Snowflake Summit 2026 (June 1-4, Moscone Center, San Francisco) brought 20,000+ attendees and 26+ new product capabilities centered on a single thesis: AI agents need enterprise data, governed context layer, and verified identity to be useful at scale. Anthropic President Daniela Amodei co-keynoted alongside CEO Sridhar Ramaswamy. Every major announcement (Snowflake CoWork, Snowflake CoCo, Snowflake Cortex Sense, Snowflake Horizon Context, AI agent identity) extended Snowflake’s stack toward the agentic enterprise.
| Field | Detail |
|---|---|
| Event | Snowflake Summit 2026 |
| Dates | June 1-4, 2026 |
| Location | Moscone Center, San Francisco |
| Theme | The Agentic Enterprise |
| Attendees | 20,000+ |
| Total new capabilities | 26+ (announced across press releases) |
| Key keynote guests | Sridhar Ramaswamy (Snowflake CEO), Daniela Amodei (Anthropic President) |
| Sessions / labs | 500+ breakout sessions, 39 hands-on labs |
Complete Snowflake Summit 2026 announcements
Permalink to “Complete Snowflake Summit 2026 announcements”Snowflake announced 26+ new capabilities across six domains at Summit 2026: AI agents and productivity (CoWork, CoCo), context and semantics (Horizon Context, Cortex Sense), security and governance (Agent Identity, AI Security Posture Management), data infrastructure and interoperability (Apache Iceberg v3, Datastream, Openflow), AI infrastructure (Snowflake Cortex Training, Adaptive Compute), and partnerships (Anthropic, AWS, Natoma). The table below is a complete inventory (Snowflake interoperability press release; individual product press releases).
| Announcement | Category | Status | What It Does | Atlan’s Read |
|---|---|---|---|---|
| Snowflake CoWork | AI Agents + Productivity | GA | Rebrand of Snowflake Intelligence; personal AI agent for knowledge workers with Artifacts, Deep Research, User Memory, User Skills, Skill Catalog, and MCP connectors for Google Drive, Salesforce, and Slack | CoWork’s answers are only as good as the semantic context it draws from, Atlan’s Enterprise Data Graph and Context Agents supply that foundation |
| Snowflake CoCo | AI Agents + Productivity | GA | Rebrand of Cortex Code; coding agent for enterprise AI development with autonomous task execution, Automations, Cloud Agents, CoCo Desktop/Mobile/Slackbot, VS Code Extension, and integrations with Vercel, Retool, and Superblocks | Atlan’s Context Engineering Studio validates the semantic layer CoCo queries before it ships to production |
| CoWork Slackbot | AI Agents + Productivity | GA | Brings CoWork into Slack for in-workflow data queries and governed dashboards | n/a |
| CoWork iOS App | AI Agents + Productivity | GA | Native iOS interface for CoWork; extends knowledge worker AI to mobile | n/a |
| Microsoft Excel Extension (CoWork + CoCo) | AI Agents + Productivity | GA | CoWork and CoCo surface directly in Microsoft Excel for in-spreadsheet data queries and code generation | n/a |
| Skill Catalog | AI Agents + Productivity | GA | Enterprise-wide discovery and reuse of proven CoWork and CoCo skills and solutions | Mirrors the context-as-reusable-asset pattern that Atlan’s Context Engineering Studio formalizes for semantic context |
| Snowflake Horizon Context | Context + Semantics | Private Preview (Wave 1) | New capability inside Horizon Catalog: Collect (5 metadata connectors, PostgreSQL, SQL Server, Tableau, Power BI, dbt), Enrich (column-level lineage, AI-generated docs, popularity metrics), Activate (hybrid semantic + keyword search, automatic semantic view discovery) | Snowflake ships 5 connectors in private preview. Atlan ships 100+ in GA. Atlan pulls Horizon Context metadata into the Enterprise Data Graph and enriches it with cross-system lineage |
| Semantic Studio | Context + Semantics | Private Preview | Enables teams to define shared business logic without SQL expertise | Atlan’s Context Engineering Studio ships the same capability, GA, with CI-integrated eval suites and context-as-code versioning |
| Semantic View Autopilot | Context + Semantics | Public Preview | Automatically creates and refines semantic views over time | Complements Atlan’s Context Agents, which generate and maintain semantic descriptions across all connected systems |
