Kexxu
AI and robotics company building computer-vision smart cameras and deep-learning solutions for various sectors.
Website: https://kexxu.com/
Cover Block
From the public record
| Field | Value |
|---|---|
| Name | Kexxu |
| Tagline | AI and robotics company building computer-vision smart cameras and deep-learning solutions for various sectors. [LinkedIn] |
| Headquarters | Amsterdam, Netherlands [Dealroom, July 2025] |
| Founded | 2020 [LinkedIn] |
| Stage | Seed |
| Business model | Hardware + Software |
| Industry | Deeptech |
| Technology | AI / Machine Learning |
| Geography | Western Europe |
| Founding team | Co-Founders (2) |
| Funding label | Undisclosed |
Links
From the public record
- Website: https://kexxu.com/
- LinkedIn: https://www.linkedin.com/company/kexxu
- App Store: https://apps.apple.com/us/app/kexxu-devices/id1663635610
The Short Version
PUBLIC Kexxu is an Amsterdam startup building computer-vision systems that pair custom hardware with deep-learning software, and it merits investor attention because the same platform appears to be stretching across industrial inspection, research tooling, and edge AI devices rather than a single narrow use case [LinkedIn] [Kexxu] [Dealroom, July 2025]. The company was founded in 2020, with Tijmen De Vries publicly identified as co-founder since October 2020, and the public record suggests an early-stage business still defining where repeatable demand is strongest [LinkedIn] [Dealroom, July 2025].
The product footprint is broader than most companies at this size: Kexxu describes smart cameras for steel factories, thermal cameras for anonymous marketing analytics, AI quality-control cameras for reusable cups, and an eye-tracking line that includes OpenEye and portable glasses with a stated six-hour battery life and companion cloud analysis tools [LinkedIn] [Dealroom, July 2025] [Kexxu]. That breadth cuts both ways, but the differentiator, according to public materials, is a turnkey stack that combines sensor hardware, embedded compute, and deployment options spanning on-premises and cloud environments [LinkedIn] [Kexxu] [Kexxu Devices on the App Store].
On team quality, the evidence is directionally positive but still light. De Vries is the clearest confirmed founder in the public record, while Jurriaan Schreuder is listed as CTO on The Org and presents himself on LinkedIn as a senior technical leader associated with Kexxu Robotics, and other profiles point to internal work on real-time eye-tracking neural networks and application engineering [The Org] [LinkedIn, May 2024] [LinkedIn].
The business model appears to be hardware plus software, with at least one publicly listed webshop price of $6,194.70 for the Kexxu Eye Helmet Fit and company positioning that spans device sales, embedded AI units, and cloud-based analysis workflows [Kexxu Webshop] [Kexxu]. Funding is the main public gap: no verifiable round, lead investor, or valuation is disclosed in the available sources, and Dealroom's visible profile only points to a 2 to 10 employee range (estimated) rather than capitalization detail [Dealroom, July 2025].
Over the next 12 to 18 months, the central question is whether Kexxu can convert a technically varied product set into a clearer commercial narrative, ideally with named deployments, repeat buyers, or tighter vertical focus. One public LinkedIn reference claims software for more than 100 AI cameras in steel factories globally, which, if borne out by independent evidence, would materially strengthen the case that Kexxu is moving from bespoke projects toward scalable industrial adoption [LinkedIn].
Unconfirmed -- This section relies heavily on company-controlled sources and directory profiles, with limited independent corroboration beyond Dealroom and some third-party profile pages.
Taxonomy Snapshot
| Axis | Value |
|---|---|
| Stage | Seed |
| Business Model | Hardware + Software |
| Industry / Vertical | Deeptech |
| Technology Type | AI / Machine Learning |
| Geography | Western Europe |
| Founding Team | Co-Founders (2) |
| Funding | Undisclosed |
The Company in Brief
PUBLIC
Kexxu appears to be a small Amsterdam computer-vision company built around a straightforward proposition: pair custom hardware with deep-learning software for niche workflows where off-the-shelf cameras are not enough. Public records place the company in Amsterdam and indicate it was founded in 2020, with Dealroom describing it as an AI company associated with the OpenEye wearable eye-tracking product line [Dealroom, July 2025]. The company website and storefront show a broader product ambition that spans eye tracking, industrial AI cameras, and edge processing hardware, although those product details are better treated as company claims than independently verified commercial scale [Kexxu].
