Keewano
KeewanoDB is an event-oriented database designed for AI agents and machine reasoning at scale.
Website: https://www.keewano.com/
Cover Block
From the public record
| Field | Value |
|---|---|
| Name | Keewano |
| Tagline | KeewanoDB is an event-oriented database designed for AI agents and machine reasoning at scale. [Keewano, retrieved 2026] |
| Headquarters | Tel Aviv, Israel [PR Newswire, September 2026] |
| Founded | 2024 [PR Newswire, September 2026] |
| Stage | Seed [PR Newswire, September 2026] |
| Business Model | API / Developer Platform |
| Industry | Deeptech |
| Technology | AI / Machine Learning |
| Geography | Middle East / North Africa |
| Growth Profile | Venture Scale |
| Founding Team | Mark Kardashov, Dima Karger, Pavel Bibergal, Vitaly Bukhovsky [PR Newswire, September 2026] |
| Funding Label | Seed |
| Total Disclosed Funding | ~$12,000,000 [PR Newswire, September 2026] |
Links
From the public record
- Website: https://www.keewano.com/
The Short Version
PUBLIC Keewano is a Tel Aviv-based data infrastructure startup building KeewanoDB, an event-oriented database aimed at AI agents and machine reasoning, and it warrants attention now because it coupled product launch with a $12 million seed round led by Hetz Ventures in September 2026 [PR Newswire, September 2026] [Calcalist Tech, September 2026]. The company’s pitch is straightforward: conventional tabular databases were built for humans asking predefined questions, while agentic systems need fast access to long, ordered behavioral histories, a gap Keewano says its architecture is designed to address [SiliconANGLE, September 2026] [Calcalist Tech, September 2026] [Hetz Ventures].
The founding story is still early, but the public record is coherent on the core points. Keewano was founded in 2024 by Mark Kardashov, Dima Karger, Pavel Bibergal, and Vitaly Bukhovsky, and coverage ties the group to prior work in gaming-scale analytics as well as earlier companies including TestProject and Devalore, both reported exits that give the team more operating credibility than a first-time founder set would typically have at seed [PR Newswire, September 2026] [Dealroom News] [FundUp].
The product differentiation rests less on a generalized AI application layer and more on the database structure itself. According to Keewano and investor descriptions, KeewanoDB stores user, device, and AI-agent actions as ordered event sequences rather than forcing teams to reconstruct context from tables; the company also says the system can query 250 million events in less than half a second, though that performance claim should still be treated as an early product assertion rather than an independently benchmarked market standard [Hetz Ventures] [Citybiz] [Israel Defense, September 2026].
The commercial model appears consistent with a developer-platform motion. Keewano offers both local and managed cloud deployment, has publicly surfaced a Head of Developer Relations opening, and lead investor Hetz says the product does not charge per event, which suggests the company is trying to remove a pricing friction that matters in high-volume event workloads [Keewano, retrieved 2026] [Hetz Ventures Careers, July 2026] [Hetz Ventures].
Funding support is meaningful for this stage, even if the operating proof points remain thin. Public sources consistently describe the round as $12 million in total seed funding with participation from Hetz Ventures, a16z speedrun, Remagine Ventures, and DIG Ventures, but they do not identify customers, production deployments, or revenue signals yet [PR Newswire, September 2026] [SiliconANGLE, September 2026] [Calcalist Tech, September 2026].
Over the next 12 to 18 months, the main question is whether Keewano can convert a technically interesting database thesis into repeatable adoption among teams running agentic, analytics-heavy workloads. The signals to watch are customer references, evidence that the claimed performance advantage holds in production, and whether the company can define a durable wedge in a market where the need is clear but the category language is still ahead of hard buying behavior [SiliconANGLE, September 2026] [PR Newswire, September 2026] [Keewano, retrieved 2026].