| Cortex Sense | Context + Semantics | GA (within CoWork + CoCo) | Runtime layer that automatically assembles data, business definitions, and operational knowledge for AI agents; 86% accuracy on structured questions with full business context vs. 24% generic models | Cortex Sense reaches within Snowflake’s boundary. Atlan’s Enterprise Data Graph reaches across 100+ systems. Cortex Sense + Atlan = agents with full enterprise context |
| Open Semantic Interchange (OSI) | Context + Semantics | Specification finalized | Open standard for sharing semantic models across data tools; 54 participating vendors including Atlan, Collibra, dbt Labs, Databricks, ThoughtSpot, Informatica, and BlackRock | Atlan is a founding member and participant in OSI; context flows between Atlan and Snowflake via the standard |
| AI Agent Identity | Security + Governance | GA | Cryptographic, verifiable identity for every AI agent; per-agent RBAC (not inherited from user credentials); complete agent audit trails; dynamic data masking by agent type | IAM (Okta, Entra, AI Agent Identity) tells you WHO the agent is. Atlan governs WHAT the agent knows and is allowed to know, the context governance layer IAM vendors don’t address |
| AI Security Posture Management | Security + Governance | Public Preview | Continuous monitoring of AI system security; ML-driven threat detection; protection against data exfiltration, ransomware, prompt injection, jailbreak; includes Data Exfiltration Policies, Multi-Party Authorization, Cortex Guard | n/a |
| Connected Audit Access | Security + Governance | Private Preview | Unified audit access within Horizon Catalog for cross-system visibility | Atlan’s Enterprise Data Graph already provides cross-system lineage and audit trail context across 100+ connectors |
| External Engine Access Management | Security + Governance | Private Preview | Controls access to Snowflake data from external compute engines (e.g., Spark, Databricks) | n/a |
| Apache Iceberg v3 | Infrastructure + Interoperability | GA (May 2026) | Broadest feature support claim: deletion vectors, row lineage, variant type (semi-structured data), default values | Atlan’s Context Lakehouse is Iceberg-native, the world’s first context store engineered natively for AI; Iceberg v3 GA makes the integration richer |
| Snowflake Storage for Apache Iceberg Tables | Infrastructure + Interoperability | GA | Single live governed copy of data across Snowflake and external lakes without data movement; Horizon Catalog powered by Apache Polaris with bi-directional read/write | n/a |
| Zero-Copy Integrations | Infrastructure + Interoperability | Public Preview | Native integrations for SAP, Salesforce, Workday, AVEVA, IBM with no data duplication | As more source systems connect zero-copy, context from those systems becomes a gap Atlan’s 100+ connectors fill |
| Snowflake Datastream | Infrastructure + Interoperability | Preview | Managed Apache Kafka-compatible streaming service native to Snowflake; drop-in Kafka replacement; streaming data inherits Snowflake governance and lineage automatically | n/a |
| Snowflake Openflow | Infrastructure + Interoperability | GA | Managed data integration service on Apache NiFi; batch + streaming; BYOC or Snowflake-managed; AI-powered migration assistance | n/a |
| Cortex Training | AI Infrastructure | Preview | Managed GPU infrastructure for fine-tuning open-weight models (Qwen, Mistral) inside Snowflake; data never moves; up to 2x more training runs per equivalent GPU budget; supports reinforcement learning and domain-specific customization | Fine-tuned models still need accurate, certified context to query, Atlan’s context layer is the semantic foundation those models operate against |
| Adaptive Compute | AI Infrastructure | GA (soon) | Automatically optimizes compute and software resource mix for AI workloads in real-time; removes manual tuning | n/a |
