AI Could Run Out of Power. The Biggest Risk to the Digital Economy Isn't in the Cloud

For years, we have assumed that the future of artificial intelligence depends primarily on three resources: data, semiconductors, and capital. The companies with the most advanced processors, the largest datasets, and the deepest pockets seemed destined to dominate the technological race. But a fourth constraint is becoming impossible to ignore: access to reliable electricity and the infrastructure required to deliver it. According to the International Energy Agency (IEA), data centers consumed approximately 415 TWh of electricity worldwide in 2024. Under its base-case scenario, demand could rise to roughly 945 TWh by 2030 and 1,200 TWh by 2035.

That would more than double electricity consumption within just a few years. Yet the real challenge is not simply producing enough electricity. In many regions, investors have the capital, technology, and development plans in place but cannot secure timely connections to the power grid. The IEA estimates that grid constraints could delay approximately 20% of planned data center projects. Building new transmission lines in advanced economies can take four to eight years. In other words, the pace of energy infrastructure development is increasingly falling behind the pace of digital innovation.

A New Category of Strategic Risk

From a risk management perspective, this creates a striking paradox. A company may secure financing, purchase state-of-the-art computing equipment, and sign contracts with customers—yet still be unable to launch operations on schedule. Traditional technology risk is now accompanied by infrastructure, regulatory, location, and energy supply risks. These risks are also highly interconnected. A delayed grid connection can postpone project commissioning, increase financing costs, jeopardize customer contracts, and undermine the economics of an entire investment.

The key question is no longer simply how much electricity costs. It is whether that electricity can be reliably delivered.

What Should Management Teams Be Asking?

When evaluating investments in digital infrastructure, executives should examine several critical factors:

• Actual grid connection capacity, not merely planned or requested capacity.
• Reliability, availability, and the hourly profile of electricity supply.
• Long-term energy costs and exposure to market volatility.
• The ability of the local power system to accommodate future demand growth.
• The financial consequences of delays in securing power.

Risk assessments should explicitly consider scenarios in which electricity is available in the broader market but cannot be delivered to a specific location when needed.  Such a scenario is no longer hypothetical.

The Strategic Implication

In an AI-driven economy, competitive advantage will depend on more than access to data, algorithms, and computing hardware. It will increasingly depend on whether companies can secure the energy infrastructure needed to operate those technologies reliably and at scale. The next major bottleneck in the digital economy may not be technological at all. It may be physical. And that means energy security is becoming a fundamental component of strategic risk management.