DNV Ventures invests in cognitive AI company Aidalos to advance trustworthy engineering and infrastructure monitoring

Founded in Trondheim, Norway, Founded in 2024 in Trondheim, Norway, Aidalos is a deep-technology company building a new kind of artificial intelligence for complex engineering problems.

The founding team pairs hands-on structural and machine-learning engineering with deep AI research: CEO Arnulf Hagen holds a PhD in AI from the MIT/NTNU ecosystem, has held senior roles at SINTEF, SAP, and DNV - which acquired his earlier venture, proNavis, in 2009 - and is an adjunct professor at NTNU.

CTO Øystein Stranden has long experience from structural engineering and software development, most recently as Chief Engineer in SAPs Asset Performance Mangement. Responsible for anomaly detection using machine learning in the context of real-time asset monitoring.

The company has developed Daedalus, a cognitive AI multi-agent platform designed to automate 50–80% of repetitive professional engineering work while keeping a human-trustworthy standard of precision. Its mission is to turn the growing flood of industrial data into insight engineers can act on with confidence.

Aging infrastructure, rising data, and the need for AI that can be trusted

Across infrastructure and heavy industry, the pressures are mounting: aging assets, labour shortages, tightening safety and monitoring regulations, and a steep drop in the cost of sensors that makes continuous digital monitoring increasingly standard. The result is an overwhelming volume of data. A single modern bridge can carry hundreds of sensors, each streaming dozens or hundreds of readings every second - far more than any engineering team can interpret manually, and of little value unless it can be turned into reliable conclusions.

In these high-risk settings, the bar for AI is unusually high. A model that answers confidently but incorrectly is not merely unhelpful - it is dangerous. What engineers need is not just speed, but explainability, traceability, and trust. Aidalos was founded to meet exactly that need. Its origins trace back to the 2021 closure of the Stavå Bridge on the E6 in Rennebu, Norway, where analysis of sensor data flagged an anomaly on the roughly 80-year-old structure. The finding led to a lane closure and continuous monitoring that held until the bridge was replaced less than a year later - and it became the case that inspired Aidalos.

Daedalus: built to doubt

Where conventional large language models rely on statistical pattern matching and tend to deliver a single fast answer, Daedalus uses a “System 2” cognitive architecture: a network of specialised agents that collaborate, cross-check, and peer-review one another’s work, fusing generative AI with physics-based digital twins and machine learning, while built from the very start to ground itself using explicitly tracked tools rather than merely relying on "black box" statistical inferences in the LLMs. The platform provides round-the-clock autonomous monitoring of sensor data alongside review of drawings and technical reports, and it is deliberately designed to question its own conclusions - to dig deeper when assessments diverge rather than accept the first, most likely answer.

“What sets us apart is that this is not one model answering quickly, but many specialised agents that work together, challenge each other, and build up a complete assessment,” says Arnulf Hagen, CEO and Co-Founder of Aidalos.

“In engineering you cannot work with ‘probably’ — you have to know. So we built a system that is willing to doubt itself, checks again, and brings in more perspectives before it reaches a robust conclusion. We are not trying to replace engineers; we want to strengthen them, so they can spend their time on the judgements that genuinely require human expertise,” he continues.

Well positioned for growth

Although still early-stage, Aidalos has built meaningful traction across markets. In Norway, the platform is used by the Norwegian Public Roads Administration (Statens vegvesen) for real-time monitoring of the Kolomoen Bridge, with further work alongside Nye Veier. Aidalos also collaborates with DNV Maritime on pilots for Agentic Class Approval, including maritime drawing quality assurance and class-rule assessment.

“In maritime classification, our engineers assess areas such as stability, hull and instrumentation largely without AI support today, and we see real potential in tools that can make that work more efficient without compromising trust. What makes Aidalos interesting is that it is a specialised intelligence layer built for the kind of complex, high-stakes engineering decisions where transparency and traceability matter. Working closely with the team helps us learn how this technology can strengthen the assurance we provide to our customers,” says Tore Torvbråten, SVP, DNV Maritime. 

In the United States, Aidalos is, in collaboration with IBM and the University of Rhode Island (URI), working together with the Rhode Island Department of Transportation (RIDOT) for automated quality assurance of bridge-inspection reports and continuous sensor monitoring, with plans to scale significantly by 2027. This project forms the backdrop for wider talks between IBM and Aidalos to establish a strategic collaboration to jointly address the enormous US transportation market

Aidalos is also a member of the MIT Startup Exchange, has presented results at U.S. Transportation Research Board conferences, and presented at the MIT AI Conference as one of very few startups in MIT’s startup program.

Because the technology is domain-agnostic, the same core platform can be configured for sectors well beyond infrastructure — from shipping and construction to finance and insurance.

Financing round and strategic partnership with DNV

Aidalos recently closed an oversubscribed financing round, with DNV Ventures joining alongside new investors.

“We are now closing a round that gives us the strength to grow further, establishing ourselves more firmly in the US while pursuing strong opportunities in Europe,” says Arnulf Hagen.

“Having DNV on board is a significant vote of confidence in what we are building,” he continues.

The capital will fund the company’s expansion in the United States, accelerate growth in Europe, strengthen the organization, and support continued product development.

“We chose to invest in Aidalos because they are developing a new kind of AI for complex engineering problems. This is particularly relevant in sectors such as maritime, energy, infrastructure, and risk management, where explainability and trust are essential. At the same time, the company has demonstrated promising pilots not only with DNV Maritime, but also internationally, through its collaboration with RIDOT and IBM and projects in the United States,” says Kaare Helle, Ventures Director at DNV Ventures.