No longer is data just a concern to accountants. Now, with AI, data promises to transform how every department in an organization operates.
“AI represents a new way of thinking about how different functions within a company operate,” said Snowflake CEO Sridhar Ramaswamy in a virtual press conference.
So Snowflake is gearing up for this wider use from its 10,000 customers. At Snowflake Summit 2025, being held this week in San Francisco, Snowflake will be unveiling a new set of tools, as well as some upgrades to the existing platform to better help its enterprise customers apply AI to their own data.
The company has overhauled its flagship warehouse platform for performance improvements, as well as unveiled a new low-management platform to streamline analysis.
To speed data ingestion, the company also introduced a new data integration platform, called OpenFlow, and a query language that is based on SQL and supercharged with AI models (“AISQL”)
And of course, no company conference would be complete without some new agents.
Snowflake Adaptive Compute
Snowflake has pressed into service 125 new product capabilities in its first quarter alone, and this show continues the work.
A new service currently in private preview, Snowflake Adaptive Compute aims to simplify resource management.
Typically, customers manually specify pre-defined virtual warehouse sizes (“extra small,” “medium,” “large,”) and managing concurrency settings or multi-cluster configurations. With Adaptive Compute, customers will specify only a set of policies, such as what queries can run, and for how long.
Snowflake itself determines what compute resources are needed and then automatically supplies them.
“We’ve learned that at scale, for many customers, the effort and the process of deciding which queries go to which compute cluster and what is the right size — it affects the ease of use and cost of operation of the platform,” said Christian Kleinerman, Snowflake EVP of product, in a press briefing.
The company is redefining its compute models, starting first with an “Adaptive Warehouse” model, and moving on to adaptive compute clusters next, Kleinerman said.
The service also takes advantage of some of Snowflake’s FinOps innovations, notably Cost-Based Anomaly Detection, now in public preview.
Found in the Cost Management interface, this service sends alerts when abnormal cost spikes occur.
Standard Warehouse – Generation 2
Those who want to retain finer control of their Snowflake usage will have a lot to look forward to as well.
Now in general availability, Standard Warehouse — Generation 2 (Gen2) updates Snowflake’s virtual Standard Warehouse with new hardware, performance enhancements and software optimizations.
Due to these various enhancements, Snowflake now boasts 2.1x faster analytics performance for core analytics workloads compared to the previous generation, 2.3x higher throughput for Concurrent BI workloads, and up to 4.4x faster Delete, Update, and Merge operations.
Open Flow
With the abundance of data that organizations now have moving about, it makes sense for Snowflake to beef up its data integration tools. To this end, the company introduced OpenFlow, a data integration service.
The company is hoping that by streamlining the process of data movement, this software will enable businesses to enable AI applications more easily.
It is built on Apache NiFi, and outfitted with enterprise governance, security, deployment flexibility, and observability.
It can work with any data source, not only on that data on Snowflake’s cloud, but from other sources as well. It works with all data types, including both structured and unstructured data. It offers out-of-the-box connectors, from commercial sources (Box, Google Ads, Oracle, Proofpoint, ServiceNow, Workday, Zendesk) and custom connectors can be built as well.
OpenFlow can be deployed in the customer’s VPC, in a Bring Your Own Cloud (BYOC) model for AWS’ S3. It can run, as a managed service, in Snowflake Snowpark Container Services (in private preview)
Snowflake OpenFlow
SQL for AI
As part of its Snowflake Cortex suite of AI tools, Snowflake has introduced a new tool, AISQL, which enables the analysis of all data using the SQL format. This works with both structured and unstructured data.
In effect, the company points out, it brings AI directly into enterprise queries.
“The goal of this is to bring AI and the power of AI to analysts and personas that are typically comfortable with database technology but may not be fully versed or deep into how AI works,” Kleinerman said.
Agentic AI
Snowflake is keeping its customers up with the agentic AI craze as well.
One service about to be previewed, called Snowflake Intelligence, will offer a conversational interface to query their structured and unstructured documents. Under the hood, it uses the Anthropic’s Claude AI service to navigate internal data repositories and deliver insights.
“I think everybody here understands the value and importance of data. But there’s a ton of work that it takes to gain those insights and to find that potential,” said Jeff Hollan, Snowflake head of apps and developer platform. “We’re super excited to deliver a solution that helps every one of those business users reach the full potential of data by having a conversation with an AI agent that is securely governed and managed inside of their Snowflake.”
A second specialized agent will also be previewed in the near future, Data Science Agent, that is being designed to help users automate routine ML model development tasks, such as data analysis, data preparation, feature engineering, and training.
According to the company, Data Science Agent uses Anthropic’s Claude to help create the ML workflows into distinct steps
Shortly, users will also be able to build their own agents, through a service called Cortex Agents.
The post Snowflake Streamlines Data Analysis for Enterprise AI appeared first on The New Stack.
Unveiling new tools and services at its user conference this week, Snowflake wants to bring AI to all the departments in your organization.