Southeast Asia is chasing the AI boom, but at what cost?

In a swanky penthouse office outside Kuala Lumpur, tech firm Zetrix AI is putting a grand plan into motion to get 1 million people across Southeast Asia using its artificial intelligence agent, Avatar, by the end of the year. Designed as an autonomous alter ego, Avatar is adept at dealing with tasks ranging from the mundane, like filling out forms, to the particular, such as helping influencers reply to thousands of social media messages. The ultimate goal, according to the company’s Gen Z mastermind, is to have these Avatars deal autonomously with other AI agents for day-to-day tasks, requiring human input only for final decisions. Zetrix hopes that by moving fast it can seize market share in a space that grows more competitive by the day, if not the hour. “AI will play a very fundamental role in our lives,” said C.Z. Wong, the 23-year-old developer of Avatar and Zetrix’s chief AI officer. “Regardless of whether we want it or not.” “Everyone has different forecasts, but personally … at the rate things are going, five years is my estimate for when AI becomes mainstream.” Thousands of companies across Southeast Asia have hitched their future to the AI bandwagon. It is a sharp-elbowed race to carve out lucrative niches before the competition, while keeping pace with near-daily advances in technology driven primarily by the United States and China, which hold a near-total monopoly on AI development. Southeast Asia’s major economies are keen to trumpet the upside, from the demand for locally made semiconductors, to the rare earths needed to make magnets used in AI processors and the data centres that store and process the vast volumes of information required to make the technology work. Malaysia shipped an estimated 465 billion ringgit (US$117 billion) worth of semiconductors in 2025, about a quarter of its total exports that year, according to government data. The country is also among the largest data centre markets in the region, with more than 140 facilities either completed or in the pipeline and investments valued at over US$6 billion. Singapore on Tuesday signed deals worth at least US$234 million with Google and ChatGPT developer OpenAI to cement its position as a regional AI hub, while Thailand last August approved a US$774 million plan to boost AI integration across key sectors such as education and agriculture. But the rush to capitalise on the AI boom has also exposed structural issues across the region. No safety net Southeast Asia is not currently producing enough clean energy to meet the zero-carbon power aspirations of chip factories and data centres sprouting up from Malaysia’s Penang tech hub to Vietnam’s Ho Chi Minh City. These vast computer warehouses also consume enormous quantities of water – threatening to deplete a resource already under siege from climate change. Meanwhile, the region risks potentially irreparable environmental damage from mining rare earth elements to feed chip factories already struggling to meet surging AI-fuelled demand. And the threat to jobs – and tax revenues – in a region where tens of millions are already barely getting by is not yet widely understood by the public, especially in countries where governance tends to sideline popular concerns. AI has displaced “some human-performed tasks”, according to a study by British multinational bank Barclays last May, even if there was no evidence then of widespread job losses. “With the application of AI still in its infancy, our findings might not adequately detect macro effects that will become more evident with time,” the bank added. For Anuar Husain, those effects threaten to make themselves felt sooner than he would like. A Grab e-hailing driver based in Kuala Lumpur, Anuar turned to the platform after losing his factory job during the Covid-19 pandemic – part of a wave of displaced workers who found a lifeline in Southeast Asia’s booming gig economy. Now, six years on, that lifeline is starting to fray. Singapore-based Grab announced in April that it would launch a public pilot of autonomous AI-driven robotaxis in the city state later this year. “So many people turned to e-hailing when they lost their jobs during Covid,” Anuar said. “What will happen to us once robotaxis take over?” Between his earnings and his wife’s salary, he said his family managed to eke out a decent living, despite the long hours on the road. That could all be about to change, however. “Once robotaxis become a reality, I’m out of a job,” he said. Southeast Asia’s gig economy employs more than 40 million people, according to multiple studies, most without pensions, health insurance or any form of redundancy protection. For them, automation represents an existential threat, with no safety net beneath. White-collar workers face a dimming outlook too. Standard Chartered on Tuesday said more than 7,000 people in its back-office centres in India, Malaysia and Poland would be replaced by AI by 2030. “It’s not cost-cutting. It’s replacing in some cases lower-value human capital with the financial capital and investment capital we’re putting in,” bank CEO Bill Winters was quoted as saying by Reuters. In March, HSBC also announced plans to cut as many as 20,000 jobs as part of an AI-focused transition. Japanese lender Mizuho said the same month it would replace 5,000 clerical roles in Japan with AI over the next decade. Even where AI adoption stops short of wholesale replacement, workers may find themselves doing more for less. Labour’s share of global output fell 0.6 percentage points to 52.4 per cent over the 10 years to 2024, according to the International Labour Organization – a trend economists link to the growing income flows enjoyed by asset owners as technology adoption scales up. Barclays, in a separate research note on Tuesday, insisted that increased AI adoption “does not imply rising unemployment”. “Indeed, because those without capital income need to work to earn income, employment could even rise as wages fall,” it said. For the millions of Southeast Asians already at the economic margins, that distinction between unemployment and working for poverty wages may feel academic. Their predicament is not helped by the slow – in many cases, absent – roll-out of basic AI skills and ethics training in the region. A World Bank