Why We Invested in EMBRASURE
Agents are becoming the primary users of the data warehouse. Agents now query data platforms 29x more often than humans do through a UI (MotherDuck, 2026). The warehouses most enterprises run today were built for analysts clicking through dashboards, not for agents querying around the clock. We believe the plumbing underneath agent workloads will be rebuilt from first principles, and that is a category creation moment in data infrastructure.
Embrasure is building the fully autonomous data warehouse for AI-native companies. Its proprietary compute engine and storage format sit on open Iceberg tables and are designed for how agents actually query: high volume, nonstop, with each question building on the last. For enterprises, Embrasure runs alongside Snowflake and Databricks with no migration required. On top of the engine, AI data engineers fix broken pipelines before anyone notices, and AI data scientists surface insights on their own.
Rockstar team as well: co-founders Jacob Kieser (CEO) and Victor Sunderland (CTO), who have built data platforms at OpenAI, Uber and Palantir.
