Keewano Raises $12M for Agent-Native Database
Keewano has raised $12 million to build a database designed around the event histories and contextual memory required by autonomous AI agents.
AI agents are creating new infrastructure requirements below the model layer. Keewano has raised $12 million for a database designed specifically around the way autonomous systems need to remember and reason about events.
What happened
The seed round was led by Hetz Ventures, with participation from a16z Speedrun, Remagine Ventures, DIG Ventures and angel investors.
Its product, KeewanoDB, stores sequences of actions and contextual events rather than organising information primarily around conventional tables and records.
The idea is to let agents inspect historical activity directly when deciding what to do next.
Why it matters
Agents often operate over long-running workflows: they open files, call APIs, change system states and respond to previous decisions.
Traditional databases can store those events, but developers still need significant logic to reconstruct context. An agent-native database could make persistent memory and event reasoning easier to build into applications.
The bigger picture
As agentic software matures, the surrounding stack is beginning to fragment into specialised tools for memory, orchestration, evaluation and permissions.
Keewano is betting databases themselves will evolve for that world. The opportunity is early and architectural choices are still unsettled, but the funding shows investors expect AI agents to create entirely new infrastructure categories rather than simply sit on top of the existing software stack.