Snowflake AI

Snowflake AI is the artificial intelligence research and product division of Snowflake, the publicly listed American enterprise data-cloud company headquartered in Bozeman, Montana with operations in San Mateo, California. Snowflake AI develops the Arctic open-weights mixture-of-experts foundation model, the Arctic-Embed text-embedding family, and the Cortex AI platform that integrates foundation-model capability into Snowflake's enterprise data-cloud platform. The Cortex AI platform supports both Snowflake-built models and third-party foundation-model integration (including OpenAI, Anthropic, Mistral AI, and other providers), and the broader Snowflake Intelligence platform extends Cortex into agentic AI workflows for enterprise customers.

At a glance

• Founded: Snowflake was founded in July 2012 by Benoit Dageville, Thierry Cruanes, and Marcin Zukowski. The dedicated Snowflake AI organization scaled substantially during the 2023 to 2024 period as Snowflake added foundation-model capabilities to the Cortex AI platform.
• Status: Subsidiary AI organization within Snowflake. Snowflake is publicly listed on the New York Stock Exchange (SNOW) since September 2020.
• Funding: Snowflake internal R&D budget. Snowflake's annual research-and-development investment exceeds $1 billion.
• CEO: Sridhar Ramaswamy, Chief Executive Officer of Snowflake since February 2024. Former senior VP at Google (head of advertising) and co-founder of Neeva (the AI-powered search engine acquired by Snowflake in 2023). Baris Gultekin leads the AI organization at Snowflake as Head of AI.
• Other notable leadership: Senior research-and-product leadership for the Cortex AI platform, the Arctic foundation-model family, and the Snowflake Intelligence agentic-AI platform. Frank Slootman served as CEO from 2019 through February 2024 before transitioning to Chairman.
• Open weights: Yes, partial. Arctic (April 2024) and Arctic-Embed (April 2024) are released open-weights through Hugging Face under the Snowflake organization. Cortex AI commercial flagships are gated through the Snowflake platform.
• Flagship outputs: Arctic (480-billion-parameter mixture-of-experts model with 16 billion active parameters, April 2024, open-weights), Arctic-Embed (text-embedding family, April 2024, open-weights), Cortex AI (the enterprise AI platform), Cortex AISQL (AI-powered SQL extensions), Snowflake Intelligence (the agentic AI platform).

Origins

Snowflake was founded in July 2012 by Benoit Dageville, Thierry Cruanes, and Marcin Zukowski to build a cloud-native data warehouse. The founders had earlier worked at Oracle and Vectorwise, and the founding thesis was that the cloud computing era required a fundamentally redesigned data warehouse architecture that decoupled storage and compute. The product launched in 2014, and Snowflake grew rapidly through the 2010s into a leading enterprise data-cloud company.

The September 2020 IPO valued Snowflake at approximately $33 billion at debut, making it the largest software IPO in history at the time. Through 2020 to 2023, Snowflake expanded the Data Cloud platform across analytics, application development, AI/ML workloads, data sharing, and other enterprise data capabilities.

The 2023 acquisition of Neeva for approximately $185 million was structurally significant. Neeva was an AI-powered search engine founded by Sridhar Ramaswamy (former Google senior VP of advertising) and Vivek Raghunathan; the acquisition brought Ramaswamy and the Neeva team into Snowflake. Ramaswamy initially served as SVP of AI at Snowflake, leading the Cortex AI platform development, before being appointed CEO in February 2024 succeeding Frank Slootman.

The April 2024 release of Arctic (the 480-billion-parameter mixture-of-experts model with 16 billion active parameters) was the principal Snowflake AI open-source contribution. Arctic was characterized at release as one of the leading open-weights enterprise-focused LLMs, with optimization for SQL co-pilot, chatbot, and other enterprise workloads. The Arctic-Embed text-embedding family released alongside as open-weights through Hugging Face.

The 2024 to 2026 period has seen Snowflake AI emphasize the Cortex AI platform expansion. Cortex AI provides foundation-model capability integrated with Snowflake's enterprise data-cloud platform, with both Snowflake-built models (Arctic) and third-party model integration (OpenAI, Anthropic, Mistral AI, Meta AI / FAIR, and other providers). Cortex AISQL extends the platform with AI-powered SQL extensions allowing data analysts to query structured and unstructured data using SQL syntax with foundation-model capability embedded.

The Snowflake Intelligence platform launched as the principal agentic-AI offering, with Cortex Code adjacent capabilities for AI-assisted enterprise development. Sridhar Ramaswamy has framed 2026 as "the year of the Great Decentralization" of AI, with predictions that big-tech AI walled gardens will give way to more decentralized AI deployment patterns.

