Key takeaways
- Swiggy has significantly improved its data processing speeds by moving to Snowflake, achieving a reduction of up to 96% in workflow times.
- The company has established a central analytical layer using Apache Iceberg, integrating data from various business units.
- Near real-time insights have empowered teams across Swiggy to access crucial operational and business data independently.
- Marketing and product engineering teams can leverage the data infrastructure for targeted campaigns and feature assessments.
- Snowflake's platform ensures governed self-service access with robust security measures, enhancing data accessibility for all stakeholders.
Swiggy has moved its central analytical workloads to Snowflake, cutting the processing time of its slowest data workflows by 90–96%. The company also reduced the runtime of its heaviest individual queries from two hours to 15 minutes.
The foodtech company built a central analytical layer on Snowflake using Apache Iceberg, bringing data from its different businesses into a unified serving layer.
Snowflake said the broader data processing pipeline, which previously took about six hours, now delivers insights in near real-time. This allows teams across Swiggy to access operational and business data without depending as heavily on central data teams.
Marketing teams can use the data infrastructure to create and launch targeted campaigns through their own tools, while product engineering teams use standardised metrics to assess experimental features before release.
Swiggy’s operational teams can also monitor service-quality metrics in real-time, while finance teams get visibility into technology costs at the workload level.
The platform also provides governed self-service access through role-based permissions, column masking, and row-level security.
Snowflake said the same controls can be applied to AI agents and applications, including permission boundaries, short-lived credentials, and audit trails.
“At Swiggy, data is valuable only when it reaches the person who can act on it, whether that is a city sales manager, restaurant owner or delivery partner,” Madhusudhan Rao, CTO of Swiggy, said in a statement. “With Snowflake, we are making trusted, governed insights easier to access while ensuring that every user and AI agent operates within the same permissions and audit framework.”
The company said the unified infrastructure was designed to address the growing complexity of serving millions of customers across food delivery, Instamart and Dineout, where fragmented data had made it harder to support peak workloads and provide timely insights.
“India's fastest-moving companies can't afford to route every data question through a central team,” said Vijayant Rai, Managing Director of India at Snowflake. “With Snowflake, Swiggy has built a centralised, governed data foundation, consistent controls, and an AI framework that holds agents to the same standards as people.”
The Swiggy deployment comes as Snowflake is seeing increased demand for its AI-focused data platform.
In its latest quarterly results, Snowflake reported 37% year-on-year growth in product revenue to $1.49 billion and raised its full-year product revenue forecast to $6.07 billion. CEO Sridhar Ramaswamy said AI accounted for “approximately half of the acceleration” in the company’s growth.
