Snowflake, Microsoft Partner to Integrate OpenAI’s Models in New AI Development Tools

BOZEMAN, Mont. — Snowflake, the AI Data Cloud company, announced an expanded partnership with Microsoft on February 26, 2025, enabling access to OpenAI‘s language models through its Cortex AI development environment. This integration will allow joint customers to develop generative AI applications while maintaining data security within Snowflake’s platform.

Microsoft has had a close relationship with OpenAI since it began investing in the developer in January 2023, shortly after the launch of ChatGPT. The new integration with Snowflake will leverage the Microsoft Azure OpenAI Service, allowing OpenAI’s models to be utilized securely in the Snowflake ecosystem. Stephen Catanzano, an analyst at Enterprise Strategy Group, highlighted the significance of this integration for enterprises. ‘It simplifies AI deployment, enhances data security, and supports multimodal AI use cases — text, audio, video — enabling businesses to leverage advanced AI capabilities without complicated integrations.’

In addition to deploying OpenAI’s models in Cortex AI, the expanded partnership includes features available in Microsoft 365 Copilot and Microsoft Teams, with the Cortex Agents feature in public preview. General availability across Microsoft applications is anticipated by June 2025.

Increasingly, enterprises are turning to generative AI applications due to their potential to enhance worker efficiency and improve information accessibility. However, to tailor these applications effectively, organizations must integrate generative AI models with their proprietary data. Catanzano noted that combining data with models is critical for instilling “intelligence” in AI systems. ‘The integration is a significant boost for Snowflake and Microsoft, particularly as enterprises seek streamlined environments for innovative AI solutions,’ he added.

Snowflake’s Cortex AI currently includes integrations with various AI vendors’ models, including those developed by Snowflake, AI 21 Labs, Anthropic, DeepSeek, Meta, Mistral, and now OpenAI. The availability of multiple language models gives developers in the Snowflake environment greater flexibility, according to Andy Thurai, an analyst at Constellation Research.

Thurai stated that while many companies now face competitive pressures to innovate, those offering comprehensive platform services, like security and data governance, are better positioned for financial success. ‘Models will become less differentiated soon enough,’ he noted. ‘Companies can make money by providing trusted services on their platforms.’

Baris Gultekin, Snowflake’s head of AI, pointed to customer feedback as a key driver for adding OpenAI’s language models directly into the Snowflake environment. ‘Customers want to run the best models in the security of Snowflake’s data boundary to focus on AI insights rather than managing data transfers,’ he explained.

Despite these advancements, analysts suggest that Snowflake still needs to catch up with competitors like Databricks in terms of AI development capabilities. Catanzano cautioned that while Snowflake has integrated significant features, its model training capabilities still trail behind those of its rivals. ‘Snowflake has made major strides, but ongoing development in custom model training and more partnerships with AI vendors will be crucial for their growth,’ he stated.

Looking ahead, Snowflake aims to focus on implementing agentic AI strategies, which can proactively assist businesses with tasks that previously required human input. Gultekin believes such advancements will become essential within the enterprise ecosystem. ‘AI agents will soon be integral to the workforce, driving operational efficiencies,’ he concluded.