Southeast Asia’s digital economy is moving beyond apps and platforms. The focus is shifting to what lies beneath: data centres, chip capabilities, cloud networks and the physical infrastructure powering AI workloads. From Singapore’s hyperscaler campuses to Malaysia’s semiconductor labs and Indonesia’s sovereign-backed data parks, the region is investing heavily in the machinery that underpins the next phase of digital growth. As generative AI models, real-time automation, and machine learning become standard tools, governments and companies alike are waking up to a hard reality: no software strategy is complete without a compute strategy.
This infrastructure race is not driven by hype. It is a direct response to real bottlenecks. Southeast Asian startups are already pushing against the limits of cloud latency, hardware availability, and regional data restrictions. AI-powered services, particularly in sectors such as finance, logistics, healthcare, and media, require rapid, secure, and scalable access to processing power. And the global cloud platforms they often depend on are starting to feel too far away, both in physical distance and regulatory alignment.
We explore Malaysia’s AI and data centre play and whether policy can keep up with demand.
Geopolitical pressure is also forcing the issue. The US-China tech rivalry, ongoing trade restrictions, and concerns over data exposure are leading many Southeast Asian economies to prioritise technological sovereignty. This means investing in homegrown capacity: data sovereignty, local chip design, AI training infrastructure, and regional cloud redundancy. The result is a wave of public-private partnerships, sovereign fund deployments, and capital-intensive infrastructure projects. For startups, this could open new pathways to scale. For governments, it is a strategic hedge against uncertainty. And for the region as a whole, it signals a shift in Southeast Asia’s position from technology consumer to infrastructure builder.
Investment in infrastructure is accelerating
Investment in data centre capacity is climbing sharply. A recent report projects that Southeast Asia’s data centre industry, valued at about USD 13.71 billion in 2024, will grow to USD 30.47 billion by 2030, at a compound annual growth rate (CAGR) of about 14.24 percent. This growth is driven by hyperscaler investments, demand from AI workloads, and sustainable design becoming more standard. Countries are starting to build local compute infrastructure to reduce latency, comply with data localisation laws, and improve reliability. For example, Microsoft committed USD 2.2 billion over four years to developing cloud and AI infrastructure in Malaysia, including establishing an AI centre and increasing capacity for local computing. Indonesia’s sovereign wealth fund, INA, is also investing heavily in digital infrastructure. It is backing data centres, subsea cables and hybrid capital in support of domestic growth, with projects like the DayOne data centre project in Batam. Perhaps most telling is the scale of new capacity planned. According to a recent Deloitte report, Southeast Asia will see approximately 397 MW of data centre capacity go live by 2028. Total capacity is expected to more than quadruple from 1.68 GW in 2024 to about 7.59 GW in the coming years. On the chip side, Southeast Asia is positioning itself more centrally in global semiconductor supply chains. Nations like Singapore and Malaysia already have strong roles in output, assembly, testing and backend manufacturing. Governments are offering tax incentives, R&D grants, and industrial policy to attract fabs (fabrication plants) and chip design capabilities.
Regional trends and political drivers
Geopolitical uncertainty is a strong motive. Trade disruptions, export controls, and concerns about dependence on foreign computing infrastructure push governments to build capacities at home. Data sovereignty is not just a policy slogan; it is increasingly important in contracts, compliance, and public trust. Demand-side pressures are equally strong. AI applications such as generative models, machine vision, and high-resolution sensors in mobility or logistics require fast, reliable compute, often close to the edge. Cloud latency can become a user experience issue. Startups producing video, AI agents, or real-time data pipelines need local infrastructure. A strategic imperative is also visible. National governments see semiconductor capability, chip design, and data centre infrastructure as components of sovereign technology. Policies, stimulus, and public-private partnerships are aligning around infrastructure that supports AI, not only for export or service delivery but for resilience. For example, INA in Indonesia is investing in digital infrastructure as part of its sovereign fund mandate.
Opportunities and obstacles for startups
Startups that can provide or tap into hardware infrastructure, data centre operations, AI chip design, or cloud-native tools have a new opportunity. Infrastructure providers might become strategic partners for AI startup ecosystems. Accessibility of local compute can reduce costs and risks associated with cross-border data transfer or latency. Regulation will become more central, especially around data privacy, data flow across borders, content of compute, and energy sourcing. Data localisation laws, environmental regulations (for power consumption, cooling, land use) will affect where infrastructure can be built and how it must be run.
Firms that do not comply will invite legal or reputational risk. However, investment and scaling in infrastructure are intensive. Building AI-ready data centres, sourcing chip fabrication capabilities, or establishing advanced cloud environments requires heavy upfront costs, stable power supply, land, cooling, and regulatory approvals. Often, government support or subsidies are required. Much of this is happening already. For example, the DayOne-INA data centre campus in Batam, Indonesia, received a USD 411 million loan from Singapore’s banks DBS and UOB, for three data centres with a capacity of about 72 MW, expected to be completed by the end of 2025 .
Investors should evaluate not just capacity but energy sourcing, sustainability, operational efficiency, and compliance. Infrastructure that can scale sustainably, with renewable power, efficient cooling, and smart design, may attract premium valuations or lower operating costs.
Ongoing risks in infrastructure build-out
Dependence on foreign hardware supply remains a vulnerability. Despite gains, chip fabrication and advanced semiconductor fabrication infrastructures are still dominated by East Asia. Supply chain bottlenecks in GPU chips, rare earths, and advanced photolithography equipment are hard to replace quickly. Energy costs and environmental constraints are material obstacles. AI computing and data centres consume large power for cooling, transmission and constant operation. In many SEA countries, power infrastructure or land for cooling is limited. Renewable energy sources or green design are being emphasised, but may not always match demand. Regulatory fragmentation remains an issue.
Differences in data sovereignty laws, cross-border flow restrictions, environmental regulations, import tariffs, or tax incentives vary widely. Infrastructure developers must navigate local, regional, and global rules. Misalignment or delay in policy can slow down projects significantly. Financing remains challenging. Even with sovereign wealth fund involvement and bank financing, many data centre and chip projects require stable long-term revenue commitments. Investors will demand strong contracts, robust demand forecasts, and risk mitigation against energy price volatility, environmental cost burdens, and regulatory change.
What to expect in the next 12 to 18 months
Over the next 12 to 18 months, the region’s progress will be measured by how quickly planned data centre capacity becomes operational, particularly in markets beyond Singapore, Malaysia, and Indonesia. Whether the projected 7.59 GW capacity by 2028 materialises will test market integrity. Attention will also turn to whether more semiconductor fabrication or chip design startups move from policy ambition to actual implementation. Singapore, Malaysia, Vietnam, and Thailand are viable candidates, but success depends on capital, talent, and IP protection. Regulation will play a defining role as data localisation, environmental standards, and hardware export rules begin to converge. Finally, sustainability will become a key performance metric. As power costs rise and climate pressure increases, data centre operators that adopt renewable energy, waste heat reuse, and efficient cooling will outperform those that do not.
What’s next for Southeast Asia?
Southeast Asia is no longer simply a consumption market for cloud and compute. It is building its own infrastructure. The race for localised AI compute, data sovereignty, and effective chip supply chains is underway. For startups, infrastructure firms, and policymakers, success will depend on combining scale with sustainability, regulatory compliance, and technological independence. If those elements come together, this growth phase could reshape the region’s role in global technology supply chains.