Single-source, plausible -- Core company, funding, and founder facts are corroborated by multiple public sources, but product performance and several operating claims remain company or investor sourced.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Business Model | API / Developer Platform |
| Industry / Vertical | Deeptech |
| Technology Type | AI / Machine Learning |
| Geography | Middle East / North Africa |
| Growth Profile | Venture Scale |
| Founding Team | Co-Founders (3+) |
| Funding | Seed, total disclosed about $12,000,000 |
The Company in Brief
PUBLIC
Keewano entered the market with a fairly specific claim: databases built for human analysts are not a clean fit for AI agents that need to reason across long behavioral histories. On its website, the company describes KeewanoDB as an event-oriented database for "machine reasoning at scale," and public company materials place the business in Tel Aviv, Israel [Keewano, retrieved 2026]. Keewano was founded in 2024, according to the company website and company profiles that repeat the same founding year [Keewano, retrieved 2026].
The public milestone record is still short, which is typical for a company at this stage. The clearest dated inflection point is September 2026, when Keewano announced the launch of KeewanoDB alongside $12 million in total seed funding, led by Hetz Ventures with participation from a16z speedrun, Remagine Ventures, and DIG Ventures [PR Newswire, September 2026]. The company website at retrieval in 2026 also shows a broader product surface around that core database, including local and cloud deployment options and analytics tooling aimed at querying event histories for AI use cases [Keewano, retrieved 2026].
Single-source, plausible -- Confirmed primarily by the company website, with milestone timing corroborated by PR Newswire.
What They Have Built
Mixed sourcing
Keewano is making a specific database claim, not a general AI-platform claim. KeewanoDB is described as an event-oriented database built for AI agents and machine reasoning, with the core idea that actions from users, devices, and agents should be stored as complete, ordered, contextual event sequences rather than reconstructed from conventional tables at query time [Keewano, retrieved 2026] [Hetz Ventures] [SiliconANGLE, September 2026]. Public coverage frames that architecture around agent workflows that need to inspect long behavioral histories and unfamiliar patterns, which is a narrower and more technically opinionated wedge than a standard analytics datastore pitch [Calcalist Tech, September 2026] [PR Newswire, September 2026].
The public product surface is broader than the database alone, although most details come from company-controlled sources. Keewano's website lists KeewanoDB Local, which runs the full engine on a single machine, and KeewanoDB Cloud, a managed version for distributed production workloads [Keewano, retrieved 2026]. It also describes Keewano Analytics for open-ended questioning across the event record, Ask for Slack for querying data inside Slack, MCP for Agents for direct agent access to analytics, and Signal for surfacing emerging threats or opportunities [Keewano, retrieved 2026]. The website also says the platform is built on GCP and Cloudflare [Keewano, retrieved 2026].
Performance is the most concrete technical claim in circulation, but it remains lightly verified. Multiple outlets repeated the company's statement that KeewanoDB can query 250 million events in less than half a second, including Citybiz and Israel Defense, which gives the metric some cross-publication consistency even if both appear to rely on launch-period company information [Citybiz] [Israel Defense, September 2026]. A separate commercial claim, that KeewanoDB does not charge per event, appears in Hetz Ventures' portfolio description and may matter if the product is targeting high-volume agent telemetry or behavioral data, though pricing mechanics beyond that were not publicly detailed in the reviewed sources [Hetz Ventures].
Unconfirmed -- Core product positioning is corroborated by public press and investor coverage, but most feature-level detail and infrastructure information in this section come from the company website or investor materials.
Market Size and Demand
PUBLIC
The market matters now because Keewano is trying to sit in the path of a real infrastructure shift: AI agents need long event histories and fast retrieval, while most public coverage still describes existing databases as optimized for human-written queries rather than machine-led investigation of unfamiliar patterns [Calcalist Tech, September 2026] [SiliconANGLE, September 2026].
The difficulty is that no cited third-party market report in the provided record sizes a dedicated "database for AI agents" category, so a conventional TAM, SAM, SOM build is not supportable from the available evidence. The cleaner approach is to treat Keewano as participating in adjacent markets that are already established in public discourse: operational analytics infrastructure, event data systems, and the emerging software stack for agentic AI applications [SiliconANGLE, September 2026] [Calcalist Tech, September 2026]. That framing is narrower than the company's own category claim, but it is also more defensible.
