Snowflake Massively Expands Types of Applications That Can Be Built, Deployed and Distributed on Snowflake

Apps are the way to democratize AI: to make it accessible to everyone and streamline customers' experiences with faster time to insights. According to a recent IDC survey, AI applications is currently the largest category of AI software, accounting for roughly one-half of the market’s overall revenue in 2023.

But to successfully build AI-enabled apps, you need secure access to a solid data foundation that encompasses a massive amount of high-quality data; has high scalability to handle ever-increasing data volumes; and integrates with other systems, apps, services and enterprise LLM models. You need a single platform that gives you the ability to build, deploy and operate your applications across clouds and regions, reliably and at scale. That’s where Snowflake comes in.

Companies of varying sizes and industries are building applications on Snowflake, from large financial institutions like Capital One Software, to consulting firms such as SNP Group and data integration platforms like RudderStack. There are now more than 160 Snowflake Native Apps on Snowflake Marketplace that you can install and use today.  

Let's take a closer look at how Snowflake enhances your AI application development journey.

The toolkit for building on Snowflake continues to expand. You can add robust capabilities to your applications with embedded analytics, AI, search, containerized workloads and more. Let’s explore the latest developments designed to expand the range of applications you can create on Snowflake.

Snowpark Container Services, soon to be generally available on AWS and in public preview on Azure, empowers app providers to efficiently build and operate sophisticated generative AI apps. It’s already making an impact for Eutelsat OneWeb, which provides global broadband satellite connectivity services and is constantly transmitting vast amounts of data for enhanced web experiences. The Eutelsat system processes 70 billion rows of data daily, in near real time from low Earth orbit, reaching servers in less than 30 minutes. With data moving at that extreme pace, it is critical to monitor performance and troubleshoot in a timely manner to ensure an undisrupted customer experience. 

To quickly address network disruptions, Eutelsat built an advanced real-time processing system on Snowpark Container Services with multiple container orchestration. “By incorporating Streamlit for real-time frontend monitoring and using Snowpark Container Services as the backend, we ensure minimal downtime and maintain high-quality internet connections for our customers,” says Miguel Morgado, Senior Product Owner for Data and Performance Hubs at Eutelsat Group. “Our Snowflake LEO Operations Support System (OSS) allows troubleshooting within a two-minute window and shifts from a reactive to a proactive operational model using AI and predictive analytics.” 

With containers running in Snowflake, there is no need to move governed data outside of Snowflake (thereby exposing it to additional security risks) in order to use it as part of the most sophisticated AI/ML models and apps. For example, by utilizing Snowpark Container Services with block storage, users can securely deploy high-performance large language models (LLMs) and other low-latency applications. Developers can also store large data sets, preloaded models, code and other files next to their containers in a reliable, persistent way. This accelerates response times and supports the execution of both proprietary and third-party services with sub-second response times, delivering an optimal user experience. 

Snowpark Container Services also works with Hybrid Tables (public preview), which offers double-digit millisecond, single-row operations, enabling use cases like storing app state, storing workflow state and serving data — all within Snowflake. To learn more about Snowpark Container Services, check out this docs page.

Search tools allow users to quickly find information or items in an app, providing great useability and efficiency. Sygnia, a renowned cybersecurity consulting and incident response company, turned to Snowflake when it needed a platform that could support high data velocity with near-unlimited data retention and give teams the fast search functionality they needed to do their jobs. Sygnia tested Snowflake Full Text Search (public preview soon) for their Managed Detection and Response (MDR) product and saw the benefits firsthand. 

“Migrating from Elasticsearch to Snowflake has allowed us to scale and support a rapidly growing managed XDR client base, and opens up new innovations in our product roadmap,” says Avri Dahan, Director of R&D at Sygnia. “The combination of Sygnia’s vast security data lake and Snowflake’s Full Text Search function with Search Optimization is yielding unrivaled efficiency in cyber incident resolution. We're excited to continue to innovate with Snowflake.” 

Snowflake’s Full Text Search feature gives developers a new token-based search function to use for log analytics and other high-volume data search applications. It enables end users to swiftly locate precise answers across large volumes of structured and semi-structured data with exact matching. Full Text Search is particularly valuable for entity searches, such as IP addresses, URLs, emails and file paths.

We're also expanding to conversational search capabilities with Snowflake Cortex Search (public preview soon) and Cortex Analyst (private preview). Cortex Analyst empowers app users to engage with structured data using natural language. Cortex Search is a serverless API designed to process questions and return the most relevant answers from a defined set of documents, such as PDFs, text files (.txt) and CSV files (.csv), utilizing a hybrid search method that combines vector and keyword searches. 

Creating next-generation full-stack and gen AI/LLM apps is just the first step. Next, you need to securely distribute and deploy them to customer accounts — and have the opportunity to market and monetize your offering to thousands of customers in the AI Data Cloud. 

On the deployment front, we are excited to announce that the Snowflake Native App Framework is now generally available on Google Cloud Platform (GCP), making it available across AWS, Azure and GCP. Providers can build their app once and publish it to customers across all three major clouds and multiple regions with a single listing, removing the operational burden of keeping your app updated in various clouds. 

Additionally, we’ve released Snowflake Native App Framework integration with Snowpark Container Services in public preview on AWS, enabling providers to build and deliver more sophisticated apps faster. You can now deploy on top of configurable GPU and CPU instances, significantly enhancing the provider's development time-to-value. 

More than 20 Snowflake Native Apps that leverage Snowpark Container Services are already available in Snowflake Marketplace, demonstrating the power and flexibility of this integration. These apps showcase a wide variety of use cases, including computer vision automation, geospatial data analysis, graph databases, ML applications for enterprises and much more. Anyone on AWS and Azure can discover the apps in Snowflake Marketplace and start using them today.

Providers can streamline the purchasing process and accelerate deal cycles with flexible on-platform monetization options, including subscription-based and usage-based pricing models. With the new Compute Pool Surcharge Monetization Model (private preview), providers can now bill consumers based on the compute resources the apps use, allowing better alignment between pricing and value.

Convenient, on-platform payment options also help consumers accelerate the procurement process and achieve faster insights. They can submit payment via credit card, ACH or wire, or simply use their committed Capacity through the Snowflake Marketplace Capacity Drawdown Program, which is now expanding internationally with private preview in the UK (only available to organizations located in select jurisdictions. See documentation for additional program criteria). 

“By offering our AI product on Snowflake, we have already seen a 30% increase in the pipeline within several weeks of the private preview of our Snowflake Native App,” says Hema Raghavan, Co-Founder and Vice President of Engineering at Kumo AI. “The ability for customers to use their committed Snowflake spend to pay for Kumo AI’s Snowflake Native App has also helped our team close deals faster.”