Snowflake’s 34% Surge Pushes Cortex AI Into Microsoft 365

Snowflake posted product revenue of $1.334 billion for the quarter that ended April 30, up 34% from a year earlier and roughly $74 million above its own guidance, with the company crediting fast-rising demand for its Cortex AI tools. The data-cloud vendor lifted its full-year outlook on the strength of the result.

For the millions of people who work inside Microsoft 365 every day, the line that mattered was not the revenue beat. It was that Snowflake’s Cortex Agents now answer questions directly inside Microsoft 365 Copilot and Teams, and that the OpenAI models doing much of that work reach customers through Microsoft’s own cloud.

Snowflake Cleared Its Own Guidance by $74 Million

The first quarter of fiscal 2027 was the company’s strongest sequential dollar gain on record. Management had guided to about $1.26 billion in product revenue; the business delivered $1.334 billion instead. Total revenue, which adds professional services, came in at $1.391 billion, up 33%.

The headline metrics tell a consistent story of acceleration rather than a one-off spike:

• 126% net revenue retention rate, a sign that existing customers keep spending more each year.
• $9.21 billion in remaining performance obligations (RPO, the value of contracted revenue not yet recognized), up 38%.
• 779 customers now generate more than $1 million in trailing 12-month product revenue, a 29% increase.
• 616 net new customers joined in the quarter, up 38% from a year earlier.

Sridhar Ramaswamy, Snowflake’s chief executive, framed the period as a turning point rather than a good run.

AI continues to be a powerful tailwind for Snowflake, and Q1 marks a clear inflection point in that journey.

The company raised its full-year product revenue target to $5.84 billion, which would represent 31% growth. That is a step up from the 29% pace it ran through fiscal 2026, and it sits well above the 26% growth the same quarter showed a year ago. You can read the detail in Snowflake’s first quarter fiscal 2027 results release.

Cortex Agents Reach Users Inside Microsoft 365 Copilot and Teams

The number that should catch a Microsoft administrator’s eye is not on the income statement. It is the count of accounts now touching Snowflake’s machine-learning layer.

What the Integration Puts in Front of Office Workers

More than 13,600 accounts are now using Snowflake’s AI capabilities, and the Cortex Code assistant alone is live in over 7,100 of them. The bridge to Microsoft’s user base comes through a native app that pipes those agents into the apps people already keep open all day. Once an administrator connects an account, business users can do the following without leaving the workflow:

• Ask plain-language questions of governed Snowflake data and get answers back inside a Teams chat.
• Pull both structured tables and unstructured files, such as contracts and invoices processed by Document AI, into a single response.
• Generate charts and summaries through Copilot without writing a database query.

Snowflake’s setup notes spell out the plumbing. When an account sits outside the Azure US East 2 region, administrators get a consent prompt acknowledging that the bot backend processes prompts and responses through that region. The full configuration steps live in the Cortex Agents setup guide for Microsoft Teams and Copilot.

Why the Distribution Reach Counts

Plenty of data tools claim a Microsoft tie-in. Few get to ride inside the chat window where a sales manager already spends the morning. By delivering its agents through Microsoft’s AppSource marketplace, Snowflake reaches non-technical staff who would never open a data warehouse console, which is exactly the audience that turns a back-office platform into daily habit.

How Azure Became Cortex’s Model Pipeline

Cortex AI is the layer that connects a company’s internal databases to large language models built by outside firms, so teams can build custom software faster. The choice of which models, and which cloud delivers them, is where Microsoft quietly gained ground this year.

In February, Snowflake and OpenAI signed a multi-year deal worth about $200 million to embed OpenAI’s frontier models natively across the data platform. Those models reach Snowflake customers in part through Azure OpenAI Service inside Azure AI Foundry, with Snowflake’s Horizon Catalog handling governance and security. A separate $200 million expansion with Anthropic, struck the prior December, brings Claude models to the same customers through Microsoft’s cloud alongside Amazon and Google options.

