Snowflake Summit 2025 focuses on simplicity and AI-readiness

Snowflake Summit 2025, the largest in its history, centered on the theme "simplicity drives results." This message core was woven into every keynote and announcement, emphasizing the data cloud’s focus on simplifying the ability to derive value from data.

The undeniable influence of AI and the resulting requirement for organizations to architect their data for AI, or otherwise become “AI ready,” was equally central. There is a clear appetite for this among Snowflake customers, with 72% of those the data cloud surveyed stating they intend to adopt agentic/autonomous AI in 2025. 

Feature announcements consistently highlighted how AI is reshaping the way we build, test and use data products. Examples span from conversational interfaces that provide custom insights to autonomous AI agents that automate functions. 

The summit presented a clear vision of a future where data management is more intuitive and deeply integrated with AI and automation. This vision aligns with our own vision at Capital One Software. We built Capital One Slingshot to simplify data cloud management and Databolt to enhance AI readiness by enabling organizations to run AI on all of their data in a secure and performant manner. 

Read on for a full recap of Snowflake Summit 2025!

Snowflake’s emphasis on simplification extends far beyond usability. It represents a strategic move to address the inherent complexities of modern data ecosystems. By reducing friction in adoption and accelerating time-to-value, the data cloud aims to alleviate key pain points for enterprise data engineers and platform leaders. 

Openflow is a new tool that makes it easy to get structured and unstructured data into Snowflake, from a multitude of popular sources. Snowconvert AI is a new feature added to Snowconvert to help map and migrate data to the platform. These are just a couple examples of native products the platform is building to make it easier to get data in.

Another major highlight was the unveiling of Cortex AISQL, which brings conversational language to SQL queries, democratizing data insights across the organization. This signifies a significant shift from traditional SQL-based data access towards a future where business users can directly query and extract insights from data using natural language. 

For data engineers, this means a greater focus on data quality, robust semantic layers and stringent data governance. The ultimate goal is to improve the accuracy, reliability, security and federation of LLM insights, as conversational interfaces become the main access point to data for non-technical users. 

The main features and capabilities announced at Snowflake Summit 2025.

Snowflake unveiled a couple enhancements to cost visibility, among the many new features announced. These enhancements include:

1. Cross-account cost views 

2. Real-time notifications for spend anomalies

3. Tag-based budgets for showbacks and chargebacks 

Slingshot enhances these new capabilities by providing a single pane of glass for data clouds, across Snowflake and Databricks. The ability to see the organization’s total spend across teams and platforms is a game changer for organizations that have scaled their operations across data clouds. 

In addition, Slingshot has built tags to enhance the capabilities of Snowflake’s native tags. Slingshot tags let users track costs and set access privileges to Snowflake objects. 

Implement zero trust on Snowflake resources by assigning least-privileges to users for Snowflake assets within Slingshot. Slingshot’s custom tags can be leveraged to meet the needs of any org structure or custom federation mode, adding flexibility and guardrails to what Snowflake offers natively. 

Adaptive warehouses were announced during the summit as a new type of compute. With adaptive warehouses, currently in private preview, users can enjoy the advantages of dynamic query routing without building custom infrastructure for compute optimization. 

Adaptive warehouses automatically route each query to an appropriately sized compute based on the query’s individual needs. As such, they represent a significant leap towards self-optimizing compute. While this promises to simplify operations for end-users, it also introduces a new challenge for platform leaders: how to effectively validate the efficiency and ROI of an adaptive system that largely operates autonomously.

Slingshot addresses this directly by providing detailed value reports that calculate the ROI seen from Slingshot optimizations. For adaptive warehouses, we make it easy for organizations to see if this new type of compute meets their needs. Users can easily set up adaptive warehouses, or migrate standard warehouses to adaptive warehouses, within Slingshot and see the effect on costs and performance. By comparing before and after metrics, data engineers and platform leaders can validate where and when to use adaptive warehouses vs Gen2 vs Gen1 standard warehouses.

Here’s a breakdown of how Slingshot enhances Snowflake's new cost and compute features:

While Snowflake is investing in its native cost visibility capabilities, Slingshot goes deeper with ML-powered forecasts, cross-platform visibility and proactive alerting. These capabilities transform cost optimization from a reactive exercise into a strategic initiative with predictable outcomes. 

