Data Infrastructure · sub-niche
Time-series databases for ML.
Time-series databases optimized for ML feature workloads, not just monitoring.
Team-sized buildTrickle — one deal per quarter
Why now
ML workloads need time-series features differently than monitoring. The slot between Influx and Snowflake is open.
What the signal looks like
Repos with Parquet / Arrow support, ML-aware aggregations, and Python / TS SDKs.
Public examples
We name publicprojects + categories only — never founders we track inside the paid product. The buyer’s edge stays inside the product.
- TimescaleDB shape
- QuestDB
- ClickHouse for time series
What this displaces
A Postgres table with a timestamp index + slow queries.
Our build-vs-invest call
Hard build. Fund only with prior database team. The wedge is a specific ML workload (financial, IoT, observability).
Common questions about this niche
- Buyer?
- ML platform teams.
- Pricing?
- Hosted DB markup.
- Moat?
- Performance + ML-specific features + ecosystem.
More inside Data Infrastructure
- Vector database engines — Vector search engines optimized for specific workloads — high-dimensional, hybrid, or local.
- Real-time feature stores — Feature stores with sub-second freshness for online ML.
- Postgres extension marketplaces — Postgres is now the AI database. The extension ecosystem is the next platform.
- Columnar warehouse alternatives — Snowflake / BigQuery alternatives optimized for a specific shape — cheap, fast, or open.