Big Tech Will Scour the Globe in Its Search for Cheap Energy

Johor is located at the southern tip of Malaysia and is famous for its beaches and mountain jungles. But Johor has a new boom industry: from data centers to power generation AIand Microsoft More than $2 billion was invested in building such a data center alone. For tech giants, electricity has become the new oil. The most advanced artificial intelligence data center may require 90 megawatts, enough to power tens of thousands of American homes. As AI applications proliferate, from chatbots to AI agents, demand continues to grow. an industrial consortium Data centers requiring 10 gigawatts (more than a hundred times today’s maximum demand) are being planned. For technology companies, ensuring cheap, reliable power is as important as silicon chips.

By 2025, big tech companies will be scouring the world for kilowatts, megawatts and gigawatts of power. At board meetings, discussions about server capacity are increasingly being overshadowed by discussions about grid capacity and the future of energy. Countries with abundant low-cost energy are taking advantage of this newfound discovery and developing policies to attract investment in AI with the enthusiasm once reserved for manufacturing.

Regions that have historically won data center arks, such as Ireland and Singapore, found that their capacity was stretched to bursting before the GenAI boom. This has created opportunities for unlikely rivals, not just Malaysia but also Indonesia, Thailand, Vietnam and Chile. Delay is more important than keeping electrons flowing.

Low-cost energy has long been a priority for businesses. Just as companies in the past clustered refineries near ports and factories near coal mines, AI companies are trying to position themselves in places with consistent access to electricity at favorable prices.

Location does ultimately matter. Half of a data center’s energy costs typically come from running cooling systems and air conditioning to prevent servers from overheating. Cool climate or coastal areas will begin to increase in demand as potential locations.

This pull to provide artificial intelligence is so powerful that big tech companies are buying dirty electricity to feed it, costing themselves and local economies Decarbonization goals at risk.

Competition among countries for data center business is fierce. Tax breaks are popular: More than half of U.S. states, including Arizona, New York and Texas, offer operators some form of tax relief or even preferential rates for purchasing land and committing to power. In Malaysia, green lane The pathway program speeds up construction approvals, eliminates red tape, and enables rapid construction of data centers and power lines.

The interplay between watts and algorithms is redrawing the map of global influence. This shift is as profound as the oil boom of the 20th century, but far less obvious. No pipelines were built and no tankers were rerouted. Instead, humble warehouses filled with servers are becoming the new geopolitical hot spots.

The extent of this shift in global influence is unclear. Real AI research — where the breakthroughs happen — will remain in research centers in San Francisco, London, Beijing and Paris. However, data centers bringing these algorithms to market will be a low-margin, stacked, sell-off business.

This electronic diplomacy will become a key pillar in the coming years. The focus of expanding artificial intelligence is no longer algorithms, but electronic products.

Countries that take advantage of this moment should be wary, however; their advantages may be fleeting as dominant economies figure out how to bring sufficient amounts of cheap, clean electricity online to incentivize domestic hosting.

For today’s energy-rich AI data center providers, the challenge is to turn this short-lived advantage into a sustainable one. Not only do they need to attract data centers, they also need to build their own durable innovation ecosystems that can thrive long after the “power boom” subsides.