Malaysia’s emergence as one of the region’s leading destinations for data centres is an important success story for our digital economy. Over the past several years, the country has attracted major hyperscalers, cloud providers and data centre operators, bringing substantial investment to Malaysia, strengthening our digital infrastructure and positioning us as an increasingly important location for the infrastructure underpinning Southeast Asia’s digital economy.

The momentum remains strong. In the first half of 2026 alone, Malaysia recorded RM95.8 billion in approved data centre and cloud computing investments, representing close to 44% of all approved investments during the period. The Malaysian Investment Development Authority (Mida) has noted that Malaysia is now among the world’s leading destinations for large-scale data centre investment, with demand increasingly driven by AI computing requirements.

We should be proud of what Malaysia has achieved, but good policy must evolve as technology, markets and national priorities change. The rapid expansion of the data centre industry has brought legitimate questions about electricity demand, water consumption, environmental sustainability, the number and quality of jobs created, local participation and ultimately how much economic value Malaysia derives from the resources committed to the industry. Rather than dismissing these concerns, we should use them to shape the next phase of our strategy.

The first phase of Malaysia’s data centre strategy was necessarily focused on attracting investment, establishing scale and positioning the country as a competitive regional data centre hub. That strategy has worked, and the question for the next phase should therefore no longer simply be how much more investment or capacity Malaysia can attract. Increasingly, we should ask how much economic and strategic value the country can create from every megawatt of electricity, litre of water and acre of land committed to the industry.

The scale of the resource requirements ahead illustrates why this matters. Based on projections by the Electricity Supply and Tariff Planning and Implementation Committee, or JPPPET, cited by the Ministry of Energy Transition and Water Transformation, annual electricity consumption by the data centre sector could increase from about 10,544GWh in 2026 to 73,274GWh by 2035, equivalent to about 31% of total annual electricity consumption. Over the same period, peak electricity demand across Peninsular Malaysia is projected to rise from 21.3GW to 33.5GW, driven in part by the rapid expansion of artificial intelligence (AI), data centres and other high-technology industries. The expected trajectory makes long-term energy planning increasingly important.

Water requires similar attention. Data centres operating in Johor, Selangor and Negeri Sembilan were using about 28.68 million litres of water per day as at January 2026, while projected requirements associated with data centre developments in those three states could reach about 445.04 million litres per day between 2025 and 2030. These figures are not directly comparable because one represents actual consumption by operating facilities at a particular point in time while the other reflects potential requirements from the development pipeline, but they demonstrate why water availability and efficiency must become integral to future planning.

The conclusion is not that Malaysia should stop developing data centres. It is that we must grow the industry better, with greater attention to resource efficiency, sustainability and the economic value generated from the national resources committed to it.

Malaysia is not starting from zero. The government has already recognised that the rapid expansion of the sector requires closer coordination between investment, electricity, water, connectivity, sustainability and economic development, and the Data Centre Task Force, or DCTF, provides the whole-of-government mechanism through which these cross-cutting issues can be considered.

The DCTF was established in February 2025 as a joint platform between the Ministry of Investment, Trade and Industry and the Ministry of Digital and is co-chaired by the two ministers. It consolidates inputs from key agencies, including the Energy Commission, Tenaga Nasional Bhd and the National Water Services Commission, and is intended both to streamline approvals and to ensure that projects have credible execution plans, secured resources, strong sustainability credentials and contributions to Malaysia’s local supply-chain ecosystem. Mida describes this mechanism as helping filter speculative proposals and prioritise AI-ready infrastructure.

This whole-of-government approach is necessary because a major data centre is no longer simply an investment project. Its development affects electricity generation and the grid, water availability, land use, high-capacity connectivity, environmental sustainability, state infrastructure planning and increasingly Malaysia’s broader industrial and AI strategies. The government therefore needs a mechanism capable of considering these issues together rather than through a series of disconnected approvals.

Malaysia has also established the Guidelines for Sustainable Development of Data Centres, which provide a national framework for more sustainable data centre development and operation. The guidelines explicitly seek to position Malaysia as a Southeast Asian data centre hub while improving energy efficiency, accelerating the use of renewable or clean energy and encouraging innovation in water efficiency.

