At the Snowday 2023 event, Snowflake, the Data Cloud company, launched an extensive series of innovations centered around the advancement of customer success. Focused particularly on simplifying the data foundation, accelerating AI success, and scaling with applications, Snowflake’s latest releases intend to derive value from AI, whose robust data strategy permeates throughout the rest of the application development process.
Snowflake’s recent developments boil down to a central theme, as Christian Kleinerman, SVP of product at Snowflake, put it, “we do not believe that organizations will be able to have a successful AI strategy and a successful application of AI without a strong data foundation.”
This sentiment manifests as enhancements to Snowflake’s Data Cloud platform that offer new avenues toward eliminating data silos and bringing AI and app development directly to enterprise data.
Iceberg Tables (in public preview soon) takes its place among these improvements as a single table type that drives the streamlined management and great performance of Snowflake for data stored externally in the open standard Apache Iceberg format—without requiring upfront ingest cost. This innovation supports data architectures that require interoperability while retaining Snowflake’s hallmark traits—ease of management and high performance, according to the company.
In conjunction with the announcement of Iceberg Tables, Snowflake Horizon is a reimagined, built-in governance solution with a unified set of compliance, security, privacy, interoperability, and access capabilities that simplifies app governance. These capabilities include:
- Additional authorizations and certifications, such as U.K.’s Cyber Essentials Plus (CE+), FBI’s Criminal Justice Information Services (CJIS) Security Policy, IRS’s Publication 1075 Tax Information Security Guidelines, and more
- Data quality monitoring
- Data lineage UI
- Differential privacy policies
- Enhanced data classification
- Trust center, an interface for discovering security risks with recommendations to resolve them
“This [Snowflake Horizon] is not a new capability. It is bringing together and organizing what we've had as industry leading capabilities…but we're simplifying how we structure these efforts and introducing the next generation of technologies,” said Kleinerman. “We're incredibly excited about Snowflake Horizon as the overarching technology that helps our customers gather and understand the data within Snowflake.”
Snowflake is further introducing a new Cost Management Interface, which increases visibility in Snowflake spend. This centralized interface enables Snowflake users to comprehend, control, and optimize their spend with various out-of-the-box capabilities.
Regarding the company’s leap into greater support for AI app development, Snowflake is announcing enhanced Python capabilities for Snowpark to supercharge productivity, collaboration, and acceleration of end-to-end AI and ML workflows. These innovations include:
- Snowflake Notebooks for Python and SQL users to explore, process, and experiment with data in Snowpark within an interactive, cell-based programming environment
- Snowpark ML Modeling API for enabling developers and data scientists to scale out feature engineering and streamline model training
- Snowpark Model Registry for empowering scalable, secure deployment and management of models in Snowflake
- Snowflake Feature Store for creating, storing, managing, and serving ML features for model training and inference
Additionally, this advancement of AI app development yields other capabilities, including the Snowflake Native App Framework for developing apps simpler with Snowflake’s building blocks, Snowpark Container Services for running any component of an app without needing to move data or manage complex, container-based infrastructure, and Database Change Management for coding declaratively and templatizing work easier.
Finally, Snowflake is announcing Snowflake Cortex, a fully managed service that empowers enterprises to discover, analyze, and build AI apps in the Data Cloud—with a particular emphasis on securely deriving value from generative AI (GenAI) more easily.
Snowflake Cortex offers serverless SQL/Python functions that run inference on conversational LLMs and execute vector search functionality, enabling rapid construction of contextually enriched applications. The solution affords users instant access to a growing set of serverless functions—including LLMs such as Meta AI’s Llama 2 model—so that organizations can produce LLM and AI apps without needing extensive AI expertise or complex GPU-based infrastructure management.
“Snowflake is helping pioneer the next wave of AI innovation by providing enterprises with the data foundation and cutting-edge AI building blocks they need to create powerful AI and machine learning apps while keeping their data safe and governed,” said Ramaswamy. “With Snowflake Cortex, businesses can now tap into the power of large language models in seconds, build custom LLM-powered apps within minutes, and maintain flexibility and control over their data—while reimagining how all users tap into generative AI to deliver business value.
Furthermore, Snowflake is introducing a myriad of new LLM experience to the Data Cloud, including:
- Snowflake Copilot, an LLM-powered assistant that brings GenAI to everyday Snowflake coding tasks
- Universal Search, an LLM-powered search capability that searches across Snowflake accounts and Snowflake Native Apps
- Document AI, an LLM-powered content extraction function that increases efficiency in document comprehension
To learn more about Snowflake’s latest enhancements, please visit https://www.snowflake.com/en/.