TL;DR
Snowflake has added Google’s Gemini models to its Cortex AI stack, allowing customers to run Gemini inside their Snowflake environments across supported clouds. The integration is available as a public preview for select Gemini versions and aims to let businesses query company data with a choice of foundational models while keeping data inside Snowflake.
What happened
Snowflake announced native support for Google’s Gemini models within its Cortex AI environment, enabling customers to run Gemini across Snowflake deployments on AWS, Azure and Google Cloud via cross-region inference. The integration is presented as part of Snowflake’s model-agnostic approach, which already includes OpenAI, Anthropic, Deepseek, Meta and Mistral. Snowflake said customers can access Gemini inside Cortex using SQL or APIs to analyze multimodal corporate data — from ERP and CRM records to documents, images, audio and IoT telemetry — without moving data outside the platform. A public preview of Gemini 3 Pro and Gemini 2.5 Flash is available through Cortex AI Functions as of Tuesday. The company said pricing details were not being disclosed but indicated usage would follow Snowflake’s broader consumption model, and that support for Snowflake Intelligence and Cortex Agents is coming soon.
Why it matters
- Brings a major Google foundation model inside a popular enterprise data platform, reducing the need to export sensitive data for LLM processing.
- Model choice inside Snowflake could simplify workflows for organizations that want to experiment with or standardize on different LLM providers.
- Running models natively across supported clouds helps enterprises maintain governance and data residency controls while using advanced AI capabilities.
Key facts
- Snowflake has added native support for Google’s Gemini models inside its Cortex AI stack.
- The integration supports cross-region inference across Snowflake environments on AWS, Azure and Google Cloud.
- A public preview includes Gemini 3 Pro and Gemini 2.5 Flash via Cortex AI Functions, available as of Tuesday.
- Cortex AI already offers access to models from OpenAI, Anthropic, Deepseek, Meta and Mistral.
- Users can interact with models directly from SQL or via APIs to analyze multimodal enterprise data.
- Snowflake stores many operational and content data types including ERP, CRM, IoT, documents, images, video and audio.
- Snowflake described its approach as model choice without re-architecting data or workflows.
- Pricing was not disclosed; Snowflake said costs align with its broader consumption-based billing model.
- Support for Snowflake Intelligence and Snowflake Cortex Agents to use Gemini is planned soon.
What to watch next
- Rollout timeline and availability of Gemini support inside Snowflake Intelligence and Cortex Agents (support coming soon).
- How enterprise customers evaluate Gemini’s performance and accuracy on their internal, multimodal datasets — not confirmed in the source.
- Specific pricing and commercial terms for running Gemini in Snowflake beyond the stated consumption model — not confirmed in the source.
Quick glossary
- Large Language Model (LLM): A type of AI model trained on vast text datasets to generate or analyze human-like language and perform related tasks.
- Foundation model: A broadly trained model that can be adapted to many downstream tasks, often forming the base for specialized applications.
- Cross-region inference: Running model predictions across multiple cloud regions so that users can access models regardless of where their data or compute reside.
- Multimodal data: Information that includes multiple types of media, such as text, images, audio, and video, which models can process together.
- Consumption-based pricing: A billing model where customers pay based on how much of a service they actually use rather than a fixed upfront fee.
Reader FAQ
Is Google Gemini now available inside Snowflake?
Yes. Snowflake added native support for Gemini inside Cortex AI and announced a public preview for select Gemini versions.
Which Gemini versions can customers access?
A public preview includes Gemini 3 Pro and Gemini 2.5 Flash via Cortex AI Functions as of Tuesday.
Will using Gemini in Snowflake require moving data outside the platform?
Snowflake says Gemini runs natively within Cortex AI so customers can process data without exporting it; further deployment details are not confirmed in the source.
How much will it cost to run Gemini in Snowflake?
Snowflake did not disclose specific pricing; it said costs will follow the company’s broader consumption-based model.
AI + ML Snowflake gets frosty with Google Gemini Munge that corporate data using the LLM of your choice O'Ryan Johnson Tue 6 Jan 2026 // 23:54 UTC The data platform Snowflake is putting…
Sources
Related posts
- CES 2026 roundup: Nvidia’s Rubin and Alpamayo, AMD Ryzen AI 400, Razer oddities
- Google Voice to receive Gemini Notes for recorded calls and transcripts
- Viral Reddit ‘whistleblower’ post alleging delivery-app fraud was AI-generated