From Energy to Intelligence
How offshore computing could reshape the geography of artificial intelligence
Artificial intelligence is often discussed as a software revolution. Most debates focus on algorithms, models, chips and digital platforms. Yet beneath the software lies a growing physical reality. Artificial intelligence requires enormous amounts of electricity.
As AI adoption accelerates, energy is becoming one of the defining constraints on future computing infrastructure. Across Europe, electricity networks are facing increasing pressure. Grid congestion is delaying industrial projects, datacenters face long waiting lists for new connections, and governments are searching for ways to balance growing demand with ambitious climate goals. At the same time, another paradox is emerging.
Europe is investing heavily in offshore wind energy, yet significant amounts of potential electricity are never generated because transmission networks cannot always absorb available capacity. When the grid becomes saturated, wind turbines are often curtailed and production is reduced.
The AI revolution may not be constrained by computing power alone. It may increasingly be constrained by access to energy.
What happens if the solution is not simply building more grids? What happens if computing moves closer to energy itself?
A Different Infrastructure Logic
This question sits at the heart of Project Enki, a Dutch initiative with potentially pan-European implications that is exploring the possibility of developing AI datacenters directly alongside offshore wind farms. The concept challenges a long-standing assumption about digital infrastructure. Traditionally, electricity travels to the datacenter. Project Enki proposes the reverse.
The idea raises a surprisingly simple question. Is it more efficient to transport vast quantities of electricity across increasingly congested grids, or to transport data through fiber-optic networks from locations where energy is abundant?
For decades, digital infrastructure was designed around bringing power to computing. Projects such as Enki explore whether the AI era may reverse that logic by bringing computing closer to power itself. Instead of transporting large volumes of electricity across increasingly congested networks, compute infrastructure could be located directly where energy is generated.
In doing so, the project attempts to address multiple challenges simultaneously. Potentially unused renewable energy could be captured more effectively. Transmission losses could be reduced. Cooling systems could utilise seawater rather than scarce freshwater resources. Grid constraints might become less restrictive.
While the economic and technical viability of offshore compute remains an open question, its conceptual significance is difficult to ignore.
The Energy Geography of AI
Artificial intelligence is increasingly becoming an infrastructure challenge rather than merely a software challenge. Datacenters require electricity, electricity requires generation capacity, and that capacity ultimately depends upon geography.
As a result, the future location of AI infrastructure may be shaped less by traditional technology hubs and more by access to reliable and abundant energy systems. This represents an important shift. For much of the digital era, technology companies concentrated around talent, finance and urban innovation ecosystems.
The AI era may introduce a new variable. Energy abundance. Regions capable of providing large-scale, reliable and affordable electricity could become increasingly attractive locations for future compute infrastructure. The relationship between energy and intelligence may become far more direct than many technology discussions currently assume.
Beyond Datacenters
The implications extend beyond individual projects. Across Europe, policymakers are debating digital sovereignty, cloud infrastructure and technological competitiveness. Much of that discussion focuses on software platforms, semiconductor supply chains and regulatory frameworks.
Yet infrastructure ultimately remains physical. Artificial intelligence depends upon electricity networks, energy generation, cooling systems, land availability and industrial planning. Projects such as Enki therefore highlight a broader reality. The future of AI may be shaped as much by infrastructure architecture as by software architecture.
Artificial intelligence is becoming a question of geography as much as technology.
Innovation increasingly depends upon the interaction between multiple systems. Energy systems. Digital systems.Industrial systems. Infrastructure systems. The most successful societies may not simply develop better technologies. They may become better at integrating the systems upon which those technologies depend.
From Energy to Intelligence
Whether offshore AI datacenters become commonplace remains uncertain. Significant technical, operational and economic challenges remain.
The next generation of AI infrastructure may be built not where talent is concentrated, but where energy is abundant.
Yet Project Enki raises an important strategic question. As artificial intelligence becomes more energy-intensive, should societies continue bringing energy to computing? Or should they begin bringing computing closer to energy?
The answer may help shape the next geography of digital infrastructure. And perhaps the next phase of Europe’s technological development.
Credit
Illustration generated by OpenAI’s DALL·E for Altair Media Europe
Caption
Project Enki explores how offshore wind energy and AI infrastructure could converge at sea. By bringing computing closer to energy generation, the initiative raises broader questions about digital sovereignty, energy systems and the future geography of artificial intelligence in Europe.