| SpaceX Grok (Cortex AI addition) | AI Infrastructure | GA | Grok models added to the Cortex AI model menu (joining Claude, GPT, Gemini, Llama, Mistral, DeepSeek) | More model choice amplifies the need for a single certified context source, the Enterprise Data Graph, so every model answers from the same truth |
| Auto-gen Agents for Data Shares and Listings | AI Infrastructure | Public Preview | Automatically generates agents for Snowflake Marketplace datasets | n/a |
| Natoma acquisition | Partnerships + Ecosystem | Intent announced May 27, confirmed at Summit | Enterprise MCP platform acquisition; Natoma provides centralized MCP gateway with identity verification, access policies, and audit controls at the tool-call level | Snowflake is building an MCP governance layer for external tool access. Atlan’s MCP server already exposes the Enterprise Data Graph to any agent, a complementary surface, not the same thing |
| Snowflake + Anthropic expanded partnership | Partnerships + Ecosystem | GA | Claude (3.5 Sonnet+) powers Cortex AI, CoWork, and CoCo; Daniela Amodei keynoted; Claude Marketplace launch partner integration | Atlan’s context layer grounds Claude-powered agents in certified enterprise semantics, Snowflake + Anthropic + Atlan is the complete stack |
| Snowflake + AWS $6B commitment | Partnerships + Ecosystem | Announced | $6B, 5-year infrastructure commitment to AWS Graviton + AI; expansion to 10 new AWS regions | n/a |
| Snowflake + Collibra expanded integration | Partnerships + Ecosystem | Preview (GA targeted Q3 2026) | Bi-directional metadata flow: Collibra governed metadata, tags, and policies into Horizon Catalog and Cortex; Snowflake technical metadata returns to Collibra; semantic models aligned to OSI published as semantic views | Atlan similarly layers on top of Snowflake Horizon, the pattern Snowflake is validating with Collibra is the same architecture Atlan ships today |
| dbt Labs + Fivetran merger | Partnerships + Ecosystem | Announced at Summit | Fivetran and dbt Labs are now a single company; joint vision: “trusted agents using open data infrastructure principles”; dbt Core v2.0 in alpha (Snowflake, BigQuery, Databricks, Redshift support) | Atlan connects to dbt via the Enterprise Data Graph, lineage and semantic models from the merged dbt+Fivetran stack flow into Atlan’s context layer |
| Thomson Reuters on Snowflake | Partnerships + Ecosystem | Announced | Enterprise AI platform on Snowflake; workloads up to 3.4x faster; moving from static reporting to near real-time insights within a governed environment | n/a |
See how Atlan layers on Snowflake as the context plane.
Read the ArchitectureWhat each major announcement means for data teams
Permalink to “What each major announcement means for data teams”Not every Summit announcement carries the same weight for enterprise data teams. Four stand out as genuinely decision-relevant: CoWork’s rebrand signals Snowflake’s intent to own knowledge worker AI workflows; Horizon Context defines where Snowflake’s semantic layer begins and ends; AI Agent Identity sets a new baseline for AI agent access control; and Cortex Sense establishes runtime context as a first-class primitive (Constellation Research; SiliconANGLE; InfoWorld).
Snowflake CoWork: from chatbot to knowledge worker operating system
Permalink to “Snowflake CoWork: from chatbot to knowledge worker operating system”CoWork is a full rebrand and capability expansion of Snowflake Intelligence, now positioned as “the personal agent for knowledge workers.” The additions that change the architecture: Artifacts (governed dashboards with live Snowflake data that persist and can be shared across the org), User Memory (AI agent memory from historical behavior for role-tailored recommendations over time), and the Skill Catalog (enterprise-wide discovery and reuse of proven agent workflows). The new interfaces (iOS app, Slackbot, Excel Extension, MCP connectors for Google Drive, Salesforce, and Slack) move CoWork from a platform feature to a cross-surface agent. The practical implication: CoWork is now Snowflake’s primary surface for governing how knowledge workers interact with enterprise data. For the full picture, see the Snowflake CoWork deep-dive (Snowflake CoWork press release).