Mission and strategy

Snowflake AI's stated mission is to bring AI capability to the enterprise data cloud, making AI accessible to data engineers, analysts, and other enterprise users without requiring deep AI expertise. The Cortex AI platform is the principal product expression of this mission, integrating foundation-model capability with Snowflake's existing enterprise data-cloud infrastructure.

The strategy combines four threads. First, foundation-model research through the Arctic and Arctic-Embed families, with selective open-source releases for research-community engagement and developer adoption. Second, the Cortex AI platform as the integrated commercial layer, providing foundation-model access (both Snowflake-built and third-party) within Snowflake's enterprise data-cloud platform. Third, Cortex AISQL as the principal interface for AI-augmented data analytics, allowing standard SQL syntax to query unstructured data and invoke foundation-model capability. Fourth, Snowflake Intelligence and Cortex Code as the agentic AI platform extending Cortex into autonomous and semi-autonomous AI workflows.

The competitive premise is that enterprise customers benefit most from AI capability deeply integrated with their data infrastructure, and that Snowflake's existing enterprise data-cloud customer base positions the company uniquely to deliver enterprise AI value. The Cortex AI integration with Snowflake's data foundation is structurally distinct from pure-play AI offerings that require enterprise customers to integrate AI capability with separate data infrastructure.

The post-2024 emphasis on agentic AI through Snowflake Intelligence positions Snowflake as one of the principal enterprise agentic AI platforms, alongside Microsoft Copilot, Salesforce Agentforce, Databricks Mosaic AI, and other enterprise AI offerings.

Models and products

• Arctic. Released April 2024. 480-billion-parameter mixture-of-experts foundation model with 16 billion active parameters. Open-weights through Hugging Face. Optimized for enterprise workloads including SQL co-pilot and chatbot applications.
• Arctic-Embed. Released April 2024. Open-weights text-embedding model family. Used in Cortex Search and other retrieval applications.
• Cortex AI. The enterprise AI platform integrating foundation-model capability with Snowflake's data cloud. Supports Snowflake-built and third-party models including OpenAI, Anthropic, Mistral AI, Meta AI / FAIR, and other providers.
• Cortex AISQL. AI-powered SQL extensions allowing standard SQL syntax to query structured and unstructured data including images, PDFs, and other unstructured content.
• Snowflake Intelligence. The agentic AI platform extending Cortex AI into autonomous and semi-autonomous workflows.
• Cortex Code. AI-assisted enterprise development platform. Part of the broader Snowflake Intelligence agentic AI capability.
• Cortex Search. Embedding-based search and retrieval capability built on Arctic-Embed.
• Cortex Analyst. AI-powered analytics agent capability for Snowflake's data customers.
• Project Snow Work. Outcome-driven AI coworker product for handling reports, slides, and workflows.

The principal commercial channels are direct enterprise sales for the Snowflake Data Cloud platform with Cortex AI included or as upgrades, partner-and-system-integrator distribution through the Snowflake ecosystem, and cloud-marketplace distribution.

Benchmarks and standing

Arctic was characterized at April 2024 release as one of the leading open-weights enterprise-focused LLMs, with reported strong performance on SQL generation, instruction-following, and other enterprise-relevant tasks. The model has been less prominent on subsequent benchmarks given the rapid evolution of open-source LLM capability through 2024 to 2026 and Snowflake's strategic emphasis on Cortex AI platform integration rather than continued frontier model capability investment.

The Cortex AI platform's commercial standing is anchored by Snowflake's enterprise customer base and the integration of AI capability with the existing Snowflake data foundation. Industry coverage frequently characterizes Snowflake AI as one of the principal enterprise AI platforms globally, alongside Databricks on the data-and-AI platform tier and alongside Salesforce AI Research on the enterprise application AI tier.

The company's standing in the global AI ecosystem is anchored on the enterprise customer base, the Cortex AI platform's data integration moat, the Sridhar Ramaswamy senior leadership and AI-strategic positioning, and the Arctic open-source contribution.

Leadership

As of April 2026, Snowflake AI's senior leadership includes:

• Sridhar Ramaswamy, Chief Executive Officer of Snowflake since February 2024. Former senior VP of advertising at Google (2007 to 2018). Co-founder of Neeva (acquired by Snowflake in 2023). Senior strategic leadership for Snowflake's overall direction and the AI platform expansion.
• Baris Gultekin, Head of AI at Snowflake. Senior research-and-product leadership for the Cortex AI platform, the Arctic foundation-model family, and the Snowflake Intelligence agentic-AI platform.
• Frank Slootman, Chairman of Snowflake (former CEO 2019 through February 2024). Provides senior board-level strategic leadership.
• Senior product-and-engineering leadership across Cortex AI, Arctic, Snowflake Intelligence, and other platform capabilities.

