At the southern tip of Malaysia is the state of Johor, famous for its beaches and mountain forests. But Johor has a new boom industry: data center-to-power generation A.Iwith Microsoft Committing more than $2 billion on such a data center. For tech giants, electricity has become the new oil. A state-of-the-art AI data center may require 90 megawatts, enough to power thousands of American homes. With the proliferation of AI applications, from chatbots to AI agents, the requirements are growing. An industry association is planning for data centers requiring 10 GW (more than a hundred times the peak demand today). Securing cheap, reliable power is now as important to tech firms as silicon chips.
In 2025, major tech companies will search the world for kilowatts, megawatts and gigawatts. At board meetings, discussions about server capacity are increasingly overshadowed by discussions on grid capacity and energy futures. Nations blessed with abundant low-cost energy are capitalizing on this new advantage and are devising policies to attract AI investments with enthusiasm reserved for manufacturing.
Regions that have historically championed the data center arc, such as Ireland and Singapore, have found their capacity to explode ahead of the GenAI boom strained. This has created opportunities for unlikely competitors, not just Malaysia but Indonesia, Thailand, Vietnam and Chile. Latency is less important than keeping electrons flowing.
Low-cost energy has long been a priority for firms: just as companies in the past co-located their refineries near ports, their factories near coal mines, AI firms try to position themselves closer. Doing so where they can get consistent power — and great prices
Location ultimately matters. Half of the energy costs in a data center typically come from running the cooling system and air conditioning to keep the servers from overheating. Cooler climates or coastal areas will begin to be more in demand as potential sites.
The draw for AI to deliver is so powerful that big tech firms are buying raw power to make it happen, both on their own and on local economies. Decarbonization targets at risk.
Countries compete fiercely for data center business. Tax breaks are popular: More than half of US states—including Arizona, New York, and Texas—offer operators some form of tax break, and even preferential rates for committing to purchase land and access power. In Malaysia, Green Lane Pathway initiatives speed construction approvals by cutting through red tape to fast-track construction—and power lines—for data centers. Data regulations concessions to allow information to flow freely.
This interplay between Watts and algorithms is reshaping the global influence map. This is a change as profound as the oil boom of the 20th century, but much less visible. No pipelines are being built, no tankers are changing course. Instead, nondescript warehouses cluttered with servers are becoming the new geopolitical hotspots.
The extent to which this changes the global effect is unclear. The real research on AI—where the breakthroughs happen—will remain in research centers in San Francisco, London, Beijing, and Paris. Data centers that take these algorithms to market, however, will be a low-margin, pile-it-high and sell-it-cheap business.
This electro-diplomacy will be the main pillar for the next few years. Scaling AI is less about algorithms and more about electronics.
However, nations seizing this moment should be careful; Their advantage may prove temporary as major economies figure out how to bring enough cheap, clean power online to spur domestic hosting.
For today’s energy-rich providers of AI data centers, the challenge lies in turning this momentary advantage into a sustainable edge. They will need to go beyond attracting data centers to create their own sustainable innovation ecosystems that can thrive long after the “electricity rush” has subsided.
