e6data Lakehouse Compute Engine

e6data, Inc. is based in San Francisco and led by co-founder and CEO Vishnu Vasanth. It builds a compute engine for lakehouses and warehouses that runs SQL, streaming and AI workloads on open tables inside the customer's own cloud, aiming to reduce cost and lock-in from data intelligence vendors.

The engine is disaggregated and Kubernetes-native and scales in 1-vCPU increments. The Query Engine serves high-concurrency SQL and AI analytics on Iceberg, Delta and Hudi tables; the Ingest Engine streams data into Iceberg tables that become queryable in about 15 seconds; and AgentFlow, in preview, builds agents for analytics and operational automation. It works alongside Databricks, Snowflake, Trino and Microsoft Fabric and deploys on AWS, GCP, Azure, on-premise, hybrid or air-gapped environments. Nasdaq is a published customer.

• Query Engine: High-concurrency SQL and AI analytics on open lakehouse tables.
• Ingest Engine: Streaming ingestion into Iceberg tables, queryable in about 15 seconds.
• Granular scaling: Compute scales in 1-vCPU increments instead of fixed warehouse sizes.
• Open formats: Works with Iceberg, Delta and Hudi and any catalog.
• Platform interoperability: Runs alongside Databricks, Snowflake, Trino and Microsoft Fabric.
• Hybrid deployment: AWS, GCP, Azure, on-premise, air-gapped and sovereign environments.

A Vietnamese bank or securities firm can serve customer-facing portfolio and transaction queries from an Iceberg lakehouse at high concurrency. A company already on Databricks or Snowflake can offload heavy dashboard workloads to cut compute cost. A regulated enterprise can run analytics on-premise without moving data to a public cloud.

The vendor claims more than 1,000 queries per second with p95 latency under 2 seconds and up to 50% lower compute cost through 1-vCPU scaling. Running compute next to the data cuts egress by about 99% and avoids moving data. A Nasdaq case study reports 15 million customer-facing queries a day, 60-75% lower total cost of ownership and p95 latency of 1.2 seconds.

As data volumes and AI workloads grow, warehouse compute bills have become a major cost line, and companies are separating storage on open formats from interchangeable query engines. Data residency rules in Vietnam and the region also favor engines that run inside the customer's own cloud or data center.

Legal entity e6data, Inc., 355 Bryant Street, San Francisco; co-founder and CEO Vishnu Vasanth. Raised a $10 million Series A led by Accel with Beenext in September 2024. Holds SOC 2 Type II and ISO 27001 certifications.