Demand drivers are visible even in the launch coverage. Multiple sources describe the product around the same problem shape: organizations generate large volumes of behavioral or operational events from users, devices, and now AI agents, and those records become more useful when preserved as ordered sequences rather than flattened tables [Hetz Ventures] [SiliconANGLE, September 2026]. Keewano's reported performance claim, querying 250 million events in less than half a second, is best read as an attempt to show that this architecture can serve production-scale investigation workloads, not just archival storage [Citybiz] [Israel Defense, September 2026].
The company is also positioned against substitute approaches rather than only against a cleanly bounded new category. One substitute is the standard analytical stack built on tabular databases and downstream transformations, which coverage says was designed around predefined questions from human analysts [Calcalist Tech, September 2026]. Another is the broader class of event pipelines and observability or product-analytics systems that already store time-ordered records, although the provided sources do not identify named vendors or direct feature comparisons. That gap matters, because the commercial question is not whether event data exists, but whether buyers will replace or augment existing systems for agent-facing retrieval.
Regulatory and macro forces are present, but only indirectly in the cited material. The strongest macro tailwind is the broader enterprise push to operationalize AI agents and machine reasoning, which has created demand for infrastructure that can supply context quickly and consistently [PR Newswire, September 2026] [SiliconANGLE, September 2026]. A second force is deployment flexibility: Keewano publicly markets both local and cloud versions of its database, which may matter for teams balancing latency, cost, and data-control requirements, even though the sources do not map those choices to specific compliance regimes or regulated industries [Keewano, retrieved 2026].
| Public market signal | What the sources support | Evidence quality |
|---|---|---|
| Category definition | Keewano is presented as a database for AI agents and machine reasoning, but no third-party market size for that exact category is cited in the record | YELLOW |
| Demand driver | Event-heavy workloads from users, devices, and AI agents create need for ordered contextual history | YELLOW |
| Substitute market | Traditional tabular analytics databases are described as the incumbent approach being challenged | GREEN |
| Deployment tailwind | Local and cloud deployment options suggest appeal across development and production environments | RED |
The table reads less like a mature market map and more like a category-formation snapshot. Public evidence supports the existence of a real technical pain point, but it does not yet support a quantified market boundary or a proven budget line dedicated to agent-oriented databases.
Single-source, plausible -- Based mainly on named launch coverage from Calcalist Tech, SiliconANGLE, PR Newswire, Citybiz, Israel Defense, plus company and investor descriptions for product framing.
Who Else Is Fighting for This
MIXED
Keewano is positioning itself against both conventional analytical databases and newer AI-data stacks by arguing that event history, rather than tables alone, is the native unit for agent reasoning [Calcalist Tech, September 2026] [SiliconANGLE, September 2026].
The public record does not name direct head-to-head competitors, which matters here because the competitive set is still better described by architecture than by a settled vendor list [PR Newswire, September 2026]. On one side sit incumbents in analytical data infrastructure, the systems built around structured tables and predefined queries that Keewano says force developers to reconstruct behavior from fragmented records [Calcalist Tech, September 2026] [SiliconANGLE, September 2026]. On another side are adjacent substitutes, meaning data lakes, warehouses, and internal pipelines that can store event logs but are not presented in the available coverage as purpose-built for agent investigation across long contextual sequences [Hetz Ventures] [Keewano, retrieved 2026].
That leaves Keewano competing first on product framing and team credibility, not yet on a publicly demonstrated customer roster. The wedge is clear enough in the sources: KeewanoDB stores actions from users, devices, and AI agents as ordered contextual event sequences, and the company says it can query 250 million events in under half a second [Hetz Ventures] [Citybiz] [Israel Defense, September 2026]. Still, the durability of that edge is less clear because the strongest differentiators in the public record are company or investor descriptions, not independent evidence of production adoption, migration wins, or ecosystem lock-in [Keewano, retrieved 2026] [Hetz Ventures].
