AI Needs More Than Smart Chips: The Energy Crisis Behind the Tech Boom

Artificial intelligence is advancing at remarkable speed, but one major obstacle could slow its progress: electricity. According to a new Goldman Sachs report, global data center power demand could jump by 170% between 2025 and 2030. The United States alone will account for more than 60% of that growth.

Tech companies are racing to build data centers and develop increasingly powerful AI systems. Yet the infrastructure that keeps these facilities running cannot expand at the same pace. In the US, around 2,600 gigawatts of proposed power projects are waiting for grid connections. Connecting a project to the grid and bringing it into operation takes a median of five years.

Billions in Investment Face Delays and Uncertainty

The stakes are enormous. Goldman Sachs estimates that spending on computing, data centers, and power infrastructure could reach $7.6 trillion between 2026 and 2031.

However, many projects face construction delays, regulatory hurdles, and resistance from local communities. Since January 2026, 86 infrastructure projects have been canceled or significantly delayed, with 76 linked directly to local opposition.

Financing adds another layer of risk. Goldman expects global borrowing by major cloud computing companies to reach $420 billion in 2027. If facilities cannot access enough electricity on time, investors may wait longer to see returns.

There is a potential upside. Slower construction could prevent companies from pouring money into infrastructure before demand catches up. Meanwhile, energy providers stand to benefit as technology giants explore ways to fund new power generation and grid upgrades.

Our Conclusion

We believe energy access will become one of the defining factors in AI’s future. Electricity constraints could affect the cost of digital services, the speed of innovation, and how quickly businesses adopt AI tools. Companies that secure reliable power may gain a significant advantage.

The next AI breakthrough may depend not only on better software, but also on whether we can supply enough electricity to run it.

Enjoyed this article? Share it with your network and explore more of our technology insights for a closer look at the forces shaping our digital future.