Israeli Startup Keewano Emerges from Stealth with $12 Million Seed Funding
Israeli startup Keewano emerged from stealth with a $12 million Seed round led by Hetz Ventures and a16z Speedrun. The company developed a native database optimized for AI agents.
Israeli startup Keewano has emerged from stealth mode, announcing a $12 million Seed funding round led by Hetz Ventures, with participation from Andreessen Horowitz (a16z) via its Speedrun accelerator, Remagine Ventures, DIG Ventures, and private investors. Founded in 2024 by Mark Kardashov (CEO), Dima Kreger (COO), Vitaly Bukhovsky (CPO), and Pavel Bibergal (CTO), the company is tackling a fundamental bottleneck in modern enterprise artificial intelligence.
Rethinking the Data Stack for AI Agents
"Organizations today are trying to connect analytical agents to existing databases that were never designed for them," explains Mark Kardashov, co-founder and CEO of Keewano. Traditional data architectures require long and expensive integration projects, after which AI agents typically process only pre-filtered data subsets, often relying on historical information rather than real-time streams. Keewano was built from the ground up for the AI era, creating a native database and storage format optimized specifically for autonomous agents.
KeewanoDB organizes information by logical entities and actual chronological events rather than traditional modeled tables with pre-defined relational schemas. This design eliminates the need for complex indexing, allowing agents to query raw operational data dynamically. According to the company, the engine processes millions of simultaneous events in milliseconds without unnecessary network overhead, returning only compact results and relevant context.
Deployment and Expansion Plans
"Keewano was built from the ground up in the AI era to solve these problems at the data layer itself, instead of trying to retrofit an existing infrastructure to be agent-first," notes Kardashov.
Operating either as a standalone database or alongside an existing enterprise data stack, Keewano ingests application events directly via SDK or API. Streamlined integrations can be completed within hours or days, while broader deployments alongside existing data lakes or message buses function as an optimization layer taking several weeks.
Headquartered in Ramat Gan with a team of approximately 20 employees, Keewano plans to expand its operations to the United States this year. The founding team brings extensive industry experience, having previously built and exited companies such as TestProject and Devalore.