Horizon Context: the semantic catalog and where its reach ends
Permalink to “Horizon Context: the semantic catalog and where its reach ends”Horizon Context is a new capability inside Horizon Catalog with three parts: Collect (5 external metadata connectors in private preview Wave 1: PostgreSQL, SQL Server, Tableau, Power BI, dbt), Enrich (column-level lineage stitching, AI-generated documentation, popularity metrics), and Activate (hybrid semantic and keyword search, automatic semantic views discovery). Semantic Studio enables teams to define shared business logic without SQL expertise; Semantic View Autopilot auto-generates semantic view DDL from existing tables. The OSI standard (54 participating vendors) provides the interoperability layer that makes Horizon Context open. The honest read: 5 connectors in private preview versus 100+ in GA is a real gap for enterprises with heterogeneous stacks. “Wave 1” implies a multi-wave rollout; the agent context layer fills the gap today. See Snowflake Horizon Context for the full architecture (Horizon Context blog; Snowflake Horizon Catalog press release).
AI Agent Identity: the new security primitive for agentic AI
Permalink to “AI Agent Identity: the new security primitive for agentic AI”Traditional IAM was designed for humans, not autonomous software agents. AI Agent Identity (now GA) delivers: cryptographic, verifiable identity for every agent before data access; per-agent RBAC that is not inherited from user credentials; dynamic data masking by agent type; and complete AI agent observability audit trails of all agent activities. As Christian Kleinerman, EVP of Product at Snowflake, stated: “Traditional security models were designed for human users, not autonomous software agents.” The implication is clear: cryptographic agent identity is now table stakes for any production agentic deployment in a Snowflake environment. See the AI Agent Identity guide for what this means for your architecture (Snowflake Horizon Catalog press release).
Cortex Sense: runtime context as a product
Permalink to “Cortex Sense: runtime context as a product”Cortex Sense is a runtime layer that automatically assembles data, business definitions, and operational knowledge for AI agents at query time. The accuracy benchmark: 86% on structured questions with full business context versus 24% with generic frontier models alone (Snowflake CoWork press release). What distinguishes Cortex Sense from a static semantic layer is that it assembles context dynamically, not from a pre-built cache. What it does not include: cross-system metadata, business glossary terms from other sources, or context from the 15-30 other tools in a typical enterprise stack. For the full boundary analysis, see Cortex Sense.
How these announcements fit together: the agentic strategy
Permalink to “How these announcements fit together: the agentic strategy”Every announcement at Summit 2026 extends a single architecture: data plane at the bottom (Apache Iceberg v3, Datastream, Openflow, Adaptive Compute), a semantic and context layer in the middle (Horizon Context, Snowflake Cortex, Semantic Studio), and agentic surfaces at the top (CoWork, CoCo). Snowflake is building from the data plane up. That direction is clear, but SiliconANGLE’s “System of Intelligence is still being built” assessment captures where the gaps remain (SiliconANGLE; Medium/DataAIChronicles).
Sridhar Ramaswamy’s keynote thesis was the organizing frame for everything that followed: “The model is not your unique advantage. It’s when you combine models with your data that things begin to shine.” The four building blocks from his keynote: Enterprise Data and Context, AI Models, Applications, and an Agentic Control Plane. Every product announcement at Summit maps to one of these four. CoWork and CoCo are the Applications layer. Horizon Context and Cortex Sense are the Context layer. AI agent governance and Cortex Training are the Control Plane layer. Iceberg v3, Openflow, and Datastream are the Data layer.
SiliconANGLE’s five-layer System of Intelligence framework places Snowflake strongest at the Data Foundation layer, actively building toward the System of Intelligence. The identified gaps from their analysis: cross-system business process logic modeling; enterprise-wide entity resolution; intent-based AI governance framework beyond data-level controls; institutional knowledge capture and reasoning. Snowflake has acknowledged these gaps by building Horizon Context, Cortex Sense, and acquiring Natoma. But acknowledging the gaps and shipping GA solutions are different timelines. The gaps SiliconANGLE identifies are Atlan’s lanes.