The AI organization includes senior product, engineering, and research talent, with recruitment from Google, Microsoft, AWS, and other technology incumbents alongside foundation-model researchers from academic and Insurgent organizations.

Funding and backers

Snowflake AI's capital structure is the publicly listed parent Snowflake. Specific R&D investment figures for the AI division are not separately disclosed, but Snowflake's annual research-and-development investment exceeds $1 billion as of fiscal year 2025 to 2026. Snowflake's market capitalization in 2025 to 2026 has been in the $40 billion to $80 billion range on the New York Stock Exchange.

The company's commercial revenue scale and cash flow support continued AI investment alongside product engineering and commercial operations. Snowflake Ventures, the company's strategic-investor arm, has positions across the broader AI ecosystem including investments in Mistral AI, Anthropic, and other organizations.

Industry position

Snowflake AI occupies a structurally distinctive position in the global enterprise AI landscape. The combination of the foundational Snowflake data-cloud platform, the Cortex AI integrated AI platform, the Arctic open-source contribution, the Snowflake Intelligence agentic-AI extension, and the Sridhar Ramaswamy AI-strategic leadership produces a profile that competes directly with Databricks on the data-and-AI platform tier and complements Salesforce on the enterprise application AI tier.

Industry coverage frequently characterizes Snowflake as one of the principal enterprise AI platforms globally, with the Cortex AI platform's data-foundation integration as the principal differentiator versus pure-play AI offerings. The post-2024 emphasis on agentic AI through Snowflake Intelligence positions the company strongly in the agentic-AI commercial market.

Strategic risks include intensifying competition from Databricks on the data-and-AI platform tier, from cloud-platform AI services (AWS, Azure, Google Cloud), and from frontier-AI labs expanding enterprise capabilities. Strategic strengths include the enterprise customer base, the Cortex AI integrated platform value proposition, the Sridhar Ramaswamy AI-strategic leadership, and the post-2024 agentic-AI commercial positioning.

Competitive landscape

Snowflake AI competes with several enterprise AI organizations:

• Databricks (Mosaic AI). Direct enterprise data-and-AI competitor. Snowflake's Cortex AI and Databricks's Mosaic AI compete head-to-head on the enterprise data-and-AI platform tier.
• Salesforce AI Research (Einstein, Agentforce). Enterprise AI competitor through the CRM-and-business-application focus.
• Microsoft AI (Copilot, Azure AI). Direct enterprise AI competitor through Microsoft 365 Copilot and Azure cloud AI services.
• Amazon AGI (AWS Bedrock). Direct enterprise AI competitor through AWS Bedrock and other AWS AI services.
• Google DeepMind, OpenAI, Anthropic, Cohere, Mistral AI. Frontier-AI labs whose enterprise APIs are integrated into Cortex AI as third-party model providers.
• Together AI, NVIDIA Research. AI infrastructure providers competing on the foundation-model training and deployment tier.
• Oracle, SAP, ServiceNow. Enterprise software incumbents deploying AI capabilities across their respective product portfolios.

Outlook

Several open questions affect Snowflake AI's trajectory in 2026 and 2027:

• The continued expansion of the Cortex AI platform and the AI-product revenue trajectory.
• The Snowflake Intelligence agentic-AI platform expansion and commercial traction against Microsoft Copilot, Salesforce Agentforce, and other enterprise agentic-AI offerings.
• The competitive dynamic with Databricks on the enterprise data-and-AI platform tier.
• The trajectory of Arctic and other Snowflake-built foundation-model research.
• Continued senior research-and-engineering talent recruitment.
• Sridhar Ramaswamy's continued strategic leadership and the realization of the "Great Decentralization" thesis for 2026 and beyond.
• The Snowflake Ventures strategic-investor portfolio evolution.

Sources

• Snowflake official site. Company overview reference.
• Snowflake: Arctic announcement blog post. April 2024 Arctic release.
• Snowflake: Cortex AI. Cortex AI platform reference.
• TechCrunch: Snowflake releases a flagship generative AI model of its own. Arctic release coverage.
• Fortune: Snowflake CEO Sridhar Ramaswamy 7 predictions for enterprise AI in 2026. 2026 strategic predictions.
• Wikipedia: Snowflake Inc.. Comprehensive company history reference.
• Snowflake on Hugging Face. Open-source model distribution.