The founding picture is narrower than the product surface. Tijmen De Vries is the only founder clearly supported in the permitted source set for this section, and public materials tie Kexxu's start to 2020 rather than to a later launch or financing event [Dealroom, July 2025]. No state filing, disclosed financing announcement, or legal-entity documentation was available in the cited materials used here, so the chronology that can be stated with confidence is limited to founding in 2020, establishment in Amsterdam, and subsequent public commercialization through the Kexxu website and webshop [Dealroom, July 2025] [Kexxu].
That limited paper trail matters because Kexxu's public presence suggests product iteration, but not yet a heavily documented corporate history. The website currently markets Kexxu Eye tracking hardware and related analysis tools, which at minimum shows an active commercial surface rather than a dormant holding page [Kexxu]. Beyond that, investors would still need direct diligence on incorporation details, cap table, and major operating milestones, since none were confirmed in the sources allowed for this section [Dealroom, July 2025] [Kexxu].
Single-source, plausible -- Confirmed in part by Dealroom and the company website; founding year, headquarters, and product presence are supported, but legal entity details and milestone history were not independently documented in the permitted sources.
What They Have Built
Mixed sourcing
Kexxu presents itself less as a single-product startup than as a small machine-perception shop shipping tailored systems across several narrow workflows. Public materials describe a stack that combines custom hardware, computer vision, and deep-learning models, with deployment offered both on premises and through cloud-connected tooling [LinkedIn]. The clearest product family is eye tracking: Kexxu markets Kexxu Eye as a portable gaze-recording device, Dealroom identifies OpenEye as a wearable eye-tracking product, and the company says the glasses provide six hours of battery life and connect to a mobile app for WiFi setup, calibration, recording, and cloud uploads [Dealroom, July 2025] [Kexxu] [Kexxu Devices on the App Store].
The broader product surface appears deliberately cross-vertical, which is interesting but also makes the core platform harder to isolate from public evidence. Kexxu and its LinkedIn profile cite smart cameras for steel factories, 360-degree thermal cameras for anonymous marketing-data collection, AI quality-control cameras for reusable cups, and AI tools for text labeling, transcription, translation, and biometric-data collection for research use cases [LinkedIn]. On the hardware side, the company also advertises a Deep Learning Unit, described as an edge device built on a GPU-accelerated embedded processing platform for industrial environments, and a 2024 LinkedIn post from Jurriaan Schreuder references an "Industrial AI camera processing unit" intended to replace external compute units [Kexxu] [LinkedIn, May 2024].
What can be said with confidence is that Kexxu is trying to own more of the system than a software-only computer-vision vendor would. A LinkedIn post attributed to Valerio Scacco states that Kexxu created software for more than 100 AI cameras running in steel factories globally, but that claim does not appear independently verified in named press coverage, so it should be treated as directional rather than settled fact [LinkedIn]. The same caution applies to commercial readiness across the rest of the portfolio: the website includes a webshop listing for a Kexxu Eye Helmet Fit priced at $6,194.70, which supports the view that at least some hardware is offered for sale, but public evidence is still thin on customer names, deployment depth, and repeatability across sectors [Kexxu Webshop].
Unconfirmed -- This section relies primarily on company website materials, LinkedIn profiles and posts, and directory listings, with limited independent corroboration from Dealroom.
Market Size and Demand
PUBLIC
This market matters now because Kexxu sits at the overlap of industrial computer vision, edge AI, and applied sensing, three areas where buyers are increasingly trying to move inference closer to the device rather than route every workflow through centralized systems [LinkedIn] [SNS Insider].
