Introducing e6data: The Kubernetes-Native Data Compute Engine That Promises Significant Cost Savings

Even with advanced tools from data warehouse providers like Snowflake and Databricks, enterprises often face challenges with mission-critical workloads.

San Francisco-based startup e6data claims to offer a solution.

Introducing e6data’s Innovative Compute Engine

e6data has recently secured $10 million in funding from Accel and other investors to develop a "reimagined" Kubernetes-native compute engine. This engine integrates into any mainstream data intelligence platform, enabling businesses to manage compute-intensive workloads with up to 5 times better performance and a 50% reduction in total cost of ownership (TCO) compared to traditional compute engines.

Despite being newer than established options like Spark, Trino/Presto, and Starburst, e6data has already attracted significant interest from major industry players, including Freshworks, who are exploring the potential cost-performance advantages.

Solving Performance Bottlenecks

Most modern data platforms—such as Snowflake, Databricks, Google BigQuery, and Amazon Redshift—rely on compute engines to process substantial data workloads. These engines, crucial for executing data transformation, analysis, and modeling, excel at traditional tasks like analytical reporting. However, they struggle with emerging use cases such as real-time analytics (e.g., fraud detection) and generative AI, which demand faster processing and higher capacity.

Vishnu Vasanth, founder and CEO of e6data, notes that spending on such workloads is surging, with increases of 100-200% annually for many large enterprises. The root cause of these performance bottlenecks lies in the outdated architecture of many compute engines. Most are built around a central coordinator that can become a single point of failure under heavy load, leading to inefficiency.

Vasanth describes this architecture as having a "central puppet master" overseeing the workload, making it prone to bottlenecks.

A New Approach to Compute Architecture

To tackle this issue, e6data's team applied their expertise in both commercial and open-source data projects to revamp the compute engine architecture. They introduced decentralized components that can scale independently, utilizing a Kubernetes-native distributed processing model that eliminates the need for centralized task scheduling.

"This innovation breaks away from the traditional top-down control structure, allowing components to operate autonomously and coordinate in a flexible, bottom-up manner, akin to a flock of starlings," Vasanth explained.

Benefits of e6data's Compute Engine

e6data claims that its compute engine delivers up to 5 times better query performance and 50% lower TCO compared to conventional engines. However, these results have been sourced from early customers like Freshworks and Chargebee, with third-party benchmarks expected soon.

Additionally, e6data differentiates itself by minimizing vendor lock-in. Traditional architectures often push customers toward a specific data ecosystem, complicating data management. In contrast, e6data integrates seamlessly into existing platforms, supporting popular open table formats (Hive, Delta, Iceberg, Hudi), data catalogs, and various SQL dialects.

"Our architecture doesn’t require data relocation, application changes, or downtime, allowing customers to implement our solution within two days," Vasanth states.

As e6data continues to innovate, it faces the potential to capture significant market interest. The global market for data and AI solutions is projected to reach $230 billion by 2025, with 60% of CXOs planning to increase their budgets in the coming year.