From June 2 to 5, 2025, San Francisco hosted a new edition of the Snowflake Summit, one of the main global benchmarks in the field of cloud data and applied artificial intelligence.
This year, under a theme focused on AI and intelligent applications, the meeting brought together thousands of technology professionals, business lead
From June 2 to 5, 2025, San Francisco became the global epicenter of data and AI innovation as it hosted a new edition of the Snowflake Summit — a premier event for the cloud data and applied artificial intelligence community.
This year’s theme focused on AI and intelligent applications, bringing together thousands of data professionals, technology leaders, and business executives to explore the latest innovations in the Snowflake ecosystem.
As a leading cloud data platform, Snowflake enables organizations to unify, process, share, and activate massive volumes of data in real time — with scalability, security, and performance at its core. Today, it’s a strategic cornerstone for data-driven companies across industries seeking to accelerate decision-making and scale AI initiatives effectively.
The 2025 Snowflake Summit introduced a wide range of innovations, including new generative AI capabilities, powerful technology integrations, expanded partner ecosystems, and enhanced features designed to maximize the business value of data in increasingly complex environments.
If you couldn’t attend the event, don’t worry — we’ve got you covered. In this article, we break down the 15 most important announcements from Snowflake Summit 2025. Whether you're leading a data team or shaping your organization’s digital transformation strategy, these insights are essential to understand how Snowflake is reshaping the future of data and AI.
Snowflake Summit is the company’s flagship annual event, where the most important updates to the Snowflake platform are unveiled. It serves as a strategic moment for showcasing the company’s vision and the future of cloud data and AI.
At Snowflake Summit 2025, under the leadership of CEO Sridhar Ramaswamy, one of the standout moments was the keynote conversation with Sam Altman, CEO of OpenAI — reinforcing the message that generative AI applied to data will be the key innovation driver in the cloud computing space over the coming years.
The event introduced a redesigned data platform aimed at making data work even more seamless — helping businesses unlock value through AI at every stage of the data lifecycle.
Ramaswamy emphasized that Snowflake is now more intuitive, enabling faster connections between business users and the data, people, applications, and AI agents they need to drive results.
In his words, the company’s mission is to “reimagine data engineering, analytics, application development, collaboration, and enterprise AI” — all while maintaining a trusted foundation for data security and privacy.
Curious why so many forward-thinking companies are doubling down on Snowflake? Keep reading to discover the innovations that are transforming how modern enterprises operate with data.
Snowflake Summit 2025 was more than a technology conference: it was a roadmap for organizations that want to lead in a world increasingly dominated by artificial intelligence, automation and data culture. Over the course of four intense days, Snowflake unveiled multiple improvements that directly impact how companies use, share, analyze and monetize their data.
Below, we explore the 15 most important news, each with a clear business impact.
Snowflake Intelligence is a new conversational interface layer that represents a before and after for business users.
For the first time, anyone —without technical knowledge— can interact with your organization's data using natural language and receive clear, actionable answers and, if desired, execute actions on that data. All this happens without writing a single line of code.
On a technical level, it works with advanced language models (LLMs) from OpenAI and Anthropic, but most importantly, it runs within Snowflake's secure environment, ensuring data protection and compliance with governance policies.
This tool transforms the way leaders make decisions: what used to require a query to the BI team or a specific report can now be resolved in seconds.
Semantic Views allow you to define business metrics once, and have them applied consistently across any BI tool or AI wizard within the enterprise.
This puts an end to one of the most common sources of friction in organizations: having multiple definitions for the same metric ("active customers", "churn", etc.).
For business leaders, this semantic layer means that decisions are made on consistent, auditable, cross-area aligned data, which improves analytical quality, avoids arguments, and strengthens confidence in internal dashboards.
With the new Adaptive Warehouses, Snowflake introduces a compute system that automatically scales according to load, without the need for pre-configuration.
Being somewhat similar to Smoothing in Microsoft Fabric, this means that when a load increases (for example, in accounting closings or campaigns), resources are adjusted to maintain performance, and reduced when they are not needed.
For CFOs and CIOs, this translates into operational efficiency without compromising performance, with billing optimized for actual usage.
Migrating SQL code, ETL processes and data structures from legacy platforms to Snowflake can be a costly and risky task.
With SnowConvert AI, the company offers a tool that automatically analyzes existing code, translates it into Snowflake format, validates it and optimizes it for execution.
This allows companies to accelerate their transition to the cloud without compromising quality or relying excessively on manual development, reducing time, costs and errors.
The new Horizon Catalog not only catalogs Snowflake's internal data, but also external sources, dashboards, models and semantic assets. In addition, it includes an AI co-pilot that enables natural language querying of the environment's metadata: which datasets contain sensitive information, who accesses which data?
These types of capabilities reinforce regulatory compliance, data security and corporate transparency, three fundamental pillars for regulated industries or listed companies.
One of the big breakthroughs of the year has been the introduction of Cortex AISQL, an extension of theSQL language that allows users to work directly with unstructured data - such as PDF documents, images or free text - within the Snowflake database.
This means that Snowflake users can now generate insights from previously difficult-to-analyze files without the need for external tools or complex processes.
For the business, this opens the door to a new class of automations: from analyzing legal contracts to automatically extracting tables from scanned documents, all in the same governed cloud environment.
