Snowflake launched a bevy of data and AI enhancements that position the company as "an agent of transformation."
The combination of launches at Summit 2025 highlight how Snowflake is expanding its lens for its platform ranging from analytics to data to AI. CEO Sridhar Ramaswamy said the company's revised mission statement is to "empower every enterprise to achieve its full potential through data and AI."
The big theme at Snowflake Summit 2025 is about stitching together data sets, tools and workflows to enable AI use cases. "Snowflake is where data does more," said Ramaswamy, noting the conference's mantra.
Ramaswamy noted that Snowflake is in a different place than it was a year ago due to strong results and a faster product cadence. "Over the past few years, we've been expanding our lens. Obviously, we started as an analytics platform, but we've been adding substantial functionality to that platform, extending it so that we can work with customers. We can help customers through more of their data life cycle," said Ramaswamy.
According to Ramaswamy, Snowflake is in a place where it can bring together AI and unstructured and structured data. Here's a look at the major announcements from Snowflake Summit 2025.
- Snowflake makes its Postgres move, acquires Crunchy Data
- Snowflake delivers strong Q1, Q2 outlook
- Snowflake adds OpenAI models to Cortex AI via expanded Microsoft partnership, integration
- Snowflake, Anthropic ink LLM partnership, delivers strong Q3, acquires Datavolo
Snowflake Adaptive Compute, platform enhancements
Snowflake announced the addition of Adaptive Compute to is data warehouse platform. The company also outlined Standard Warehouse-Generation 2 with 2.1x faster analytics performance over the previous generation, interoperability features and AI security and governance features in Snowflake Horizon Catalog.
Here's the breakdown:
- Snowflake Gen2 warehouses are retooled with upgraded hardware and enhanced software and generally available.
- Snowflake Adaptive Compute, which is in private beta, automatically scales resources and routes queries efficiently behind the scenes. Customers won't have to manually manage warehouse sizes, concurrency settings and multi-cluster configurations. Adaptive Compute will power Adaptive Warehouses.
- Snowflake Chief Product Officer Christian Kleinerman said this adaptive compute approach eliminates complexity from the platform. "We're bringing increased ease of use to the platform," said Kleinerman. "We are materially improving the price performance of the platform of Snowflake."
- Gen2 warehouses feature simplified conversion to adaptive warehouses without downtime.
- Interoperability across catalogs and engines with Catalog-Linked Databases, which will enable customers to automatically sync Snowflake Horizon Catalog with Apache Iceberg objects managed by any Iceberg REST Catalog including Apache Polaris and AWS Glue.
- Expanded Data Discovery with Universal Search, which is in private preview. Customers will be able to discover data in external relational databases including Postgres SQL, MySQL without leaving Snowflake.
- Pfizer was cited as an early adopter of Adaptive Compute.
AI-Driven data governance, security and compliance including the ability to ask security questions via Snowflake Cortex AI as well as a set of AI observability tools.
Snowflake Openflow
The company launched Snowflake Openflow, which simplifies the movement of data to where it can be used. Openflow is the evolution of Snowflake's Datavolo acquisition.
Snowflake said Openflow is a multi-modal data ingestion service that will give users the ability to connect any data source and architecture.
Openflow, which is powered by Apache NiFi, will enable customers to meld data engineering practices into their Snowflake workflows.
With Openflow, Snowflake is pursuing what it calls limitless interoperability to create data that's ready for AI uses. Openflow has multiple integrations with Box, Google Ads, ServiceNow, Workday and Zendesk to name a few.
Snowflake Intelligence, Data Science Agent
At Snowflake Summit 2025, executives also moved to outline how the company is leveraging AI for customer use cases and democratizing data for business users.
The company launched Snowflake Intelligence, which will be in public preview soon, to offer users a conversational experience with data agents that can traverse dashboards, structured and unstructured data stores and analytics tools to deliver answers.
In a demo, Snowflake showed how users in different functions--finance, marketing, manufacturing--could ask questions and run SQL queries without analysts and technical expertise in a compliant manner.
Snowflake Intelligence leverages the platform's connectors, Openflow and Cortex Knowledge Extensions that can bring in third party content from Snowflake partners.
According to Snowflake, Snowflake Intelligence uses large language models from Anthropic and OpenAI running inside Snowflake perimeter with Cortex Agents.
Snowflake also launched Data Science Agent, which will be in private preview. Data Science Agent automates machine learning model development tasks, and AI workflows to boost productivity.
Data Science Agent uses Anthropic's Claude model to break down problems into steps including data analysis, data prep, feature engineering and training.
Cortex AISQL, SnowConvert AI
Snowflake launched Cortex AISQL, which embeds generative AI into customer queries to analyze and build pipelines with SQL syntax. Snowflake claimed that Cortex AISQL can lead up to 60% cost savings when filtering or joining data.
According to Snowflake, Cortex AISQL can "effectively turn every data analyst into an AI engineer." Cortex AISQL uses models from Anthropic, Meta, Mistral and OpenAI and combines them with Snowflake's SQL Engine.
The secret sauce for Cortex AISQL is that analysts can query both structured and unstructured data to analyze multi-modal data, eliminate silos, consolidate tools and enrich customer tables.
SnowConvert AI is designed to make data warehouse, business intelligence and extract transform and load (ETL) migrations faster. Snowflake said SnowConvert AI makes code conversion and testing phases in migrating from legacy systems 2x to 3x faster.
Snowflake Marketplace enhancements
Snowflake launched Cortex Knowledge Extensions on Snowflake Marketplace to add content and data from The Associated Press, CB Insights and Stack Overflow. These third-party content partners can be used to enrich AI apps with real-time news.
Via Snowflake Marketplace, customers can share Semantic Models to integrate AI-ready structured data within their Snowflake Cortex AI apps.
Snowflake also rolled out Agentic Snowflake Native Apps on Snowflake Marketplace. Agentic Native Apps are interoperable and can deliver standalone experiences on customer or provider data or be used as building blocks for apps created on Cortex Agents or in Snowflake Intelligence.