The Ministry of Digital and Malaysia Digital Economy Corporation (MDEC) are now developing the overall policy for the next phase of Malaysia’s data centre and AI compute industry, while the DCTF provides the wider whole-of-government platform through which the cross-cutting implications and policy direction can be coordinated. The objective is not to put the brakes on investment, but to provide the country and industry with a clearer trajectory for future growth, the infrastructure required to support it and the outcomes Malaysia increasingly expects from new investment.

Energy will necessarily be an important part of that work. Malaysia needs to move from assessing electricity requirements largely project by project towards more integrated planning of data centre and AI compute demand, generation capacity, grid development, renewable and low-carbon energy, land and connectivity. We also need to distinguish clearly between proposed capacity, approved capacity, contracted capacity, operational capacity and actual utilisation so that infrastructure is developed against credible demand rather than speculative requirements.

At the same time, Malaysia should explore how future large-scale AI compute developments can increasingly meet part of their resource requirements through dedicated and alternative utility solutions, rather than relying solely on existing public infrastructure. For electricity, this could include investor-developed generation, renewable energy, battery energy storage, microgrids and other dedicated supply arrangements, subject to Malaysia’s energy, grid and environmental requirements. This can help support additional compute capacity while reducing pressure on existing systems and protecting the reliability and affordability of electricity for households and other industries.

The same principle should apply to water. Future developments should increasingly consider reclaimed and recycled water, dedicated non-potable water infrastructure, alternative sources and advanced cooling technologies that reduce dependence on treated potable water. The location and scale of future capacity should also be matched more closely to local resource availability.

A useful principle for the next phase is therefore that incremental demand should, where feasible, be accompanied by incremental resource solutions. Large new AI compute investments can help stimulate additional energy generation, renewable power, storage, reclaimed-water infrastructure, advanced cooling and other resource-efficient solutions rather than simply adding demand to existing systems.

Sustainability should therefore not be viewed simply as a regulatory cost. If Malaysia can combine reliable infrastructure, competitive costs, cleaner energy, efficient water solutions and strong environmental standards, sustainability itself can become part of our competitive proposition as a regional hub.

There is another reason why this policy evolution is timely: the data centre itself is changing. AI requires substantial specialised computing capacity and modern facilities increasingly support GPU-as-a-Service, AI training and inference, AI factories, high-performance computing and advanced cloud platforms. Mida’s Data Centre Nexus 2026 reflects this change, broadening the discussion from data centre investment alone to AI-driven digital infrastructure, sustainable and resource-efficient development, local ecosystem spillovers and technology accessibility.

Our ambition should therefore evolve from a regional data centre hub to a sustainable Regional AI Compute and Data Centre Hub. This is not about abandoning the data centre strategy, but about building on its success and moving Malaysia further up the value chain. The economic opportunity from AI compute is potentially much larger than the activity taking place inside the facilities themselves because the infrastructure can anchor new industries around it while enabling Malaysian businesses, researchers and innovators to build new products and services on top of it.

This gives us a simple economic strategy: build around the infrastructure and build on top of the infrastructure. The first creates new Malaysian industries and supply chains around AI compute, while the second uses access to that compute to create Malaysian research, innovation, companies, intellectual property and products. This approach is reflected in the AI compute policy concept now being developed.

Large concentrations of AI compute create equally concentrated demand for electricity, energy technology, cooling, water, electrical equipment, engineering, connectivity and specialist services. Rather than viewing these requirements solely as costs or constraints, Malaysia should use them to create new markets for Malaysian companies and, over time, new industries with regional export potential.

Energy is one of the clearest examples. The growing electricity requirements of AI compute can stimulate investment in renewable generation, battery energy storage systems, smart grids, microgrids, power electronics and energy management systems. Mida has already highlighted how rising AI and data centre demand is creating opportunities for renewable energy, advanced engineering services, battery technologies, skilled technical talent and local suppliers.

Cooling offers a similar opportunity. Higher-density AI workloads create significantly greater thermal management requirements, driving global demand for precision cooling, liquid cooling, direct-to-chip systems and sophisticated monitoring and control equipment. Malaysia should not aspire simply to import these technologies. Our engineering companies, manufacturers, universities and technology firms should be encouraged to develop capabilities and intellectual property in this emerging industry, initially serving facilities here and eventually exporting Malaysian solutions across Asean and other markets.