The public market data here is thin, so the cleanest framing is the adjacent market Kexxu appears to sell into rather than a company-specific TAM claim. The only cited third-party sizing figure in the record is for the smart camera with NPU market, which SNS Insider projects to reach $76.79 billion by 2035 at an 11.84% CAGR [SNS Insider]. That is not a direct measure of Kexxu's addressable revenue, but it is a reasonable analog for a business that combines camera hardware, on-device compute, and deep-learning software across industrial and research use cases [LinkedIn] [Kexxu].
| Market segment | Sizing claim | Relevance to Kexxu |
|---|---|---|
| Smart camera with NPU market | $76.79B by 2035, 11.84% CAGR [SNS Insider] | Closest public analog to Kexxu's mix of vision hardware and embedded AI |
The cited market figure points to a category with room for multiple architectural approaches. For Kexxu, the important detail is less the headline number than the fact pattern underneath it: demand appears to favor integrated systems that can capture, process, and act on visual data in context, whether the setting is a steel line, a greenhouse, a farm, or a research environment [LinkedIn] [Kexxu]. The same product set also suggests that Kexxu is not pursuing a single monolithic market, but a collection of narrower budgets tied to inspection, analytics, and observational research [LinkedIn] [Dealroom, July 2025].
A few demand drivers are visible from the product mix itself, even if public customer data is absent. In industrial settings, the appeal of on-premises or edge-deployed vision systems is straightforward: lower latency, tighter control over where data is processed, and the ability to run in environments where constant cloud connectivity is impractical or undesirable [LinkedIn] [Kexxu]. In research and media workflows, the attraction is different, namely lower-friction capture and analysis for eye tracking, labeling, transcription, and translation tasks that were historically labor intensive [Kexxu] [LinkedIn].
The adjacent markets matter because they widen the set of substitutes. Kexxu's eye-tracking products sit near user research hardware, behavioral science tools, and sports analytics equipment, while its industrial AI cameras sit closer to machine vision, quality-control automation, and agricultural sensing systems [Kexxu] [LinkedIn]. That breadth can help a small company find niche demand, but it also means each wedge faces a different buyer, procurement cycle, and incumbent alternative.
Regulatory and macro forces cut both ways. On one hand, privacy-conscious sensing can benefit vendors that process data locally or frame analytics as anonymous, as Kexxu does in marketing-related thermal camera use cases [LinkedIn]. On the other, any company spanning biometrics, industrial monitoring, and cloud-connected devices will face a patchwork of compliance expectations around data handling, workplace monitoring, and research use, none of which are trivial for a small team to absorb [Dealroom, July 2025] [Kexxu Devices on the App Store].
The market signal is real, but it is still easier to validate category demand than Kexxu's precise share of it. The public record supports a favorable backdrop for embedded vision systems; it does not yet show which vertical, if any, is large enough for Kexxu to build repeatable go-to-market around.
Single-source, plausible -- Section relies on one named third-party market report for sizing and otherwise on company and profile sources to map adjacencies and demand drivers [SNS Insider] [LinkedIn] [Kexxu].
Who Else Is Fighting for This
Market map
MIXED Kexxu appears to sit less as a single-product vendor and more as a small applied computer-vision shop, which means its real competition changes by use case rather than by one clean peer set [LinkedIn] [Dealroom, July 2025].
| Company | Positioning | Stage / Funding | Notable Differentiator | Source |
|---|---|---|---|---|
| Kexxu | Amsterdam-based AI and robotics company building computer-vision smart cameras, custom hardware, and deep-learning systems across industrial, research, and media workflows | Seed, funding undisclosed | Combines on-premises and cloud deployment with custom hardware across multiple narrow use cases, including eye tracking and industrial vision | [LinkedIn] [Dealroom, July 2025] |
| Kornea.AI | Named competitor in the available research set | Publicly unverified from provided materials | Competitive relevance is implied by overlap in AI vision positioning, but the supplied evidence does not establish product scope or funding history | [STRUCTURED FACTS] |
The competitive picture splits into at least three segments on the public record. In industrial vision, Kexxu points to steel-factory cameras, reusable-cup quality control, and AI cameras for factories, greenhouses, and farms [LinkedIn] [Kexxu]. In research and behavioral measurement, it markets Kexxu Eye and OpenEye as wearable eye-tracking tools, while in adjacent media and analytics workflows it references transcription, translation, product research, and anonymous marketing-data collection [Dealroom, July 2025] [Kexxu] [LinkedIn].