Constellation Research’s Mike Ni captured the strategic shift precisely: Summit 2026 shows Snowflake “shifting from data format ownership to customer relationship ownership.” The announcements prioritize meaning, trust, permissions, context, and governance over format. That is the right direction. It is also the direction Atlan has been building for years.
Deep dives: each announcement explained
Permalink to “Deep dives: each announcement explained”Each of the pages below covers one Summit announcement in full: architecture, limitations, how it fits the enterprise AI stack, and where Atlan complements or extends it. Start with whichever announcement is most relevant to your current evaluation.
-
Snowflake CoWork: the agentic enterprise’s knowledge worker layer, What changed from Snowflake Intelligence, how Artifacts and User Memory work, and why the Skill Catalog matters for enterprise knowledge management at scale.
-
Snowflake CoCo: the coding agent built for enterprise AI development, How CoCo (formerly Cortex Code) differs from a code copilot: autonomous task execution, Cloud Agents, and the Automations engine that runs workflows without local infrastructure.
-
Cortex Sense: the runtime context layer and where its boundary ends, The 86% accuracy benchmark explained, what runtime context assembly means technically, and what Cortex Sense still needs from outside Snowflake’s boundary.
-
Snowflake Horizon Context: the semantic catalog and its first five connectors, Full architecture breakdown of Collect/Enrich/Activate, what the five Wave 1 connectors tell you about the roadmap, and where Horizon Context sits relative to a full-stack context layer.
-
AI Agent Identity: the new Snowflake security primitive, Why cryptographic identity for agents is different from human IAM, how per-agent RBAC works, and what audit trail completeness means for regulated enterprise deployments.
-
Snowflake Semantic Views: the core primitive Snowflake is building its context layer on, Why Snowflake’s CPO called semantic views a core primitive, how Semantic View Autopilot changes the creation workflow, and what OSI means for portability.
-
Atlan + Snowflake architecture, The answer to “where does Atlan fit if I already have Horizon Context and Cortex Sense?” Data plane vs. context plane, with a concrete architecture breakdown.
-
Snowflake Cortex Training: fine-tune open-weight models inside Snowflake, What managed GPU fine-tuning means for engineering teams, which models are supported, and what the 2x training-run efficiency claim means in practice.
-
Apache Iceberg v3 GA: what actually changed for your data lake, Deletion vectors, row lineage, variant type, default values: what each feature unblocks for AI workload patterns on open table formats.
Related Atlan resources:
- Context layer for Snowflake, The full Atlan architecture for Snowflake environments, including the enterprise data graph, Context Agents, and Context Lakehouse.
- What is context engineering?, The discipline Summit 2026 validated: why AI accuracy depends on the quality of context layer for Snowflake Cortex, not just the model.
- Snowflake data governance explained, How governance works natively in Snowflake and where the Enterprise Data Graph extends it.
- Snowflake Summit 2025 announcements, Year-over-year context: what Snowflake committed to in 2025 and what shipped by Summit 2026.
See how Atlan extends Snowflake's context layer to 100+ systems.
Atlan for SnowflakeWhat Snowflake Summit 2026 means for enterprise AI
Permalink to “What Snowflake Summit 2026 means for enterprise AI”- Snowflake Summit 2026 launched 26+ new capabilities under one organizing thesis: agentic AI requires governed enterprise context layer, not just compute.
- CoWork and CoCo (rebrands of Snowflake Intelligence and Cortex Code) are now Snowflake’s primary agent surfaces: one for knowledge workers, one for developers.
- Horizon Context, Cortex Sense, and Semantic Studio show Snowflake building its own semantic layer from the data plane up.
- AI Agent Identity (GA) sets a new security baseline: every agent needs cryptographic identity, per-agent RBAC, and a complete audit trail.