AI ceases to be a "side project" and becomes an integrated and accessible capability for multiple areas of the enterprise.
Snowflake now supports the secure sharing of semantic models and AI-ready datasets across departments — and even between separate organizations.
This fosters collaborative data ecosystems, allowing partners, subsidiaries, or joint ventures to build common solutions without duplicating efforts, while maintaining data governance and security.
For enterprises with complex ecosystems, this unlocks faster innovation and alignment across business units
Thanks to new partnerships with media and content platforms (USA Today, Stack Overflow, CB Insights, etc.), Snowflake will enable your AI applications to consume up-to-date external knowledge without leaving the secure environment.
These Cortex Knowledge Extensions ensure respect for intellectual property and provide appropriate attribution.
The result: more contextualized and useful artificial intelligence that not only responds with internal data, but also understands the external environment, ideal for competitive intelligence, market analysis or advanced technical support.
Snowflake introduces a new type of application: Snowflake Native Apps with agentic capabilities. That is, apps that incorporate built-in AI agents to execute specific tasks within the Data Cloud. These apps can automate data flows, respond to queries or generate complex analytics.
Most importantly, companies will be able to install and use them without the need for in-house development, which accelerates innovation and reduces barriers to entry.
The Data Science Agent is an artificial intelligence solution aimed at data science teams. It is another AI agent that automates many of the repetitive and technical tasks that normally slow down the development of AI models.
From data preparation and cleaning, to variable engineering, to model training, everything can be executed by describing in natural language the goal of the project.
This not only improves the productivity of the technical team, but also accelerates the time-to-value of AI models, allowing them to reach production sooner and generate value faster.
It also democratizes access to artificial intelligence within organizations that do not yet have large or mature data science teams.
With Openflow, Snowflake launches its own open data ingestion service, based on Apache NiFi, which allows information to be moved from virtually any source to Snowflake in a visual, automated and managed way.
Already available on AWS, this service eliminates technical dependencies and reduces integration times for new sources.
Snowflake now natively integrates the dbt (data build tool), widely used in data engineering.
This means that data engineering teams can create, version and execute transformations directly within the platform, leveraging modern features such as AI Copilot (AI code assistant) and version control via Git.
The business value is in the simplification: the entire process of extracting, transforming and loading (ETL/ELT) runs within the Snowflake ecosystem, with no external dependencies or fragmented processes, reducing errors, accelerating delivery and strengthening governance.
Snowflake has also improved its support for Apache Iceberg, a key format for lakehouse data architectures. It is now possible to work with external Iceberg REST-compliant catalogs, and read or write data to Iceberg tables directly from Snowflake. In addition, capabilities such as Merge on Read and support for semi-structured data (VARIANT) are added.
This means that Snowflake no longer just manages its own data, but can act as an intelligent hub that connects to external data lakes, facilitating open and scalable data strategies.
One of the common challenges for companies implementing AI is the lack of visibility into the behavior of models once in production.
Snowflake solves this with an observability suite specifically for generative AI, which allows you to monitor how AI agents operate, what data they use, what responses they generate and how often they are triggered.
This functionality not only increases transparency and confidence in models, but also improves the ability to detect errors, optimize costs and comply with regulatory requirements. In an environment where traceability is critical - for example, in industries such as banking or healthcare - this provides a clear competitive advantage.
With new capabilities in Snowflake Trail, technical teams can gain full visibility into what is happening in their environment: ingest with Openflow, use of AI agents, containers in Snowpark, app behavior, etc.
This allows them to identify errors, bottlenecks or cost variances before they impact the business, and ensure efficient and secure use of the entire platform.
Snowflake Summit 2025 delivered a wave of innovation for data-driven enterprises, centered around:
| Update | Summary Description | Expected Impact |
|---|---|---|
| Snowflake Intelligence | AI-enabled conversational agents to explore data in natural language. | Democratizes data analysis without writing code. |
| Data Science Agent | Co-pilot for data scientists with ML workflow automation. | Accelerates model development and improves data science productivity. |
| Openflow | Apache NiFi-based, managed, multi-modal data ingestion. | Eliminates silos and manual tasks in ETL. |
| Semantic Views | Semantic layer with unified metrics for AI and BI. | Ensures analytical consistency and reliable results. |
| Standard Warehouse Gen 2 | New generation of faster warehouses at no extra cost. | Improves performance up to 2.1× with no tuning required. |
| SnowConvert AI | Free tool that converts legacy code to Snowflake with AI. | Facilitates and accelerates migrations from legacy platforms. |
| Adaptive Compute | Adaptivewarehouses that intelligently scale and route queries. | Automates infrastructure and reduces human intervention. |
Snowflake Summit 2025 confirmed a bold vision: turning Snowflake into the core enterprise platform for AI, automation, and open data ecosystems.
From conversational agents to intelligent co-pilots, from performance breakthroughs to governance enhancements, every update brings Snowflake closer to becoming the intelligent hub for enterprise data and AI.
In short, Snowflake is becoming smarter, more open, and more powerful — a platform built not just to manage data, but to activate its full potential through generative AI, process automation, and seamless integration.
As these innovations move from preview to general availability, the focus will shift to how businesses operationalize them to create a competitive advantage.
As Snowflake says: “Data does more.” And in 2025, it certainly will.