Water constraints can likewise stimulate investment and innovation in reclaimed-water systems, treatment technologies, recycling, monitoring and more efficient cooling technologies. The physical infrastructure itself also creates demand for transformers, switchgear, uninterruptible power supplies, racks, modular systems, fibre, cabling, sensors, control systems and precision engineering. Malaysia already possesses significant electrical and electronics and engineering capabilities, and the next phase should deliberately connect those strengths to the rapidly growing AI infrastructure market.

This is a broader conception of localisation. It should not simply mean requiring international investors to purchase more existing products locally. It should mean using the scale of Malaysia’s AI compute industry to develop competitive Malaysian technologies, suppliers and companies. Mida has already begun emphasising this approach through local vendor participation, collaboration with universities and technical institutes, high-value employment and stronger domestic supply chains.

A Malaysian cooling company that begins by supplying a facility in Johor and subsequently exports its technology across Asean creates far greater long-term value than a one-off local procurement requirement. The same could be true of Malaysian companies developing energy storage systems, water technologies, electrical equipment, power management solutions, precision engineering or AI infrastructure software. In this sense, some of the challenges created by the next generation of AI infrastructure can themselves become the industries of the next generation of Malaysia’s economy.

These new industries will require Malaysian talent, and the scale of investment being developed in the country creates an opportunity to build a new generation of electrical and mechanical engineers, AI infrastructure specialists, cloud architects, cybersecurity professionals, cooling specialists, energy engineers, data scientists and AI researchers. Talent development should therefore be designed into the next phase of policy rather than treated merely as a spillover we hope will occur.

The policy should encourage stronger partnerships between data centre operators, hyperscalers, Malaysian universities, Technical and Vocational Education and Training (TVET) institutions and industry. Universities should not simply supply graduates to the industry. They should become partners in research, technology development and innovation through industry-supported laboratories, specialist curricula, apprenticeships, industrial placements, professional certifications and joint research programmes.

Technology transfer similarly should be an important consideration. When Malaysia facilitates a major global technology investment, we should increasingly ask not only how much capital is invested or how many direct jobs are created, but what knowledge, technology and capability will remain in Malaysia because that investment happened here.

Our ambition should be for Malaysia to become known not only as a country capable of hosting advanced AI infrastructure, but also one with the talent capable of designing, operating, improving and innovating around it.

The second economic opportunity is what Malaysians can build using the infrastructure. There is a fundamental distinction between hosting AI compute and having access to AI compute. Malaysia could host billions of ringgit of world-class GPU infrastructure while our universities, researchers, start-ups and small and medium enterprises (SMEs) remain unable to afford the computing capacity they need. If that happens, we will have succeeded as an infrastructure location but fallen short of our ambition to become an AI Nation.

Our universities and research institutions need affordable access to advanced compute so that Malaysian researchers can undertake world-class AI research. Start-ups need access so that promising Malaysian ideas are not constrained simply because their founders cannot afford the computing resources needed to build, fine-tune and deploy solutions. SMEs need access so that AI adoption does not become the privilege of the largest companies, while the government and Malaysian technology companies also need access to develop services, applications and products.

Mida’s current policy direction similarly identifies wider access to AI, cloud computing and advanced technologies for local enterprises, start-ups and researchers as part of the next phase of the data centre ecosystem.

This is where the proposed National AI Compute Exchange, or NACX, could play a catalytic role. NACX does not necessarily require the government to purchase and own large quantities of GPUs. Instead, it could provide a trusted mechanism connecting available computing capacity from hyperscalers, Malaysian data centre and cloud operators, universities, government infrastructure and other providers with Malaysian demand.

Different users could have different access arrangements. Strategic government projects and public research could receive supported access, universities and researchers could benefit from preferential access, Malaysian start-ups could receive discounted compute and SMEs could be supported in their adoption journey, while larger domestic, regional and global users could access capacity commercially. The policy concept under development specifically identifies government, universities and researchers, start-ups, SMEs and industry among NACX’s intended users.

The principle is straightforward: AI compute located in Malaysia should help build AI capability in Malaysia. NACX is therefore not the data centre strategy itself, but one potential mechanism for converting Malaysia’s infrastructure advantage into a Malaysian innovation advantage.

Ultimately, access to compute is not the end objective. It is an enabler of something much larger: Malaysia’s ambition to move from being predominantly a consumer and host of technology to becoming a producer of globally competitive technology, products and services.