- Apache Iceberg v3 GA, Adaptive Compute, Datastream, and Openflow build the infrastructure substrate the agentic layer runs on.
- The OSI standard (54 vendors, finalized spec) signals the industry has accepted that semantic interoperability requires open formats, not proprietary lock-in.
Every announcement at Summit 2026 is Snowflake moving up the AI stack from the data plane. Atlan is the context layer for Snowflake that meets it: 100+ connectors, 690,000+ AI-generated descriptions, and a Context Lakehouse engineered natively for AI, all serving certified enterprise context to Snowflake’s agents via MCP server for Snowflake today.
FAQs about Snowflake Summit 2026
Permalink to “FAQs about Snowflake Summit 2026”1. What did Snowflake announce at Summit 2026?
Permalink to “1. What did Snowflake announce at Summit 2026?”Snowflake announced 26+ new capabilities at Summit 2026 (June 1-4, Moscone Center) spanning six domains: AI agents and productivity (CoWork, CoCo), context and semantics (Horizon Context, Cortex Sense, Semantic Studio, OSI), security and governance (AI Agent Identity, AI Security Posture Management), infrastructure and interoperability (Apache Iceberg v3, Datastream, Openflow), AI infrastructure (Cortex Training, Adaptive Compute), and partnerships (Natoma acquisition, Anthropic expanded, AWS $6B). Anthropic President Daniela Amodei co-keynoted alongside CEO Sridhar Ramaswamy under the organizing theme of the agentic enterprise (Snowflake interoperability press release).
2. What is the difference between Snowflake CoWork and the old Snowflake Intelligence?
Permalink to “2. What is the difference between Snowflake CoWork and the old Snowflake Intelligence?”CoWork is a full rebrand and capability expansion of Snowflake Intelligence, now positioned as the personal agent for knowledge workers. The functional additions that matter: Artifacts (governed live-data dashboards that persist and can be shared), User Memory (personalization from historical behavior), Deep Research (multi-step reasoning across structured and unstructured data with citations), and the Skill Catalog (enterprise-wide discovery and reuse). New interfaces (iOS app, Slackbot, Excel Extension) move CoWork from a platform feature to a cross-surface agent that follows knowledge workers into every workflow (Snowflake CoWork press release).
3. What happened to Cortex Code? Is it the same as CoCo?
Permalink to “3. What happened to Cortex Code? Is it the same as CoCo?”Yes. CoCo (Snowflake CoCo) is the direct rebrand of Cortex Code. The additions beyond the rebrand: autonomous task execution (not just code generation), Cloud Agents (background execution without a local laptop), and Automations (recurring event-driven workflows). CoCo had 7,100+ users before the relaunch, making it the fastest-growing Snowflake product at the time of the rebrand. The multi-surface interfaces (Desktop, Mobile, Slackbot, VS Code Extension, Excel Extension) and integrations with Vercel, Retool, and Superblocks complete the picture (Snowflake CoCo press release).
4. What is Snowflake Horizon Context and how is it different from a data catalog?
Permalink to “4. What is Snowflake Horizon Context and how is it different from a data catalog?”Horizon Context is a new capability within Horizon Catalog that adds semantic meaning and active context for AI and BI, not just metadata storage. The three-part architecture (Collect/Enrich/Activate) moves beyond traditional catalog functions: Collect ingests from 5 external systems in private preview Wave 1, Enrich adds column-level lineage and AI-generated documentation, and Activate serves context dynamically to AI agents and BI tools via hybrid semantic and keyword search. Traditional data catalogs index metadata; Horizon Context is designed to activate it at agent query time (Horizon Context blog; Snowflake Horizon Catalog press release).