That breadth is the interesting part, but it also complicates direct comparison. Incumbents in machine vision tend to compete with deeper vertical sales coverage and more established deployment footprints, while adjacent substitutes can include standalone eye-tracking hardware vendors, general camera integrators, and software-only AI tooling for labeling or analysis. Kornea.AI
Kexxu's clearest edge today is product flexibility. The company presents itself as willing to pair custom hardware with deep-learning models and to deploy both on premises and in the cloud, which is a practical advantage in industrial settings where latency, privacy, or connectivity constraints matter [LinkedIn]. A second, narrower edge may be accumulated application knowledge from specialized deployments, especially if the claim of software running on more than 100 AI cameras in steel factories holds up, though that specific figure traces to an individual's LinkedIn statement rather than independent reporting and should be treated cautiously [LinkedIn].
That edge also looks perishable unless it converts into one of the harder moats: proprietary data, repeatable distribution, or deeply embedded workflows. Small applied AI and hardware teams can win early because they move faster on custom jobs, but they are often exposed when a customer asks for global support, long procurement cycles, formal compliance, or multi-site servicing. On the public record, Kexxu has not yet shown named customers, channel partners, or disclosed financing that would signal those scaling advantages [Dealroom, July 2025] [LinkedIn].
The main exposure is straightforward. A better-capitalized competitor, including Kornea.AI Kexxu is also vulnerable to category specialists in eye tracking if buyers prioritize validated research tooling over general computer-vision breadth, and to industrial incumbents if buyers prioritize service coverage and procurement certainty over customization.
The most plausible 18-month scenario is that the field separates by focus. Winner if X: Kexxu, if it can turn its current spread of use cases into one repeatable wedge, most likely an industrial vision niche where on-device processing and custom hardware matter and reference deployments can compound [LinkedIn] [Kexxu]. Loser if Y: Kornea.AI, if the category shifts toward deeply integrated edge deployments and it lacks comparable hardware-software packaging, though the evidence on Kornea.AI
Inferred, not confirmed -- Kexxu positioning is supported by company and directory sources, but the section relies heavily on company-described capabilities and only one named competitor is provided in the fact set, with limited corroborating detail.
Opportunity
PUBLIC
The prize here is not a niche device business, but the chance to become a specialized machine-perception supplier across several applied markets where customers want a working vision system rather than a pile of components.
The clearest upside case is that Kexxu becomes a compact but defensible full-stack computer-vision vendor for hard-to-serve workflows, especially those that sit between industrial inspection, research instrumentation, and edge AI deployment [LinkedIn] [Kexxu]. That outcome is reachable, at least in outline, because the public record shows more than a single product concept: Kexxu describes on-premises and cloud AI solutions, industrial smart cameras, eye-tracking hardware, and an edge processing unit built for industrial settings [LinkedIn] [Kexxu] [Kexxu Devices on the App Store]. Even allowing for the fact that much of this evidence is company-sourced, the pattern matters. A startup that can package hardware, embedded compute, model deployment, and analysis software into one workflow has a better shot at sticky adoption than one selling a stand-alone sensor.
That matters most if the company can standardize around a repeatable wedge. The strongest visible candidate is applied vision in environments where general-purpose camera vendors are not enough, such as steel operations, factory equipment, greenhouses and farms, or research settings that need portable gaze capture and analysis [LinkedIn] [Kexxu]. One public LinkedIn post claims software for more than 100 AI cameras running in steel factories globally, which, if representative of a real production footprint, would point to the kind of installed base that can anchor adjacent products, support services, and replacement cycles [LinkedIn].