This is where the AI compute strategy connects directly with Made by Malaysia. We want Malaysian researchers using compute located here to develop Malaysian intellectual property, Malaysian start-ups using it to build AI applications and platforms, Malaysian SMEs developing AI-enabled products and services, and Malaysian technology companies scaling into Asean and global markets. At the same time, we want international companies investing here to help develop Malaysian suppliers, engineers, researchers and technology capabilities as part of their presence in the country.

Made by Malaysia should also extend beyond software. A cooling technology developed by a Malaysian engineering company can be Made by Malaysia, just as a Malaysian energy management platform, water-efficiency technology, piece of AI infrastructure equipment, AI model, application or platform can be Made by Malaysia.

The objective is therefore to create a broader economic chain in which data centres enable AI compute, AI compute creates new Malaysian ecosystems and talent, affordable access enables Malaysian research and innovation, and that innovation produces Malaysian companies, intellectual property, products and exports.

That is how infrastructure becomes economic transformation.

This is the opportunity before us. The overall policy being developed by the Ministry of Digital and MDEC, with the DCTF providing the whole-of-government platform for coordination, should bring these elements together into a coherent long-term direction for Malaysia’s data centre and AI compute industry. It should address the immediate challenges of electricity, water and sustainability while looking beyond them to dedicated and alternative utility solutions, localisation, talent, technology development, trusted cross-border data flows, compute access and downstream economic value.

Industry needs visibility of Malaysia’s expected compute demand, electricity and water availability, sustainability requirements, suitable locations, infrastructure development plans and the economic outcomes the government increasingly expects from investment. The government, meanwhile, needs better visibility of genuine future demand so that electricity generation, grid capacity, water infrastructure, land and connectivity can be planned efficiently. Malaysians should also have confidence that the national resources committed to this rapidly growing industry will translate into wider opportunities for our companies, workers, researchers, universities and entrepreneurs.

This creates a new compact between Malaysia and the industry. Malaysia should continue to welcome high-quality data centre and AI compute investment and provide an attractive environment from which to serve domestic, Asean and global demand. At the same time, the next generation of investment should progressively deliver more for Malaysia through sustainable and, where appropriate, dedicated resource solutions, deeper local supply chains, more Malaysian technology and high-value skills, stronger university-industry partnerships, greater technology transfer, trusted regional data flows, better access to compute and more downstream Malaysian innovation.

This is not about slowing investment. It is about increasing the value of investment while ensuring that growth remains sustainable for Malaysia.

Malaysia has built a strong foundation as a regional data centre hub, and we now have an opportunity to use that foundation to create something substantially more valuable. Our ambition should be to evolve into a competitive, sustainable and trusted Regional AI Compute and Data Centre Hub that serves Asean and global demand while deliberately creating greater economic and strategic value for Malaysia.

The policy now being developed should provide a clear pathway for that transition. It can address legitimate concerns about electricity, water and sustainability while opening the door to dedicated and alternative utility solutions; use the growth of AI infrastructure to develop new Malaysian ecosystems in renewable energy, storage, cooling, water technology, electrical equipment, advanced manufacturing and engineering; deepen localisation and Malaysian talent; strengthen Malaysia’s position as a trusted jurisdiction for regional data and AI workloads; and give our universities, researchers, start-ups, SMEs and companies greater access to the computing capacity needed to innovate.

Malaysia’s leadership in developing the Asean Guide for Sustainable Data Centre Development, led through the Ministry of Digital and MDEC, and the Malaysia-led Asean Regional Framework on Cross-Border Cloud Computing demonstrate that we are already helping to shape the regional policy environment in which this future industry will operate.

This is ultimately how the infrastructure strategy connects to Made by Malaysia. Our ambition is not merely to host the infrastructure of the global AI economy, but to use that infrastructure to build new Malaysian industries around it and new Malaysian products, intellectual property and companies on top of it.

The first phase was about attracting infrastructure. The next phase must be about turning that infrastructure into Malaysian capability, innovation, productivity and higher economic value. From capacity to compute, from compute to capability, and from capability to Made by Malaysia.

Ultimately, Malaysia’s data centre strategy must not only be measured by facilities built, billions invested or megawatts installed, but by the value it creates for the government and businesses — and above all, by the positive difference it will bring to the daily lives of all Malaysians.

Gobind Singh Deo is Minister of Digital