5. Did Snowflake make any acquisitions at Summit 2026?
Permalink to “5. Did Snowflake make any acquisitions at Summit 2026?”Snowflake announced its intent to acquire Natoma, an enterprise MCP (Model Context Protocol) platform. Natoma provides a centralized MCP gateway that enforces identity verification, access policies, and audit controls at the tool-call level, governing what actions AI agents can take against APIs, workflows, collaboration systems, and operational systems, not just data. The intent was announced May 27, 2026, and the acquisition rationale was confirmed on the Summit stage (Natoma acquisition press release).
6. Where does Atlan fit in Snowflake’s Summit 2026 announcements?
Permalink to “6. Where does Atlan fit in Snowflake’s Summit 2026 announcements?”Atlan layers on top of Snowflake as the context plane, delivering at GA scale what Summit’s announcements are building toward. Specifically: Atlan ships 100+ connectors versus Horizon Context’s 5 in preview; Context Engineering Studio is GA versus Semantic Studio in private preview; Context Agents have generated 690,000+ descriptions across 50+ enterprise customers versus Cortex Sense’s runtime enrichment within Snowflake’s boundary. Atlan is a founding member of the Open Semantic Interchange working group, a Snowflake partner, and Snowflake’s 2025 Data Governance Partner of the Year. Summit 2026 validates Atlan’s category (Atlan product stats; Snowflake Partner of the Year reference).
Sources
Permalink to “Sources”- Snowflake. “Snowflake Pioneers New Open Framework for Interoperable Enterprise Data and AI.” snowflake.com/en/news/press-releases/snowflake-pioneers-new-open-framework-for-interoperable-enterprise-data-and-ai/. June 2026.
- Snowflake. “Snowflake CoWork Powers the Agentic Enterprise as the Personal Agent for Knowledge Workers.” snowflake.com/en/news/press-releases/snowflake-cowork-powers-the-agentic-enterprise. June 2026.
- Snowflake. “Snowflake CoCo Redefines Enterprise AI Development.” snowflake.com/en/news/press-releases/snowflake-coco-redefines-enterprise-ai-development. June 2026.
- Snowflake. “Snowflake Advances Trusted AI with Snowflake Horizon Catalog.” snowflake.com/en/news/press-releases/snowflake-advances-trusted-ai-with-snowflake-horizon-catalog. June 2026.
- Snowflake. “Horizon Context: Building Governed Context for the AI Enterprise.” snowflake.com/en/blog/horizon-context-governed-context/. June 2026.
- Snowflake. “Snowflake and Anthropic Accelerate Enterprise AI Adoption.” snowflake.com/en/news/press-releases/snowflake-and-anthropic-accelerate-enterprise-ai-adoption. June 2026.
- Snowflake. “Snowflake Expands AWS Collaboration with $6B Commitment.” snowflake.com/en/news/press-releases/snowflake-expands-aws-collaboration-with-6b-commitment. June 2026.
- Snowflake. “Snowflake Announces Intent to Acquire Natoma.” snowflake.com/en/news/press-releases/snowflake-announces-intent-to-acquire-natoma. May 27, 2026.
- SiliconANGLE. “Snowflake moves up the AI stack, but the System of Intelligence is still being built.” siliconangle.com/2026/06/02/snowflake-moves-ai-stack-system-intelligence-still-built/. June 2026.
- Constellation Research. Ni, Mike. “Snowflake Summit 2026: Context, custom model training, Iceberg V3.” constellationr.com/insights/news/snowflake-summit-2026-context-custom-model-training-iceberg-v3. June 2026.
- InfoWorld. “What Snowflake Summit 2026 signals about enterprise AI.” infoworld.com/article/4178831/what-snowflake-summit-2026-signals-about-enterprise-ai.html. June 2026.
- Medium/DataAIChronicles (Vedprakash). “Inside the Snowflake Summit 26 Opening Keynote: What Was Said, What Shipped, and Where Data Is Heading.” June 2026.
- Thomson Reuters. “Thomson Reuters Powers Trusted Enterprise AI at Scale on Snowflake.” snowflake.com/en/news/press-releases/thomson-reuters-powers-trusted-enterprise-ai-at-scale-on-snowflake/. June 2026.
