DATABASES Signal 259
How Physical Intelligence unified its robotics data stack with Postgres managed by ClickHouse
↗
Physical Intelligence consolidated robotics data workloads onto ClickHouse-managed Postgres for unified OLTP and OLAP processing.
Engineers building robotics foundation models now see a viable path to handle both transactional and analytical workloads on a single stack without sharding or separate clusters. The shift reduces operational overhead but requires retooling pipelines to fit ClickHouse’s columnar storage model. If the pattern spreads, it could lower the barrier for startups entering high-cardinality robotics data domains.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
A single ClickHouse Cloud instance now serves both OLTP and OLAP for robotics foundation models, replacing a previous RDS Postgres bottleneck.
Data volumes grew from tens to hundreds of billions of rows after migration, indicating the stack scales for high-cardinality sensor data.
The unified stack simplifies exploration but demands schema and query redesign to align with ClickHouse’s columnar strengths.
THE CLUSTER