PrivateLink endpoints now support multiple hostnames. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-29-privatelink-endpoint-multiple-hostnames-azure
AI Extract/Parse Document: CSE Network Restrictions
Network restriction support for AI document extraction with client-side encryption. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-27-ai-extract-parse-document-cse-network-restrictions
AI Extract/Parse Document: GCP Dammam Region
AI document processing now available in GCP Dammam region. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-27-ai-extract-parse-document-gcp-dammam
CoWork Automations
Automations for Snowflake CoWork now available. https://docs.snowflake.com/en/release-notes/2026/other/2026-07-15-cowork-automations
https://docs.snowflake.com/en/LIMITEDACCESS/snowflake-cowork/automations
Cortex Search Regional Availability
Updated regional availability for Cortex Search services. https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-search/cortex-search-overview#label-cortex-search-overview-regional-availability
Community and Medium Articles
Hands-On: Exploring Multi-Omics Analysis with Snowflake CoCo https://medium.com/snowflake/hands-on-exploring-multi-omics-analysis-with-snowflake-coco-bc5cf61022be
Recursive Queries in Snowflake https://medium.com/snowflake/recursive-queries-in-snowflake-b18880b87d3f
Set Up Playbook: Unleashing Notebooks Python in Workspaces https://medium.com/snowflake/set-up-playbook-unleashing-notebooks-python-in-workspaces-a1c1d0bee757
GitLab CI/CD Best Practices for Snowflake Workspaces https://medium.com/snowflake/gitlab-ci-cd-best-practices-for-snowflake-workspaces-0b113b8161f0
The Browser Inside My IDE Replaced My QA Workflow https://snowflakechronicles.medium.com/the-browser-inside-my-ide-replaced-my-qa-workflow-16ecd9a3357d
Industry News
MCP Protocol Blog Update https://blog.modelcontextprotocol.io/posts/2026-07-28/
NVIDIA Open Secure AI Alliance https://blogs.nvidia.com/blog/open-secure-ai-alliance/
AAIF.io - AI Applications and Infrastructure Forum https://aaif.io/
Tools and Fun
Matrix Rain - A matrix rain animation for your browser. https://anshs.github.io/matrix-rain/
Turbo Fieldfare - Open source tool from the community. https://github.com/drumih/turbo-fieldfare
Video
Snowflake Engineering Deep Dive https://www.youtube.com/watch?v=Q1ZuGgzsTO4
SNACK-AI: The 2026 Pattern
As we look forward in 2026, SNACK-AI is the architectural pattern for powering AI applications at any scale - from mobile phones and robots to massive enterprise clusters.
SNACK-AI integrates:
Snowflake Platform
NiFi (Apache)
Apache Iceberg
Cortex AI (Agents/Search/MCP)
Kafka (Apache Kafka)
Read: SNACK-AI: The 2026 Pattern (Medium) https://medium.com/@tspann/snack-ai-d8c5dd692929
Code: Official SNACK-AI GitHub Repository https://github.com/tspannhw/SNACK-AI
Weekly Context: SNACK-AI Overview https://github.com/tspannhw/FLiPStackWeekly/blob/main/2025/BestOf2025/SNACKAI.md
Thanks
https://sessionize.com/tspann https://github.com/timothyspann
