The Pulse
Keewano Raises $12M and Launches KeewanoDB for AI Agents
Keewano has raised $12 million and launched KeewanoDB, a database designed to preserve complete event histories for AI agents. The Tel Aviv startup says the system can query a quarter-billion events in under half a second.
AI.info Team ·
250 Million Events in Under Half a Second
Keewano says its new database can query a quarter of a billion events in under half a second, a performance claim the Tel Aviv startup places at the center of its bid to build infrastructure for AI agents.
The company announced KeewanoDB’s general availability on September 15, 2026, alongside $12 million in seed funding. Hetz Ventures led the round, with participation from a16z speedrun, Remagine Ventures, DIG Ventures and angel investors.
KeewanoDB is designed to keep the complete event history of a user, device or AI agent together, in sequence and with its surrounding context. The company argues that conventional analytics systems often split those events into rows, columns and precomputed aggregates, forcing an agent to reconstruct what happened before it can reason about why an outcome occurred.
Keewano Stores the Sequence, Not Just the Result
Most business databases answer questions that people already know how to formulate: how many users churned, which products sold and when a metric changed. Questions about behavior over time are harder because the relevant evidence may be distributed across many event types and systems.
Keewano’s approach is to preserve each entity’s event sequence as a unified history. An agent can then inspect what happened before and after a particular action, compare behavioral paths and search for patterns that were not defined in advance through a dashboard or reporting schema.
“We were trying to use AI agents to answer business questions like ‘which of our users are on the same path as the ones who churned last month?’,” said Mark Kardashov, Keewano’s co-founder and chief executive. “You could see what happened on any dashboard, but why it happened lived somewhere the databases couldn't reach.”
Kardashov said the company concluded that supporting those questions required a new storage format rather than another layer on top of an existing analytics database. Keewano says its system can process trillions of events in parallel and return compact, context-ready results for agents.
A Database for Questions Dashboards Do Not Ask
Keewano presents KeewanoDB as a way to investigate why a customer churned, which sequence of actions preceded an outcome, or what separates one user group from another. Those tasks depend on event order and context, not only on aggregate counts.
The company also says the database can help agents surface emerging risks by examining complete behavioral histories. Its pitch is aimed at organizations that currently discard surrounding data because capturing more event types can increase instrumentation, storage and query costs.
KeewanoDB does not use per-event pricing, according to the company. Keewano says that model is intended to let customers retain and query more events without making every additional event a separate cost decision.
The fully managed Keewano Cloud service can run alongside an existing data warehouse or replace it, with standard integrations for organizations that want agents to work from the same underlying activity data. Customers can connect their own AI agents or use agents provided by Keewano.
Funding Comes From Gaming-Scale Analytics
Keewano was founded in 2024 by Kardashov, Pavel Bibergal, Vitaly Bukhovsky and Dima Karger. The founders bring experience from gaming, business software and infrastructure, areas where a single user can produce long and highly detailed event histories.
Kardashov and Bukhovsky previously built TestProject, which Tricentis acquired in 2019, and Devalore, which Abra acquired in 2022. Bibergal previously served as CTO at Plarium, while Karger led one of the company’s large studios, according to Keewano’s announcement.
Hetz Ventures managing partner Judah Taub said the company is addressing a data problem exposed by agent use rather than adding another feature to a traditional analytics stack.
“Every wave of AI adoption exposes the layer underneath it that wasn't built for the new workload,” Taub said. “We think machine reasoning is exposing exactly that gap in analytics infrastructure, and Keewano is addressing it directly rather than patching around it.”
The Claim Keewano Must Now Prove
Keewano’s launch rests on a specific technical bet: AI agents need raw, ordered event histories more than they need another collection of flattened tables and prepared aggregates. The company says that architecture produces faster queries, more complete context and lower costs for agent-driven analysis.
Those claims will matter most in production systems where event volumes, data quality and query complexity vary widely. The launch announcement provides Keewano’s benchmark for 250 million events, but does not describe an independent test or a direct comparison under identical conditions with competing databases.
For now, Keewano is selling a database built around the premise that an agent should be able to inspect what happened without first asking another system to reconstruct it. The company’s immediate test is whether customers find that preserved sequence valuable enough to change how they store operational data.