Where the company appears strongest today is talent concentration around large-scale analytics and infrastructure. The founding group is associated in public sources with Plarium, TestProject, and Devalore, and multiple sources tie the team to prior exits and gaming-scale analytics work [Dealroom News] [FundUp] [PR Newswire, September 2026]. In practical terms, that can matter more than branding in an early database company, because the first contest is often whether technically demanding buyers believe the team understands scale, latency, and failure modes better than a generic application-layer startup would. That edge can persist if the team ships quickly and converts its event-oriented model into developer adoption, but it is perishable if larger database vendors absorb similar agent-facing query patterns into existing platforms.
The exposure is equally visible. Keewano does not yet appear, from the cited public material, to own a distribution channel, a named systems-integrator network, or a set of reference customers that would let it outflank entrenched data platforms in enterprise procurement [PR Newswire, September 2026] [SiliconANGLE, September 2026]. Its open Head of Developer Relations role points to the likely bottleneck: the company still needs developer mindshare and implementation trust, both of which incumbents usually have in abundance [Hetz Ventures Careers, July 2026]. A reasonable 18-month scenario is that Snowflake becomes the winner if enterprise buyers decide agent analytics can be handled inside existing warehouse estates, while Keewano is the loser in that case if its differentiation remains technical but unproven in broad production use. The reverse case is narrower but credible: Keewano can emerge as the winner if agent-native workloads genuinely require event-sequence retrieval and context assembly that conventional table-first systems handle inefficiently, especially among engineering teams willing to adopt a new data layer for AI-specific use cases [Calcalist Tech, September 2026] [SiliconANGLE, September 2026].
Opportunity
PUBLIC
If Keewano executes, the prize is not a niche developer tool but a new systems layer for how AI agents store, retrieve, and reason over behavioral history at production scale [PR Newswire, September 2026] [SiliconANGLE, September 2026].
The headline opportunity is straightforward: become the default event database for agentic applications that need ordered history, not just rows and columns. That is a large claim, but the public record at least points to a credible wedge. Keewano launched with a product explicitly designed around complete event sequences for users, devices, and AI agents, rather than conventional tabular storage, and multiple publications repeat the company's core argument that agents investigating unfamiliar patterns need a different data architecture than human analysts asking predefined questions [Hetz Ventures] [Calcalist Tech, September 2026] [SiliconANGLE, September 2026]. The technical promise is also concrete enough to matter commercially, with coverage citing query performance of 250 million events in under half a second, if borne out in customer environments [Citybiz] [Israel Defense, September 2026]. Paired with a $12 million seed led by Hetz Ventures and participation from a16z speedrun, Remagine Ventures, and DIG Ventures, the setup looks more like an early infrastructure bet with platform ambition than a narrow feature company [PR Newswire, September 2026] [Calcalist Tech, September 2026].
The upside paths are still conditional, but they are legible from the product surface already in market.
