Asia data centre boom sparks $1 trillion investment rush

Asia Pacific's data centre development pipeline has surged to a record 26.5GW, marking an unprecedented expansion that will nearly triple the region's current operational capacity of approximately 15GW by 2030. This explosive growth is particularly pronounced in Southeast Asia, where the pipeline capacity is nearly four times the existing operational footprint, driven by a combination of hyperscale demand and cloud adoption.

Malaysia and Thailand are emerging as pivotal hubs in this capacity boom, with Malaysia leading Southeast Asia’s upcoming IT load pipeline at over 6GW—largely centered in Johor—while Thailand’s future pipeline is estimated at around 3.5GW, anchored by Bangkok. Singapore, despite supply constraints, continues to maintain a robust presence with 45 operational data centres and benefits from regional initiatives like the Johor-Singapore Special Economic Zone, which facilitates expansion into Johor.

This rapid pipeline growth is underpinned by a surge in pre-leasing activity, which rose 115% between the first half of 2025 and 2026, signaling strong occupier demand and early commitments securing about 7.8GW of future capacity. Markets such as Indonesia and Thailand have seen fivefold and ninefold increases in pre-leasing respectively, providing investors with greater revenue visibility and reinforcing the need for substantial capital investment in land, infrastructure, energy supply, and specialized equipment.

While the development is widespread, it remains concentrated in a handful of key markets—Japan, Malaysia, Australia, India, and Indonesia—which are expected to account for nearly 77% of the projected $280 billion capital expenditure through 2030. Japan alone represents almost one-quarter of this investment, underscoring its strategic importance in the Asia Pacific data centre landscape and highlighting the regional focus on hyperscale and cloud-driven infrastructure expansion.

Asia Pacific's data centre market is experiencing an unprecedented surge in real estate investment, reaching a record USD $92.5 billion, which is a cornerstone driving the region's asset valuations toward nearly $1 trillion by 2030. This massive capital influx is complemented by the 14 largest publicly owned global data centre operators ramping up their capital expenditure from under $450 billion in 2025 to an estimated $750 billion in 2026, underscoring the scale of infrastructure and real estate spending fueling this expansion.

Investor confidence is further evidenced by the robust financing environment, with Asia Pacific data centre operators raising over USD $43 billion in publicly announced debt financing during 2025 and 2026. Institutional investors, as noted by Gordon Marsden of Cushman & Wakefield, increasingly regard data centres as essential components of modern infrastructure portfolios rather than niche assets, a shift that supports sustained capital deployment across the region.

The capital expenditure required to support Asia Pacific’s burgeoning data centre pipeline is substantial, with more than USD $280 billion needed through 2030. This figure is amplified by the rise of AI-driven data centres, which demand 25% to 35% more investment due to higher power-density requirements and specialized infrastructure such as advanced cooling systems, influencing how investors allocate capital in this rapidly evolving market.

Strong pre-leasing activity, which surged 115% between the first halves of 2025 and 2026, highlights robust occupier demand and enhances revenue visibility for investors. With about 7.8 gigawatts of future capacity already secured—particularly in hotspots like Australia, Malaysia, India, and Japan—this trend reflects escalating asset valuations and a vibrant deal-making environment, as evidenced by the increase from 69 deals totaling $18.42 billion in 2023 to 37 deals worth $24 billion in early 2026 alone.

Power availability and infrastructure resilience have emerged as the decisive factors determining which data centre projects progress from announcement to execution, overshadowing even land availability. Leading hyperscale operators such as Oracle, Meta, and Microsoft are securing dedicated or contracted power years in advance to mitigate execution risks, often leveraging behind-the-meter and off-grid solutions like on-site fuel-cell generation to circumvent grid delays. In emerging markets, sovereign-backed initiatives in countries like Lesotho, Turkey, China, and South Korea are playing a pivotal role in addressing power constraints, enabling the development of large-scale AI data centre infrastructure.

Despite their critical role, many Asia-Pacific data centres are constructed with standard commercial building designs, which introduces significant structural vulnerabilities and underestimates their risk profile. Features such as large open compartments and flat slab areas are ill-suited to withstand intensifying climate hazards like typhoons, floods, and earthquakes—risks that are escalating in hotspots like the Philippines, Tokyo, Mumbai, and Singapore. This mismatch extends to insurance coverage, which remains fragmented and often calibrated against office benchmarks, leaving operators underinsured against business interruption and simultaneous losses across buildings, equipment, and power plants.

The data centre power market is rapidly expanding, projected to reach nearly $71 billion globally by 2034, with Asia Pacific leading growth at a 15.5% CAGR fueled by cloud adoption, 5G rollout, and soaring AI workloads. Operators are increasingly investing in fault-tolerant Tier IV infrastructure and integrating sustainability measures such as renewable energy, battery storage, and AI-enabled power management systems that optimize energy consumption and predict failures. This strategic shift toward resilient and efficient power infrastructure is essential to support the high-density computing demands and continuous uptime requirements of modern data centres.

In the US, power availability remains a critical bottleneck for data centre expansion, with proximity to robust power infrastructure dictating site viability. Capital expenditures for data centre construction and GPU infrastructure surged by 50% from $400 million in 2025 to an expected $600 million in 2026, underscoring the escalating costs tied to power and infrastructure demands. Additionally, operators face operational challenges in maintaining GPU uptime and managing their lifespan and repair cycles, highlighting the imperative for resilient infrastructure and sophisticated maintenance capabilities to sustain continuous compute availability.

AI, cloud computing, and hyperscale providers are the primary engines propelling the explosive growth of the data center market, with investments from industry giants like Nvidia, Dell Technologies, and Broadcom fueling demand for high-density computing environments. This surge is driving a sustained hyperscale investment cycle aimed at expanding capacity to meet the intensive processing needs of AI and high-performance computing workloads, fundamentally reshaping infrastructure requirements across key regional markets.

The rapid escalation in AI workloads is catalyzing technological innovation within data centers, particularly the adoption of liquid cooling systems that address the substantial heat generated by GPU-intensive operations. This shift not only enhances energy efficiency and Power Usage Effectiveness but also aligns with corporate sustainability goals, marking a critical evolution in data center design to support the massive computational demands of AI and machine learning.

Capital expenditure on data center infrastructure is reaching unprecedented levels, with the 'Mag 7' tech companies alone spending up to $630 billion annually and projections estimating a global investment of $3 trillion by 2030. This financial surge supports an anticipated buildout of 70 to 80 gigawatts of compute power, reflecting the scale of AI-driven demand that is not only tripling data center capacity but also transforming the sector into one of the hottest investment opportunities worldwide.

While AI-driven demand is pushing data center absorption to record highs—doubling year-over-year to 25 GW in North America during H1 2026—this rapid expansion introduces operational and community challenges. Hyperscale providers are tasked with solving complex power and location issues, focusing investments in energy-rich frontier markets like Texas and Ohio, even as local support lags behind national enthusiasm for AI leadership, highlighting the need for greater transparency and collaboration to sustain responsible growth.