| Scenario | What happens | Catalyst | Why it's plausible |
|---|---|---|---|
| Industrial inspection standard | Kexxu turns its camera-plus-edge-compute stack into a repeatable quality and monitoring product for steel, factory equipment, and agricultural environments | A reference deployment converts into a standardized package using the industrial AI camera processing unit and existing factory camera work [LinkedIn] [LinkedIn, May 2024] [Kexxu] | The company already markets smart cameras for steel factories and AI cameras for factories, greenhouses, and farms, suggesting at least a cross-vertical product thesis rather than a one-off build [LinkedIn] [Kexxu] |
| Research and analytics instrument platform | Kexxu Eye and OpenEye become a lower-cost entry point for labs, product research teams, and media users that need portable eye tracking with cloud analysis | Wider adoption of the portable glasses, mobile app controls, and browser-based analysis workflow | Kexxu publicly offers eye-tracking glasses, a mobile control app, and cloud analysis tools, and sells at least one eye-tracking configuration through its webshop, which indicates a productized motion rather than pure consulting [Dealroom, July 2025] [Kexxu] [Kexxu Devices on the App Store] [Kexxu Webshop] |
| Full-stack edge AI integrator | Kexxu becomes the outsourced machine-perception team for smaller industrial and research buyers that lack in-house vision expertise | Customers choose integrated hardware, model deployment, and analysis over assembling third-party parts | Kexxu consistently describes itself as building custom hardware and deep-learning models for on-premises and cloud use, which supports the case for a bundled offering [LinkedIn] [The Org] |
The compounding mechanism would come from packaging and data rather than raw model novelty. If Kexxu wins an initial deployment in a setting like steel inspection or reusable-cup quality control, it can reuse hardware design, edge compute architecture, annotation workflows, and customer-specific model tuning across adjacent sites and neighboring verticals [LinkedIn] [Kexxu]. On the eye-tracking side, the same logic applies in a different form: hardware sales can pull through usage of cloud analysis tools, while accumulated recordings and labeling workflows may improve product usability and shorten setup time for the next buyer [Kexxu] [Kexxu Devices on the App Store]. That is not a network effect in the consumer sense, but it can still create a practical moat through deployment know-how, switching friction, and an installed base of devices in the field.
The size of the win is easiest to frame against the market Kexxu is already touching. SNS Insider projects the smart camera with NPU market to reach $76.79 billion by 2035, growing at an 11.84 percent CAGR [SNS Insider]. If Kexxu were to build a durable position as a narrow but credible supplier inside that broader stack, even a 0.1 percent share of that 2035 market would imply roughly $76.8 million in annual revenue (scenario, not a forecast) based on that market estimate [SNS Insider]. A higher-end outcome, where Kexxu owns a meaningful slice of one or two vertical workflows and monetizes both devices and recurring software, could justify strategic value well beyond a hardware multiple, because buyers in industrial and research tooling often pay for integrated systems that remove operating complexity rather than for components alone [LinkedIn] [Kexxu].
Unconfirmed -- This section relies heavily on company website and LinkedIn product claims, with limited independent corroboration beyond Dealroom and one third-party LinkedIn deployment reference.
Sources
From the public record
[LinkedIn] Kexxu | LinkedIn | https://www.linkedin.com/company/kexxu
[Dealroom, July 2025] Kexxu company information, funding & investors | Dealroom.co | https://app.dealroom.co/companies/kexxu
[Kexxu] Kexxu | https://kexxu.com/
[The Org] Kexxu | The Org | https://theorg.com/org/kexxu
[LinkedIn, May 2024] Jurriaan Schreuder - CEO/C++ - Kexxu Robotics | LinkedIn | https://www.linkedin.com/posts/jurriaan-schreuder-968b7731_new-product-page-in-the-making-at-kexxu-activity-7198319308450480129-Iyet
[Kexxu Devices on the App Store] Kexxu Devices on the App Store | https://apps.apple.com/us/app/kexxu-devices/id1663635610
[Kexxu Webshop] Kexxu Eye Tracking Glasses €799 - Kexxu Webshop | https://kexxu.com/?v=7516fd43adaa
Articles about Kexxu
- Kexxu's Smart Cameras Are Already Watching Steel Mills and Tracking Eyes — The Amsterdam-based AI and robotics company is building a portfolio of specialized computer vision products, from wearable eye trackers to industrial quality control systems.