AI is expensive. Processors are expensive, data centers are expensive, power and water are expensive, data acquisition is expensive. Giants like the U.S. and China can bear these costs. But can other smaller regions—like Southeast Asia, home to the largest group of unconnected people in the world outside of Sub-Saharan Africa—keep up?
Yet experts at the Fortune Innovation Forum in Kuala Lumpur, Malaysia, last week were hopeful that smaller countries could invest in AI that works for them, even as they pointed out many of the constraints that still held back investment.
“There’s an opportunity to really leverage what has come to be known as ‘small AI,’ which is much more targeted, potentially suitable for offline use, and doesn’t necessarily compete with some of the large innovations we’re seeing [come] out of larger countries,” Mahesh Uttamchandani, regional practice director for digital for East Asia, South Asia and the Pacific at the World Bank, said.
Jon Omund Revhaug, Asia head for Telenor, agreed that there was “ample opportunity” for smaller countries to invest in sovereign AI.
Countries like Singapore, Malaysia and Thailand are trying to build their own AI industries, whether by encouraging the development of new AI models more aligned with local conditions, investing in infrastructure like power and data centers, or passing regulations to maintain data sovereignty.
Yet there’s still a lot of work to be done.
“We just need more data centers. We need to build more in Southeast Asia,” Lionel Yeo, Southeast Asia CEO for ST Telemedia Global Data Centers, said.
He admitted that a growing data center sector also needs electrical power to keep it running. “How do we secure the power all the way from upstream to downstream?,” he asked. “We have to look at collaboration across the supply chain,” he suggested, and work with “regulators to solve for power grids [and] solve for transmission and distribution.”
Water is another constraint. Singapore briefly paused data center construction in 2019 due to concerns about water overuse. The Malaysian state of Johor, too, is also warning that water might remain constrained until mid-2027, even as it tries to attract new investments in data centers and other AI infrastructure.
Yet water “opens up an opportunity for cross-border collaboration,” Uttamchandani said. “Not every country is going to warrant its own data centers,” he argued, and so resources like water and power could perhaps be shared between countries.
Talent is another issue. “There aren’t enough people with the skill sets to put [servers and data centers] together. They’re not in the right places around the world,” Wendy Tan White, CEO of Intrinsic, said.
And some of this work can’t be automated. “One of the biggest problems about putting together data centers is cable handling. At the moment, that’s still only done by human beings. There is no other way to do it,” she said.
Still, “Asia has an opportunity,” White said. “At the moment, [it’s] partly the center of manufacturing, but it has got population decline coming, and it’s dealing with geopolitics. I think it could really take a forward stance here in regulation and policy.”
Asian governments are starting to take steps to encourage more investment. Uttamchandi highlighted a recent decision in the Philippines that eliminated the need for its legislature to approve new entrants into the telecoms market. “There’s a lot of legacy legislation [and] regulation on the books that may act as a detractor,” he said.
But, at some level, supply is just not going to be able to meet the demand–which will lead to a certain amount of “self-moderation,” Yeo argued. “Everyone’s rushing to build data centers to cater to AI, but the infrastructure, the talent, the power is not going to keep up with it.”
“Businesses will have to find a way to live with the infrastructure and make themselves more efficient so they can make AI work,” he said.