Pomegra
Berlin's Kazimi raises an oversubscribed €2.2M pre-seed led by Market One Capital to help app developers detect bots now making up 57% of web traffic.
- Kazimi's pre-seed was oversubscribed, with IBB Ventures and mobile industry executives joining lead investor Market One Capital.
- Bot traffic hit 57% of all internet activity in 2026 per Cloudflare, up from 51% two years prior, driven by an 8,000% surge in agentic AI activity.
- The company verifies real users via cryptographic proofs that require no personally identifiable information.
Lead
Kazimi, a Berlin startup founded in 2024, closed an oversubscribed €2.2 million pre-seed round on August 24, led by Warsaw-based Market One Capital. IBB Ventures and a group of mobile industry executives participated. Valuation was not disclosed. The company provides mobile app developers with privacy-preserving technology to identify and filter bot activity - a problem that has grown from an advertising nuisance into a structural threat to the data pipelines that power AI-driven marketing.
What Does Kazimi's Technology Actually Do?
The product uses cryptographic proofs to verify that interactions inside a mobile app originate from a real human, not an automated script. Crucially, it does not rely on personally identifiable information. That matters because the identifiers fraud vendors historically depended on - device fingerprints, persistent user IDs - were deprecated by Apple's App Tracking Transparency framework and Google's Privacy Sandbox initiative between 2021 and 2024. Incumbents lost their primary detection signals. Kazimi's architecture does not depend on them.
The four co-founders - Chase Gummer (CEO), August Joseph (CTO), Ville Mikkola (COO), and Andreas Naumann (CPO) - built fraud-detection systems at Adjust, the mobile measurement firm AppLovin acquired in 2021. Their institutional knowledge of where legacy detection breaks down shapes the privacy-first design. Current clients include mobile gaming studios, rewards platforms, and fintech apps. In those sectors, bot operators have direct financial incentive: a bot that completes a task payout is not an annoyance, it is theft.
Why Is Bot Traffic a Mobile Problem Now?
Bot traffic crossed the majority threshold in 2026. Cloudflare's live tracking puts automated activity at just over 57% of all internet traffic, up from 51% two years earlier. The driver is agentic AI: bots acting autonomously on behalf of users grew 8,000% during 2025 alone, per HUMAN Security's annual reporting. Mobile ecosystems were slower to absorb this than open web infrastructure, but rewards platforms and performance marketing channels are now primary targets because they attach financial payouts to completed user actions.
Traditional detection approaches in mobile assumed access to identifiers that no longer exist. Behavioral analytics tools can still flag anomalies, but they operate downstream of where contamination enters the data. Kazimi's approach intercepts bot activity at the signal level, before corrupted data reaches attribution pipelines and distorts the targeting models that app developers use for paid acquisition. When 30% of reported installs are bots, the entire dataset training the next campaign is wrong at the source.
The Investor Angle
Market One Capital manages more than €140 million across multiple funds and has backed over 80 European companies, three of which reached unicorn valuations. The firm's typical focus is marketplaces and network-effects businesses. Mobile advertising qualifies on both counts: verified signal quality is the scarce resource that both buyers and sellers of app installs need, and Kazimi sits at that clearing point. IBB Ventures, the Berlin state-backed fund, adds local institutional support and deepens ties to the German startup ecosystem.
The round's oversubscribed status at the pre-seed stage signals that the founders attracted more committed capital than the €2.2 million they chose to take. No detail was provided on how much additional interest was declined or at what terms.
Outlook
The structural conditions that created Kazimi's market are not reversing. Compute costs for bot operations continue falling, open-source agent frameworks lower the technical barrier further, and performance marketing's cost-per-action model keeps financial incentive intact. For app developers, the cost of ignoring bot contamination has shifted from wasted ad spend to corrupted AI training data - a harder problem to price but a more durable one to solve.
The near-term question is whether Kazimi can sell cryptographic verification as a standalone product or whether it needs to integrate into existing attribution and measurement platforms to reach buyers at scale. Incumbents who decide to acquire rather than rebuild would find a €2.2 million pre-seed an efficient entry price. Kazimi will use the capital to scale its technology and expand internationally.