report last June urged governments to drive AI skills uptake, noting the added complexity posed by varying education levels in low-skilled, manual workforces. “It is essential that skills development goes hand in hand with policy reforms that create jobs in order to foster a virtuous cycle between capacities and opportunities,” the report said, while also calling for comprehensive policies to support displaced workers and protect society from harmful AI use. No slowing down In a different kind of analysis, the region’s energy crunch resulting from the US-Israel war on Iran – which has choked off some 25 per cent of the pre-war global seaborne oil trade and a fifth of LNG exports amid the Strait of Hormuz’s near-total closure – would be a reason to slow down. But rather than curbing AI’s rise in Southeast Asia, some in the technology industry think the crisis will only boost demand. “I don’t see it [AI adoption] slowing down, I only see it accelerating,” said Vivek Chatrath, Malaysia country lead for cloud solutions at Microsoft. “The big constraint right now is that companies are facing a situation where their profitability is getting squeezed. The margins are getting squeezed.” If the energy crisis ended up fuelling inflation and hurting companies’ profits, Chatrath said the question being asked in boardrooms would be: “can I find a better way to actually [do] some of the work that humans do?” That logic was on vivid display at the Semicon Southeast Asia conference in Kuala Lumpur earlier this month. On the exhibition floor, where Taiwanese company Delta Electronics had a prime central booth, conversations about the future rarely strayed far from AI. Li Li Mow, Delta’s Southeast Asia automation head, dismissed the idea that the genie could be put back in the bottle. “It’s not reversible,” she said. “I can’t go back from a world where I can simulate a machine to having to build a machine just to do trial and error to see if I produced the right thing or not.” For Delta, whose “Digital Twin” technology is designed to let manufacturers test and optimise entire production lines in virtual environments before a bolt is ever turned in the real world, the shift from autonomous robots to agentic AI – systems that can plan, decide and act across multi-step tasks – represents a natural progression. “From Industry 1.0 to 5.0, from autonomous to agentic,” Li said. “I really don’t see it stopping.” Yet the gulf between industrial inevitability and commercial reality is wide, particularly for the small and medium-sized businesses that form the backbone of most Southeast Asian economies. A survey in February by global management consultancy McKinsey & Company found that nearly two-thirds of the region’s largest corporations had already fully adopted AI or were actively scaling up its use. But it also suggested that translating productivity improvements into profit while integrating new tools into existing systems was proving difficult on constrained budgets. For smaller operators, the obstacles are even harder to surmount. Ivan Ooi said his Penang-based IO Technology had seen a surge in demand for automotive electronics parts used in electric vehicles since the energy crunch prompted a shift towards EVs. But an AI upgrade for his modest 7,000 sq ft (640 square-metre) factory is a financial stretch too far. “If demand hits a certain level we will need to add more machines, maybe upgrade our machines,” Ooi said. “But we are not looking at AI machines right now.” Ready to burst? Behind the soaring valuations and conference-floor optimism, fears of the AI boom becoming a bubble ready to burst are building. AI demand has pushed the combined market capitalisation of the “Magnificent Seven” – Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta and Tesla – past US$23 trillion, a figure that accounts for roughly a third of the entire S&P 500 and is more than 37 times larger than the whole of Indonesia’s stock exchange, until recently the biggest in Southeast Asia. To believers, these numbers reflect genuine transformative value. To sceptics, they carry an uncomfortable echo of past booms and busts. Michael Burry, the independent US investor who became famous for predicting the 2008 housing crash that kick-started the global financial crisis, sounded the alarm earlier this month. “They are going straight up because they have been going straight up on a two-letter thesis (AI) that everyone thinks they understand,” he wrote about he wrote about the meteoric rise of US tech stocks’ valuations in a social media post on May 8. “Feeling like the last months of the 1999-2000 bubble,” he added, referring to the so-called dot-com bubble and stock market crash caused by extreme speculation in internet-based companies that wiped out at least US$5 trillion in value from the US Nasdaq Composite Index. Zetrix co-founder Fadzli Shah takes the long view. He believes a shakeout is inevitable. “Like everything, there will definitely be a boom-to-bust cycle,” he said. “It’s like the mobile phone space, where eventually only two players made it. You have Apple and you have Android. The rest just died out.” Whether any Southeast Asian company will find itself on the right side of such a consolidation is unclear. But those building the AI infrastructure do not see demand tailing off any time soon. For S.N. Cheong, an electrical engineer with global construction firm Fluor, the abstractions of market cycles feel less relevant than the physical reality taking shape around him. He has watched multibillion-ringgit data centres and wafer fabrication factories rise in Johor and Penang. Steel, concrete, fibre optics: the infrastructure of the AI age is going up fast and the people commissioning it are not building for a boom they expect to end. “Common people like you and me are still not using AI all that much,” Cheong said. “We still haven’t found the killer app that will get everyone using it.” “When that happens, the situation will be indescribable.”

Southeast Asia is chasing the AI boom, but at what cost?

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}A regional dash for digital dominance risks running headlong into an energy crunch and an existential threat to jobs

A regional dash for digital dominance risks running headlong into an energy crunch and an existential threat to jobs