I remain strongly positive on AI. The pace of development, the scale of capital flowing into the sector, and the productivity upside it promises are all significant.
Equity markets have responded accordingly, with leadership concentrated in semiconductors, cloud platforms, and software. The concentration reflects real earnings power and strong visibility.
At the same time, the investment landscape around AI is broadening.
The next phase is taking shape, I believe, partly through the physical systems required to support it.
The data already signals how large this shift could be. The International Energy Agency (IEA) projects that global data centre electricity consumption could surpass 1,000 terawatt-hours by 2026, more than double recent levels.
This would bring demand close to the current annual consumption of Japan. In the United States, data centres account for roughly 4% of total electricity usage today. Projections from grid operators and industry groups suggest that share could approach 8% before 2030.
Those figures represent a structural increase in demand rather than a cyclical upswing. Meeting it requires sustained investment in generation, transmission, and distribution infrastructure.
Grid access is already emerging as a constraint. In several major markets, developers are facing delays measured in years to secure connections for new data centres. Utilities are responding with increased capital expenditure focused on transmission upgrades, substations, and network resilience.
Annual investment across US utilities is running at well over $150 billion, with a growing allocation tied to data centre demand and electrification more broadly.
This is shifting the earnings profile of the sector. Historically viewed as stable but low-growth, utilities are now positioned to benefit from a multi-year expansion cycle linked to AI, electrification, and industrial demand.
Visibility on long-term investment programmes is improving, and regulatory frameworks in many regions allow for cost recovery and returns on capital deployed.
Energy supply is only part of the equation. Transmission capacity is becoming just as important. Building generation without the ability to move electricity efficiently across regions limits the system’s effectiveness.
As such, companies involved in grid infrastructure, high-voltage transmission, and network management are gaining strategic importance.
Water is another factor moving into focus.
Hyperscale data centres rely on cooling systems that can consume several million litres per day, particularly in high-temperature environments. Research from University of California, Riverside highlights the broader water footprint associated with AI, including indirect usage through power generation. In areas facing water stress, this is influencing planning approvals, operating costs, and long-term viability.
Operators are adapting by prioritising locations with reliable access to both electricity and water, alongside favourable regulatory conditions.
This is creating a geographic dimension to the AI build-out. Regions able to provide stable infrastructure at scale are attracting investment, while others face constraints that limit development.
From an investment perspective, this introduces a second layer to the AI theme.
The first layer, meaning compute, semiconductors, and software, remains central, of course. Alongside it, infrastructure is becoming a critical enabler with its own return profile.
Opportunities are emerging across several segments. Utilities with exposure to grid expansion and renewable integration stand to benefit from rising demand and increased capital deployment.
Companies involved in energy storage and load balancing are positioned to support system stability as consumption patterns become more complex. Transmission specialists and equipment providers are essential to expanding network capacity. Cooling technology and water management systems are also likely to see increased investment as efficiency becomes a priority.
These areas share a common characteristic: they sit at points where demand is rising faster than existing capacity. That dynamic tends to support pricing power, long-term contracts, and predictable cash flows.
Policy is reinforcing the trend. Governments are placing greater emphasis on energy security, grid resilience, and domestic infrastructure. AI development is becoming part of that framework, with countries competing to attract data centre investment and the associated economic activity. Infrastructure readiness is increasingly part of that competition, shaping both public and private capital allocation.
The broader implication is clear. AI is driving a convergence between digital innovation and physical infrastructure. Growth in compute capacity cannot be separated from the systems that power and sustain it.
Markets have begun to recognise this, but positioning remains concentrated in the most visible parts of the theme. As capital requirements increase and bottlenecks become more apparent, investment is likely to spread across a wider set of sectors.
I see this as a continuation of the same underlying trend rather than a shift away from it.
AI remains the core driver. The difference is that its expansion is now placing measurable demands on energy, water, and infrastructure networks.
Therefore, I believe that investors who incorporate those dynamics into their positioning are likely to capture a broader share of the opportunity.
Nigel Green is deVere CEO and Founder
Also published on Medium.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
