Forget Silicon Valley. The real AI revolution is happening underground
Across Europe, the race for artificial intelligence (AI) talent is accelerating; however, much of that talent is being funnelled toward problems of convenience, not consequence. The brightest minds are optimising ad clicks, refining e-commerce funnels, and training models to recommend content more effectively. Meanwhile, some of our most foundational systems - like urban water infrastructure are buckling under age, climate stress, and decades of underinvestment.
It’s time to ask tougher questions about where innovation truly matters - and why AI still isn’t reaching the systems that keep our societies functioning.
A misallocation of talent
Europe is facing a well-documented shortage of skilled AI professionals. According to McKinsey, the continent could face a tech talent gap of up to 3.9 million workers by 2027. Yet public infrastructure rarely makes the top of the list for recruitment. In sectors like water and sanitation, where problems are complex but invisible, the talent gap is especially stark.
Many utilities lack the data pipelines, sensor networks, or analytical support to implement intelligent solutions
It's not for a lack of challenges - far from it. Take infiltration and inflow (I&I), a chronic problem in which rainwater or groundwater seeps into cracking sewer systems, overwhelming capacity, driving up energy consumption and operating costs, and increasing the risk of service failure. It’s a textbook example of a preventable problem that often goes undetected - not because the technology doesn’t exist, but because focus and funding rarely reach these domains.
If we believe that AI should serve the public good, then we must confront the fact that its deployment remains skewed toward sectors with greater visibility, profit potential, or perceived prestige.
The infrastructure AI forgot
Water utilities, especially those at the municipal level, are in a bind. Most are managing aging assets under tight budgets and new climate extremes. They need not just repairs, but foresight: the ability to predict vulnerabilities, detect early signs of stress, and respond quickly.
In theory, AI is perfectly suited to help. But in practice, many utilities lack the data pipelines, sensor networks, or analytical support to implement intelligent solutions. And many AI developers are unfamiliar with the gritty, real-world complexities of public infrastructure, or the critical need for continuous, up-to-date, and rigorously validated data streams to make models actionable.
This disconnect - between technical capability and real-world deployment - is precisely where smarter collaboration is needed. Not flashy product rollouts, but durable partnerships between infrastructure experts and intelligent system designers who understand the operational stakes.
A quiet shift, beneath the surface
Across Europe, partnerships between data infrastructure providers and nimble AI startups are creating scalable solutions to long-standing challenges in public utilities
Encouragingly, this shift is beginning - albeit quietly, and often out of sight. Across Europe, partnerships between data infrastructure providers and nimble AI startups are creating scalable solutions to long-standing challenges in public utilities.
One such collaboration — between Pluvion, a German AI startup, building plug-and-play solutions for the water industry and KISTERS a long-established provider of environmental data infrastructure, illustrates this emerging model. Pluvion’s first application uses AI to locate infiltration and inflow based solely on water level and rainfall data. Crucially, its performance, depends on access to a robust, validated data backbone and a web platform that visualizes the AI results.
That’s where infrastructure players like KISTERS come in - not as competitors, but enablers. Their platforms provide the scale, consistency, and quality control that AI models need to move beyond prototyping and into everyday utility operations.
Reframing innovation
What these collaborations reveal is that the problem isn’t a lack of AI potential in public utilities, but a lack of visibility, funding, and narrative support. Policymakers often talk about digital transformation in energy and transportation, but water remains an afterthought. AI funding schemes rarely take into account the unique barriers faced by small or rural water utilities.
We need to broaden the definition of innovation - to include not just what's new, but what's most needed
We need to broaden the definition of innovation - to include not just what's new, but what's most needed. That means backing projects that may not make the cover of Wired, but which strengthen critical infrastructure, reduce emissions, prevent service disruptions, and protect the health of millions.
If Europe’s climate resilience strategy is to succeed, infrastructure intelligence must be part of the equation and that includes sewers, stormwater networks, and wastewater treatment plants.
A call to action — and allocation
None of this is to say that AI shouldn’t be applied to advertising or search or virtual assistants. But if we're going to talk about responsible AI, or innovation in the public interest, then we need to address the misallocation of talent and investment.
There’s a different kind of revolution happening, one that’s not driven by billion-dollar valuations or venture rounds, but by engineers working with communities to solve chronic infrastructure problems using the tools already at hand. These aren’t prestige projects, but they are urgent.
And if we’re serious about building resilient, climate-smart cities, then we need to ensure that AI doesn’t just scale what’s profitable — but what’s necessary.
It’s time to shift the spotlight.
The real AI revolution is already happening.
It’s just happening underground.