Data centres dominate western European FDI. And McKinsey estimates nearly 7 trillion dollars will be invested globally in data centre infrastructure by 2030 (40% expected to occur in the US). The Middle East is quickly emerging as a data centre powerhouse, with regional capacity expected to triple, from 1GW in 2025 to 3.3GW over the next five years.

For all the boom there is however a fundamental uncertainty that requires an assumption check: AI is seen to speed material recovery, optimise renewable energy eco-systems, and reduce energy intensity of industrial processes, but as outlined in the International Energy Agency (IEA, 2025) report on energy and AI there is potential for massive ‘rebound’ effects.

AI is a high stakes game and viewed as an equaliser in international trade. Cost, will feature high with intensive energy and water as key inputs. But speed also defines winners: sector priorities being speedy connections to grids over long-term sustainability choices. This matters since if AI is to unlock economic growth that is material and energy-intensive, as current trajectories suggest, we face real negative climate impacts through waste, cost and risk to threaten financial return, jobs, security, and resilience.

Is data centre energy and water demand crowding out the ability to decarbonise other high value investments and job sectors?

How does this square with ESG, sovereign security and resilience?

What can we do about it? What assumptions aren’t working?

Data centres compete for resources with green industrialisation and alternative fuels, that further compete for energy and water. Each feedback loop turns on itself, creating economic fragility, weakening resilience and creating risk as recent blackouts in Chile and the Iberian Peninsula illustrate how quickly energy interruptions cascade affecting millions of businesses and people (IEA, 2025).

The OECD have developed several frameworks and toolkits that provide stated holistic metrics to assess the quality of an investment but what they actually offer is structured indicator sets, principles and checklists that in practice build isolated project‑level rating systems. Consequently agencies rely on isolated basic indicators like job creation, project numbers or capital expenditure without understanding the relationship between isolated data points and events: measurement that can be magnitudes out from reality.

From my experience working with clients, when dealing with ambiguity, it’s helpful to not look at each project in isolation but how they interact. Merely stating a holistic approach and assuming some whole if we zoom out far enough, risks introducing preexisting biases and decision-making frameworks, since true holistic approaches consider the environment and interconnected activities. Such distortions shift focus away from actual systemic relationships, overshadowing context, and realistic measurement.

This matters because a defined holistic view can reveal when a portfolio that looks successful at project level is actually destroying value through hidden waste, unnecessary cost, unmanaged risk, and missed opportunities.

We are in the most power-hungry industrial drive in history on aging, overloaded grids and as datacentres, decarbonising industries, new fuels, desalination compete, the load is growing in every direction. The usual renewable suspects don’t scale quickly enough, have large waste challenges, and natural limits on wind and sun requiring storage to cover intermittency, while frontier data centres can go from MW demand to GW demand in minutes.

It’s time to think differently. Challenge assumptions. Go back to fundamentals. True systemic solutions offer shortening cycle times and cost by using better what we have already, reducing waste, repeating excellence and consistent high-quality production because true systemic solutions have recurring outputs that help insure against uncertainty and ESG divergence.

As a result sustainability shifts up a gear to become sustainable viability: such an approach introduces cross investment relationships that can develop and support not just AI, but cross sector such as infrastructure, materials, industrials, energy, water, new fuels…conjoining the components needed to underpin and bolster resilience and security. Reducing cost, waste and risk.

Bottom line: to tackle the complexity of coming problems, skip the bottlenecks and boost sustainably viable development, experience has shown it is good practice to challenge assumptions, check fundamentals, seek relational holistic approaches: focus on execution and iteration.

Christopher Gleadle

CEO SV-Electra