Keewano raises $12M seed and launches KeewanoDB for AI agent reasoning

Tel Aviv startup Keewano announced general availability of KeewanoDB, which it calls the first database built for machine reasoning at scale, and revealed $12 million in total seed funding led by Hetz Ventures with participation from a16z speedrun, Remagine Ventures, DIG Ventures, and angels. The company says the product keeps each entity’s full event sequence together and in order so AI agents can query behavior paths without rebuilding context from flat tables.

Same-day company primary that a new analytics layer is shipping for AI agents that need full behavior context — not another chatbot wrapper. A $12M seed with Hetz and a16z speedrun behind a purpose-built event database is a concrete FINANCE/SOFTWARE signal that agent workflows are forcing a rethink of how companies store history.

Keewano, a Tel Aviv startup, announced via PR Newswire on 15 Sep 2026 the general availability of KeewanoDB and revealed total seed funding of $12 million. Seed is early venture funding — a first institutional check, not a later growth round. The release is datelined TEL AVIV, Israel, Sept. 15, 2026, and timed 15:00 IDT — Israel Daylight Time, three hours ahead of UTC. That company PRIMARY via PR Newswire is the filing event. These are company claims. This desk did not see the term sheet.

The round was led by Hetz Ventures, the company said, with participation from a16z speedrun, Remagine Ventures, DIG Ventures, and angel investors. File the $12 million total, the Hetz lead, and those named participants as Keewano’s. This desk is not inventing a valuation, an ownership percentage, or a check size. None was printed beyond the $12 million total seed.

What the product is meant to do, as Keewano tells it: KeewanoDB stores information about every action a user, device, or AI agent undertakes — live, complete, in order, and with context — so analysis can run inside the database and return compact results agents can reason over. Machine reasoning here means AI agents answering “why did this happen” questions from behavior history, not just showing a dashboard total. File that product picture as the company’s. This desk did not run the database. The company’s “first database built for machine reasoning at scale” line stays a company claim.

Architecture, still the company’s: KeewanoDB keeps each entity’s complete event sequence together and in order so an agent can read answers from the raw data rather than reconstructing order at query time. An event sequence is an ordered log of what a user, device, or agent did over time. Traditional analytics stores, the company says, flatten events into rows, columns, and precomputed totals built for human dashboards. The company claims querying a quarter of a billion events in under half a second. File that speed line as Keewano’s. This desk did not time a query and is not independently verifying the benchmark.

Pricing and deployment, company: KeewanoDB does not use per-event pricing. Fully managed Keewano Cloud can connect alongside an existing data warehouse or replace it, using standard integrations. A data warehouse is a traditional analytics store built for human dashboards and predefined reports. Organizations can plug in their own AI agents or use Keewano’s. File those terms as the company’s. This desk did not see a price list.

Founders, as the same release has it: Mark Kardashov, chief executive; Dima Karger; Pavel Bibergal; and Vitaly Bukhovsky — co-founded in 2024. Kardashov and Bukhovsky’s prior exits, the company says: TestProject, acquired by Tricentis in 2019, and Devalore, acquired by Abra in 2022. Bibergal was CTO at Plarium; Karger led one of its largest studios. File the names, the 2024 founding, and those prior-exit lines as Keewano’s. This desk did not audit the exits.

Named voice on the release: Judah Taub, Managing Partner at Hetz Ventures, who led the round citing an analytics-infrastructure gap for machine reasoning. File the name and title as the company’s. His quotes are color only and stay in Sources. This is not investment advice.

Plain English for the rest of the card: seed = early venture funding. Machine reasoning here = AI agents answering “why did this happen” questions from behavior history. Event sequence = ordered log of what a user, device, or agent did over time. Data warehouse = traditional analytics store built for human dashboards and predefined reports.

PRIMARY here: Keewano’s 15 Sep 2026 PR Newswire company release — Tier A company source, the original record. The company site is product-home context, not a second originating newsroom. The $12 million total seed, the Hetz lead, a16z speedrun / Remagine Ventures / DIG Ventures / angels, the KeewanoDB general-availability announce, the event-sequence architecture, the quarter-billion-events-in-under-half-a-second claim, the no-per-event-pricing / Keewano Cloud pitch, the 2024 founders and prior-exit lines, and the Taub name are company-attributed. The company’s “first” claim stays company-attributed. NOT claimed: a valuation, ARR, customer logos, independently verified query speed, that this desk tested KeewanoDB, a stock tip, or investment advice. Distinct from the already-filed cornelis-205m-active-compute-fabric, euclyd-200m-series-a, and synapse-analytics-13m-series-a.