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Agent memory system of record | KeewanoDB becomes the database teams choose when AI agents need persistent, queryable event history across product, security, and operations workflows | Broader enterprise adoption of agentic systems increases demand for infrastructure built around long event sequences rather than static tables [SiliconANGLE, September 2026] [Calcalist Tech, September 2026] | The company is already positioned specifically for AI agents and machine reasoning, not general analytics, which gives it a clearer category claim than a generic database launch [PR Newswire, September 2026] [Keewano, retrieved 2026] |
| Developer platform with workflow pull | Keewano expands from the core database into an adopted toolkit that includes cloud deployment, analytics, Slack access, and agent-facing interfaces | Product bundling lowers adoption friction for engineering and data teams that want one event stack rather than separate storage and reasoning layers [Keewano, retrieved 2026] | The public product lineup already spans Local, Cloud, Analytics, Ask for Slack, MCP for Agents, and Signal, which suggests management is designing for platform breadth from the outset [Keewano, retrieved 2026] |
| Security and operations wedge | Keewano wins first in security, monitoring, or operational use cases where ordered event trails and fast retrieval are especially valuable | A category-tipping early deployment demonstrates that event-native storage improves investigation speed or model performance in high-volume environments [Israel Defense, September 2026] [Hetz Ventures] | Public messaging around threats, opportunities, and behavioral histories implies the architecture can map to security and operational reasoning tasks, not only product analytics [Keewano, retrieved 2026] [Hetz Ventures] |
What compounding looks like here is less a consumer-style network effect and more an infrastructure flywheel. If Keewano can land developers on the core database, adjacent products such as Keewano Cloud, Analytics, Ask for Slack, MCP for Agents, and Signal can increase switching costs and make the platform more deeply embedded in daily workflows [Keewano, retrieved 2026]. The posted Head of Developer Relations role is a small but useful signal that distribution may be built through developer adoption rather than only top-down enterprise sales, which is often how early infrastructure products create durable pull before large contracts arrive [Hetz Ventures Careers, July 2026]. The pricing posture cited by Hetz Ventures, that KeewanoDB does not charge per event, also matters if accurate, because it aligns the product with workloads where data volume itself is the moat and where customers resist cost structures that penalize higher usage [Hetz Ventures].
The size of the win is hard to anchor precisely because no public market-sizing source is in the record and no direct public comparable is supplied in the source set. Even so, the scenario can be stated plainly: if Keewano becomes a standard data layer for agentic systems, then the company could plausibly grow into a multibillion-dollar infrastructure asset (scenario, not a forecast), because it is attacking a foundational layer, storage and retrieval for machine reasoning, rather than a single application workflow [PR Newswire, September 2026] [SiliconANGLE, September 2026]. That outcome still depends on proof that the performance claims generalize beyond launch messaging and that adoption emerges before larger database vendors package similar features, but the prize is large enough to justify attention.
Single-source, plausible -- Relies on multiple public sources for funding, positioning, and team, but the core product breadth and some commercial implications remain primarily company or investor described.
Sources
From the public record
[Keewano, retrieved 2026] Keewano | The Database for Machine Reasoning at Scale | https://www.keewano.com/
[PR Newswire, September 2026] Keewano Raises $12M and Launches KeewanoDB, the First Database Built for Machine Reasoning at Scale | https://www.prnewswire.com/news-releases/keewano-raises-12m-and-launches-keewanodb-the-first-database-built-for-machine-reasoning-at-scale-302878769.html
[Calcalist Tech, September 2026] Keewano raises $12 million Seed to build a new database for the age of AI agents | https://www.calcalistech.com/ctechnews/article/rje2hh8kfx
[SiliconANGLE, September 2026] Startup Keewano launches agent-focused database and pulls in $12M in funding | https://siliconangle.com/2026/09/15/startup-keewano-launches-agent-focused-database-and-pulls-in-12m-in-funding/
[Hetz Ventures] Hetz Ventures Portfolio: Keewano | https://www.hetz.vc/portfolio/keewano
[Dealroom News] Keewano raises $12M to build a database for AI agents | Dealroom News | https://dealroom.co/news/150899-keewano-raises-12m-to-build-a-database-for-ai-agents/
[Citybiz] Keewano Raises $12M and Launches KeewanoDB, the First Database Built for Machine Reasoning at Scale | https://www.citybiz.co/article/903158/keewano-raises-12-million-launches-database-built-for-ai-agent-workloads/
[Israel Defense, September 2026] Keewano Raises $12M to Build a Database for AI Agents | https://www.israeldefense.co.il/en/node/70840
[Hetz Ventures Careers, July 2026] Hetz Ventures Careers: Head of Developer Relations (AI Infrastructure) | https://careers.hetz.vc/companies/keewano/jobs/76297525-head-of-developer-relations-ai-infrastructure
Articles about Keewano
- Keewano's Database Stores 250 Million AI Events for a Half-Second Query — The Tel Aviv startup, founded by a team with prior exits, is betting that AI agents need a new kind of memory to reason at scale.