For platform engineering leaders, this translates into fewer unexpected budget surprises, significantly faster anomaly detection, and the crucial ability to course-correct spending patterns before costs escalate out of control.

Snowflake unveiled new tools that strengthen the data foundation by making it easier to ingest and govern data. Snowflake Horizon is a data governance solution that includes access controls for AI, automatic threat detection, automatic IP blocking, a copilot (in private preview) and more. Crucially, Horizon can now discover assets outside of Snowflake and integrate with popular tools like Airflow and dbt. 

Openflow is another newly announced feature, which simplifies data ingestion and adds tracking and observability. The combination of the two suggests that Snowflake has  ambition to evolve into a central "platform of platforms" for the entire data ecosystem. The data cloud is actively positioning itself as an orchestrator and unified environment for data ingestion, transformation, development and governance.

The developer experience received a major boost with dbt projects in Snowflake (in private preview). This feature enables users to build, test, and deploy dbt pipelines directly within Snowflake. It is complemented by Workspaces, a new file-based development environment for authoring and managing code artifacts, including dbt projects. 

The recent acquisition of Crunchy Data and the announcement of Snowflake Postgres indicate the platform aims to provide a fully managed Postgres service. Data engineers are now able to run Postgres-powered applications directly on Snowflake. In addition, Snowflake Postgres aims to simplify data management and accelerate AI development by offering familiar tools with enterprise-level security and governance within the Snowflake.

Features like these clearly represent Snowflake's push towards making data engineering practices more accessible and integrated within the platform. 

Snowflake announced new features that provide faster insights and shorter time to value using AI. Snowconvert, a tool to facilitate data migration, was showcased along with its new ability to test, convert and assess code for migration using Snowconvert AI. 

The recently launched Gen2 warehouses was also featured. These warehouses use faster hardware and represent the next step in the evolution of the standard virtual warehouse. With this option, users retain control over compute and can still size their warehouses. This is in contrast to  adaptive warehouses that abstract away the complexity of infrastructure management for simplicity. Overall, Gen2 warehouses offer performance boosts for current workloads, while adaptive represent the future of automated compute. 

Lastly, Snowflake’s AI capabilities have been expanded with Cortex AISQL, which enables natural language queries in SQL with support for multi-model data spanning text, images and audio files. Cortex agents are designed for orchestrations and instructions, enabling the creation of powerful agentic applications that provide data insights. Add Snowflake Intelligence to the mix for permission-aware access to agents, allowing users to ask questions in natural language and receive answers based on their access, powered by semantic views.

The combination of Cortex AISQL, Cortex Agents and Snowflake Intelligence strongly suggests Snowflake has put a strategic focus on enabling end-to-end AI application development and deployment directly within its platform. 

Snowflake Summit 2025 underscores the necessity of building robust, AI-ready data foundations. This was highlighted by the fact that 72% of Snowflake customers anticipate implementing autonomous AI within the current year. 

The pervasive and central focus on AI throughout the summit signifies that the role of data engineers is expanding dramatically to become critical enablers of enterprise-wide AI initiatives. This means shifting priorities and focus from traditional ETL processes to AI-ready data architecture and the implementation governance and guardrails for AI.

With Snowflake increasingly abstracting compute choices while simultaneously providing more granular native cost visibility, the core challenge for data platform leaders is shifting from simply how to optimize to how to intelligently manage and validate automated optimization. 

Slingshot can help you do just that. It provides enterprises with essential tools for intelligent optimization and simplifies data cloud management, enabling users to automate and validate optimization efforts, see platform use and spend and govern access to their data in one place. 

The overarching narrative of the 2025 Snowflake Summit, powerfully reinforced by Slingshot's enhancements, points towards a future where data management is seamlessly integrated across the entire data lifecycle. This integration is increasingly driven and optimized by AI and intelligent automation, drastically reducing manual effort and boosting overall efficiency. 

It is crucial for data and platform leaders to embrace solutions that offer this holistic, intelligent and integrated approach to ensure their data infrastructure is not only scalable but also cost-effective. The strategic partnership between Capital One Software and Snowflake provides enterprises with the essential tools and insights needed to confidently navigate this initiative.

Interested in learning more about Slingshot? Book some time with the team to chat about how Slingshot can help you maximize your investment in Snowflake.