In a power plant or chemical refinery, the safety dashboard and the cybersecurity dashboard are often separate screens, managed by different teams with different priorities. A pressure spike might be a process anomaly for one engineer and a potential intrusion vector for a security analyst, but the correlation is rarely automatic. Indurex, a startup founded in Amsterdam last year, is betting that this separation is the critical vulnerability. Its platform, SafeGuard AI, is built to ingest data from both sides of that divide, from distributed-control systems and safety-instrumented systems to OT networks and IT security tools, and present a unified risk score [Perplexity Sonar Pro Brief, retrieved 2026]. The company calls this category "Engineering Intelligence," and it just secured a €7 million seed round to prove the concept with critical infrastructure operators across Europe and the Middle East [EU-Startups, October 2026].
The Wedge Between Safety and Security
Industrial operators have long treated process safety and cybersecurity as distinct disciplines, each with its own vendors, data silos, and response protocols. This creates blind spots. An unusual valve command might be logged by a process historian, while a suspicious network connection is flagged by a security tool, but no system connects the two events to suggest a coordinated attack on a physical safety system. Indurex's wedge is to build the connector. Its platform aggregates data from engineering systems like alarm managers and asset-management software alongside cybersecurity telemetry from OT and IT networks [Perplexity Sonar Pro Brief, retrieved 2026]. AI agents then correlate these signals, identifying anomalies and engineering deviations that could indicate a safety incident, a cyber incident, or, most critically, a blend of both.
The technical premise is that true operational resilience requires this combined context. For a buyer like an energy utility, the promise is fewer false alarms and faster, more informed incident response. Instead of two teams investigating separate alerts, they can look at a single prioritized list of issues where the engineering data informs the security risk, and vice versa [SecurityWeek, January 2026]. The platform is designed to integrate with existing third-party OT security solutions, aiming to be an intelligence layer rather than a rip-and-replace project [Perplexity Sonar Pro Brief, retrieved 2026].
A Team Built for the Industrial Grid
The founders bring a depth of field-specific experience that is rare for a company founded in 2025. CEO Jalal Bouhdada was previously the founder and CEO of Applied Risk, an industrial cybersecurity firm that was acquired by the assurance and risk management giant DNV [RocketReach, retrieved 2026]. He now also serves as Global Segment Director for Cyber Security at DNV [Manufacturing Digital, retrieved 2026]. Co-founder and COO Maarten Oosterink has a background that spans Shell and the Dutch National Cyber Security Centre (NCSC-NL) [RocketReach, retrieved 2026]. Company materials characterize the broader founding team as having more than a century of combined experience across industrial operators, OEMs, and standards bodies [Perplexity Sonar Pro Brief, retrieved 2026]. This pedigree is a significant asset for selling into regulated, risk-averse industries like oil and gas, where vendor credibility is paramount.
The early backing reflects confidence in this team and thesis. The €7 million (approximately $8 million) seed round closed in October 2026 was led by Contrarian Ventures, a climate tech and industrial tech investor, with participation from Axeleo Capital and Curiosity [EU-Startups, October 2026]. Contrarian stated that Indurex already had critical-infrastructure customers at the time of investment [Perplexity Sonar Pro Brief, retrieved 2026]. The capital is earmarked for research and development and commercial expansion, with a roadmap that targets Europe and the Middle East first, followed by the United States and Asia-Pacific [PRNewswire, October 2026].
| Founder | Role | Key Background |
|---|---|---|
| Jalal Bouhdada | CEO | Founder/CEO of Applied Risk (acquired by DNV); Global Segment Director for Cyber Security at DNV. |
| Maarten Oosterink | COO | Previous experience at Shell and the Dutch National Cyber Security Centre (NCSC-NL). |
The Integration Hurdle at Scale
The ambition is clear, but the path to widespread adoption is lined with technical and operational hurdles. The primary challenge is data integration. Every industrial site is a snowflake of legacy equipment, proprietary protocols, and customized configurations. Ingesting clean, contextualized data from distributed-control systems, process historians, and a myriad of security tools is a massive systems integration task. While Indurex promotes its AI agents, the foundational work of mapping data relationships and ensuring reliable feeds is largely a manual, consultative process at this stage.
Furthermore, the platform's value is contingent on breaking down organizational silos, which can be a heavier lift than any software integration. Convincing safety engineers and cybersecurity teams to share responsibility and align on a joint operational view requires changes to processes, budgets, and internal culture. The company's early beachhead with forward-leaning critical infrastructure operators is promising, but the real test will be moving from bespoke deployments to a repeatable, scalable product that can be configured by a customer's own team.
A Technical Breakdown of the Stack
From an infrastructure perspective, SafeGuard AI sits as a data correlation and analytics layer. It does not replace the underlying control systems or security appliances. Instead, it pulls feeds via APIs or connectors. The core technical work involves normalizing time-series data from engineering systems with event logs from security tools, then applying models to establish a baseline of "normal" operations. Deviations are scored for risk based on their potential impact on safety and security parameters.
The sober assessment for scale is latency and completeness. In a real-time industrial environment, a delayed correlation is a useless correlation. The platform must process high-volume, high-velocity data streams without introducing lag. More subtly, its risk models are only as good as the data they receive. Missing context from an unintegrated subsystem or a misinterpreted proprietary alarm code could lead to a critical blind spot. For Indurex, the next twelve months will be about proving that its unified view is not only possible in a demo environment but is operationally reliable and actionable in the controlled chaos of a live plant.
Sources
- [Perplexity Sonar Pro Brief, retrieved 2026] Indurex company and product overview | https://indurex.ai/
- [EU-Startups, October 2026] Amsterdam’s Indurex raises €7 million to strengthen cybersecurity and safety across cyber-physical systems | https://www.eu-startups.com/2026/10/amsterdams-indurex-raises-e7-million-to-strengthen-cybersecurity-and-safety-across-cyber-physical-systems
- [SecurityWeek, January 2026] Indurex Emerges From Stealth to Close Security Gap in Cyber-Physical Systems | https://www.securityweek.com/indurex-emerges-from-stealth-to-close-security-gap-in-cyber-physical-systems/
- [RocketReach, retrieved 2026] Jalal Bouhdada and Maarten Oosterink professional background | https://rocketreach.co
- [Manufacturing Digital, retrieved 2026] Jalal Bouhdada role at DNV | https://manufacturingdigital.com
- [PRNewswire, October 2026] Indurex Raises €7 Million to Bring Engineering Intelligence to Critical Infrastructure and Industrial Operations | https://www.prnewswire.com/news-releases/indurex-raises-7-million-to-bring-engineering-intelligence-to-critical-infrastructure-and-industrial-operations-302916960.html