Artificial intelligence (AI) and data-center expansion can double Asia-Pacific data-center capacity by 2030, from 32 GW to 57 GW, and can pressure power grids worldwide, in a “generational shift” in electricity demand.

Jones Lang LaSalle (JLL), a commercial real-estate and investment-management firm based in the United States, made the anticipation in a commentary on Tuesday. JLL said the global capacity could reach 200 GW, driven by hyperscale cloud and AI demand.

A decades-long trend of flat or falling electricity demand had reversed, making reliable power the main bottleneck in the AI economy, JLL added.

Transmission networks built for large and centralized power stations are struggling to absorb a decentralized influx of renewable energy, creating congestion in the United States, Europe, and Asia-Pacific, JLL said. Interconnection queues for new projects now stretched to four years or more in some regions, with some areas pausing new connections entirely, even as hyperscalers announced $200 billion of spending for 2026, up 51 percent from 2025.

Steven Jack, head of energy and infrastructure advisory for Europe, the Middle East, and Africa at JLL, said utilities that had forecast modest growth were now facing figures nearly double their earlier estimates. Grid congestion was creating a scarcity premium for assets with secured grid access. Without a grid connection, an energy developer had no project, he said.

Battery storage is emerging as essential infrastructure to balance constrained grids and intermittent renewable output, JLL said.

Grid constraints can prompt new tie-ups between energy developers and data-center operators, with some technology firms buying renewable assets outright to secure supply, JLL said. In liberalized Asia-Pacific markets such as Australia, India, Japan, and the Philippines, the status of a grid connection has become the first question for investors, it said.

Data center boom to lift S&SEA growth, but local gains hinge on ecosystem development – Moody’s