Cortex AI now spans models from seven vendors, including Snowflake’s own Arctic system plus AI21 Labs, DeepSeek, Meta and Mistral. On top of the raw models sit packaged tools: Document AI, which reads disorganized files like legal contracts and invoices and extracts data points, and Cortex Code, a conversational agent that takes plain-language commands and executes complex database operations with full view of a user’s Snowflake setup. An industry analysis of the expanded Microsoft partnership notes that Snowflake now hosts both Anthropic and OpenAI systems under one governance umbrella.

The Fabric Question Microsoft Cannot Sidestep

Here is the awkward part of the relationship. The more Snowflake threads itself through Copilot, Teams and Microsoft’s cloud, the more it competes with the analytics platform Microsoft sells to the same buyers.

That platform is Microsoft Fabric, a bundled suite that folds data engineering, warehousing, real-time analytics and Power BI into one Azure-native service. Snowflake takes the opposite stance, running across Amazon Web Services, Microsoft’s cloud and Google Cloud so a customer is not locked to one vendor. The two products meet head-on in most enterprise data deals, and many large firms end up running both.

The honest read is that Microsoft is both Snowflake’s distribution channel and its rival, and both sides have decided the partnership is worth more than the overlap. You can see Microsoft’s side of the pitch on the official Microsoft Fabric platform overview, which leans hard on Power BI and tight Microsoft 365 ties. For an admin choosing between them, the deciding factor is usually how committed the organization already is to one cloud.

A $6 Billion AWS Pact Complicates the Cloud Math

While Microsoft captures the front-end attention, Amazon still anchors much of the back end. Alongside the quarterly numbers, Snowflake disclosed a multi-year agreement with Amazon Web Services valued at about $6 billion to speed enterprise AI adoption.

That figure dwarfs the two $200 million model deals and underlines a split worth watching: the heavy compute commitment flows to Amazon, while the everyday user experience increasingly surfaces through Microsoft’s apps. Snowflake gets to play all three hyperscalers against each other, which is the whole point of staying cloud-neutral.

The next test arrives quickly. If the Copilot and Teams integration drives real adoption among office workers this quarter, Microsoft strengthens its claim as the place enterprise AI gets used, even on data it does not own. If usage stays thin and the $6 billion of Amazon compute does the heavy lifting, the Microsoft tie-in looks more like a convenient shop window than a strategic moat.

Frequently Asked Questions

Is the Snowflake Cortex Agents integration for Microsoft 365 Copilot and Teams available now?

Yes. The integration is generally available across Snowflake’s public cloud regions and is deployed through the Microsoft AppSource marketplace. An administrator connects a Snowflake account, after which business users can query data from inside Teams and Copilot.

What does Snowflake Cortex AI actually do?

Cortex AI connects a company’s governed Snowflake data to large language models from seven vendors, including OpenAI, Anthropic, Meta and Mistral, so teams can build AI applications and agents without moving sensitive data out of the platform.

Does Snowflake run on Microsoft Azure?

Yes. Snowflake runs on Microsoft’s cloud as well as Amazon Web Services and Google Cloud. OpenAI models inside Cortex AI reach customers in part through Azure OpenAI Service in Azure AI Foundry.

How fast is Snowflake growing?

Product revenue rose 34% year over year to $1.334 billion in the first quarter of fiscal 2027. The company raised its full-year product revenue target to $5.84 billion, which would mark 31% annual growth.

Is Microsoft Fabric a competitor to Snowflake?

Yes, and a partner at the same time. Microsoft Fabric competes for the same enterprise data workloads, yet Microsoft also distributes Snowflake’s agents through Copilot and Teams. Many enterprises run both products side by side.

Which AI models can Snowflake customers use today?

Customers can choose among models from seven providers: Snowflake’s own Arctic, plus AI21 Labs, Anthropic’s Claude, DeepSeek, Meta, Mistral and OpenAI’s GPT family.