“Snowflake helps every organization easily combine data and AI to create business value.”

Snowflake, a global AI data cloud company, held a press conference at the AC Hotel on the 24th and shared the content presented at the 'Snowflake Summit 2025' held in San Francisco, USA from the 2nd to the 5th.

Before the full-scale launch, CEO Choi Ki-young stated that the core of this Snowflake Summit 2025 is the strategic expansion of functions in response to the needs of corporate customers, and that the first direction presented was to increase the return on investment (ROI) by strengthening the efficiency of data architecture.

In addition, he said that building a platform capable of sophisticatedly managing governance, which has become complex due to the influx of massive unstructured data beyond traditional typical data, has emerged as a major task, and that in a reality where unstructured data accounts for more than 80% of the total, the ability to freely access and process various data types has also been addressed as important.

In addition, the integration of AI technology is one of the key topics, and it should be introduced across all areas beyond piecemeal introduction. He said that the scenario where AI works organically was highlighted, and ultimately the demand for an application-based collaboration environment to be built quickly and easily was reflected.

In particular, the 'Marketplace', where thousands of data models and apps are registered, is a strategic choice by Snowflake to practically respond to these business needs, and users can utilize a platform where they can easily share and collaborate on data, just like an app store.

Next, Executive Director Louis Li introduced the major products announced at Snowflake Summit 2025.


Executive Director Louis Li stated that the core of the 'Snowflake Summit 2025' presentation was △ compute innovation △ data engineering advancement △ AI·ML integration △ analytics expansion △ strengthening of the app·collaboration ecosystem.

First, it was announced that the 'Adaptive Compute Elective Warehouse', which provides automated compute resource allocation, has been introduced.

Adaptive Warehouse analyzes workload patterns in real time and automatically routes them to the optimal compute pool. This allows users to maximize cost efficiency.

In addition, the '2nd generation standard warehouse' with comprehensive hardware and software upgrades has appeared.

Using new instances based on Z1 and Gen2 chips, BI workload concurrency processing performance has been improved by 2.3 times, and large-capacity deletion, update, and merge speeds have been improved by up to 4.4 times.

Credit costs will increase by 35%, but performance-to-cost on a workload basis will be significantly improved, resulting in lower total cost of ownership (TCO).

In the data engineering sector, we announced 'OpenFlow', an ingestion-based data ingestion management service.

We recently acquired Datavolo's technology and implemented it as a fully managed environment within Snowflake.

You can flexibly collect and load structured, unstructured, and semi-structured data with various deployment options, and a real-time monitoring dashboard is provided as standard so you can immediately understand data inflow statistics.

The continuous data collection feature, ‘Snowpipe Streaming’, has also been redesigned through a Rust-based SDK.

It supports 10GB of data collection per second per table and query latency of 5 to 10 seconds, further enhancing the real-time analysis environment.

It simplifies complex streaming pipelines and ensures cost savings with a transparent billing model.

The 'Workspace' IDE, which increases developer productivity, provides SQL, Python, and notebooks in a single integrated interface.

existing A web-based coding environment that can replace VS Code, and also includes file-based project management features.

In the AI·ML integration area, ‘Cortex AI’ and ‘Snowflake Intelligence’ were unveiled.

Cortex AI supports simultaneous processing of unstructured and structured data within SQL statements, and distributes 'Agent API' that can call external LLM functions in the form of APIs, enabling integration with Microsoft 365, MS Teams, etc.

In particular, Snowflake Intelligence is an enterprise-specific chatbot platform that supports natural language questions and answers by orchestrating agents from internal and external data sources.

In addition, 'SnowConvert AI' has expanded the automation scope by more than 95% by adding AI-based code conversion, verification, and validation functions to the migration process from existing databases to Snowflake.

Document understanding and query tool 'Document AI' is expected to see increased usability among domestic customers with the addition of Korean language support.

The application and collaboration ecosystem that has expanded to the marketplace is also noteworthy.

AI-ready datasets and semantic models are shared in a pre-embedded form, and developers can use them to create and monetize conversational analysis apps and agents.

“With this announcement, Snowflake has further concretized its vision of an ‘AI Data Cloud,’” said Executive Director Louis Lee. “We support all organizations to easily combine data and AI to create business value.”