Snowflake's journey in the data space began in 2014 as a cloud-based data warehouse provider, to have now evolved into a multi-cloud data platform that aims to streamline how businesses manage, store, and process data. Speaking with AIM, Murad Wagh, the director of sales engineering at Snowflake, said that at the core of the company's strategy are the three pillars -- data foundation, AI/ML, and data applications.
"We started with a core data foundation that solved the infrastructure management problem that enterprises faced with on-prem solutions like Teradata and Hadoop. By offering a cloud-native, multi-cloud data platform, Snowflake allowed businesses to focus on data outcomes rather than infrastructure challenges," explained Wagh.
Over time, Snowflake introduced features like data collaboration, enabling companies to securely share data across organisations without having to extract it from the platform. This paved the way for AI and machine learning use cases.
Wagh emphasised that Snowflake's evolution was driven by customer needs. "We solved the data platform problem, but realised that people wanted to do more like collaborating with their data and build models on it, without moving it out of the platform's governance boundaries."
To facilitate AI workloads, Snowflake introduced Snowpark, a framework that enables data scientists and engineers to use popular programming languages like Python, Java, and Scala natively within the platform. This expanded the platform's capabilities beyond SQL and unlocked new possibilities for AI and machine learning.
"With Snowpark, you're not taking data out of Snowflake, making the process more efficient and keeping data within the governance framework," Wagh noted.
"It's far from dead," said Wagh when speaking about SaaS and how Snowflake continues to thrive, with over 10,000 customers and growing. The idea of AI as a Service is evolving, but SaaS remains very much alive. Snowflake's customers appreciate the focus on outcomes, not infrastructure, which allows them to execute large-scale projects with minimal teams in record time.
The platform's seamless upgrades and high availability, built on hyperscaler regions with multiple availability zones (AZs), ensure that customers experience no downtime -- a major advantage that contributes to its ongoing success.
This shift has allowed businesses to build, train, and deploy machine learning models directly on Snowflake's platform, eliminating the need for external processing environments. "We now offer native support for ML operations, including APIs, a model registry, and more, which allows companies to fully manage their AI workflows inside Snowflake," Wagh added.
"We're focused on enabling businesses to get more out of their data without having to worry about the underlying infrastructure," said Wagh. This is what differentiates Snowflake's offering from competitors like Databricks, and others, which the company sees as healthy competition.
One of Snowflake's key differentiators is that it's a fully SaaS-based service focused on business outcomes, whereas other platforms offer Platform-as-a-Service (PaaS) with more control knobs and infrastructure management.
Snowflake's simplicity and efficiency [customers don't have to worry about tuning or managing infrastructure] set it apart. The marketplace and collaboration tools, like data clean rooms and real-time data sharing, also provide unique value. For example, Razorpay uses Snowflake's data sync product to share transaction data in real-time with its customers, a feature that has resonated well with the market.
Large enterprises, such as financial institutions, may still benefit from maintaining a monolithic system, while other interfaces or outcomes are better suited for microservices. "At Snowflake, we don't necessarily have a stance on this. Instead, we provide support for customers regardless of their architecture choice," said Wagh.
Snowflake's Cortex is its suite of generative AI capabilities, including open-source models like Llama and frameworks like NVIDIA's NeMo. "With Cortex, we allow customers to bring models to Snowflake, eliminating the need to extract data to external LLMs. Everything -- security, hosting, upgrades -- is managed by Snowflake," Wagh explained.
It also offers several powerful AI functionalities, such as text summarisation, translation, sentiment analysis, and more. Developers can access these functions using SQL or Python, making it easy to build AI-driven applications. "We've created a simple interface where you can, for example, call a function to summarise large blocks of text with just a few lines of code," said Wagh.
When it comes to Snowflake Horizon, the platform for data governance, it's becoming an essential tool for customers. Horizon includes capabilities like data discoverability, classification, and tagging, which allow users to automatically classify and tag columns that may contain sensitive information, such as Personally Identifiable Information (PII).
Once this data is tagged, policies can be set up easily, such as masking specific data columns from developers while allowing analysts to view the data in clear text. Horizon also provides other governance features like role-based access control, row and column masking, encryption, and multi-factor authentication (MFA).
Another example is Cortex Analyst, a feature that enables natural language interaction with structured data, allowing non-technical users to query datasets without writing complex SQL. This capability aims to democratise data access within organisations. "It's breaking down barriers," Wagh explained. "Sales reps can now ask questions like, 'Show me my performance over the last quarter,' and get answers without needing an analyst to write SQL."
