In March 2026, Iranian drone strikes damaged Amazon Web Services facilities in the United Arab Emirates and Bahrain. Infrastructure built to support the Gulf’s artificial-intelligence ambitions had entered a regional battlefield, and Amazon later warned that full recovery could take several months. (reuters.com)
Southeast Asia is not the Gulf. Yet the lesson should travel quickly.
UNCTAD recorded US$244 billion in foreign direct investment into South-East Asia in 2025, making it the largest recipient subregion in developing Asia. The figure exceeded the US$188 billion received by Latin America and the Caribbean and was more than three times Africa’s US$70 billion. Digital infrastructure, especially data centres, drove much of the increase in announced greenfield investment. (unctad.org)
Governments see AI as a route to industrial upgrading. Global firms see demand. Investors and sovereign wealth funds see assets capable of producing returns and strategic influence. The danger is that many are still evaluating the boom through an outdated model of investment risk.
Stay ahead of the geopolitical week.
MD Briefing delivers expert analysis across five global fronts — the Indo-Pacific, energy, geoeconomics, European security, and the Middle East — every Monday morning. Free.
The assumptions beneath the boom
Traditional analysis divides an investment into separate files. Financial teams test returns and exit value. Technology specialists review scalability and cybersecurity. Political-risk advisers examine regulation and stability.
An AI asset refuses to remain inside those boundaries.
A data centre draws on electricity, water, land, international connectivity and specialised chips. An enterprise platform may run on an American cloud, use models developed elsewhere, integrate open-source components maintained across several jurisdictions and process data governed by national law. Each dependency may look manageable on its own. Their convergence creates the exposure.
Procurement teams can sign a cloud contract without considering export controls, while energy agencies approve supply without examining how future scarcity might affect the business model. Data regulation, public opposition and political transition then move through separate institutional channels. A single shock can connect these risks faster than any committee responsible for only one part of the asset.
The Gulf attacks provided the most visible illustration. Physical damage became an outage for customers, a recovery-cost problem for the provider and a new pricing signal for investors. Infrastructure promoted as an instrument of economic diversification acquired strategic meaning inside a conflict. (reuters.com)
ASEAN faces a different security environment, but its technology assets are also embedded in political relationships that conventional financial models often leave outside the calculation.
ASEAN is a multiplex political economy
ASEAN is commonly presented as one high-growth market. In practice, capital enters a portfolio of political operating systems.
Authority is distributed differently across political leaders, bureaucracies, regulators, state-owned enterprises, courts, local governments, business groups and civil society. Energy systems and industrial capacity vary just as widely, as do relationships with American, Chinese and other technology providers.
Viewed through a Multiplex Digital Ecosystem, the fate of an AI investment is shaped by the shifting bargain among the state, society and capital—not by policy, markets or technology in isolation. A national strategy may set direction; execution depends on how that direction is negotiated across the ecosystem.
Three markets show why this matters.
Malaysia has made data centres central to its industrial strategy. Between 2021 and mid-2025, the Malaysian Investment Development Authority approved RM144.4 billion in data-centre and cloud-computing investment. The boom then pushed energy and water deeper into the investment equation. Malaysia’s Guideline for Sustainable Development of Data Centres connects future expansion with power and water efficiency and greater use of clean energy. The opportunity remains substantial, but returns increasingly depend on whether national ambition can be reconciled with local resource politics. (mida.gov.my)
Indonesia exposes a different problem: scale combined with fragmented execution. Government Regulation No. 28 of 2025 places business licensing under the OSS risk-based system, while digital infrastructure continues to intersect with sectoral regulation, energy supply and local implementation. In its 2025 performance report, Komdigi put national data-centre capacity at approximately 370 megawatts and called for priority data-centre zones, integrated cross-sector licensing and more sustainable energy supplies. The diagnosis is revealing: market expansion is moving faster than the institutional capacity to assess its combined exposure. (jdih-storage.bkpm.go.id)
Vietnam offers stronger strategic continuity but a dense interaction between infrastructure expansion and state data priorities. The 2023 Telecommunications Law opened data-centre services to full foreign investment. The Law on Data, effective from July 2025, strengthened the architecture for national data infrastructure, while the revised Power Development Plan VIII governs the energy build-out on which new digital capacity depends. Investors gain access to a growing market, but their room to operate is shaped simultaneously by power allocation, data governance and national technology strategy. (beta-en.mic.gov.vn)
Across these three markets, the same technology asset acquires a different risk profile because authority, resources and political legitimacy are organised differently.
The Strategic Blindness Cascade
The cost of overlooking those relationships rarely arrives as six separate problems. It develops as a cascade.
Operational disruption usually comes first. Access to a facility, provider, chip or cable is interrupted, and what looked like redundancy during normal conditions suddenly determines whether the damage stays local. Financial consequences follow: insurers reconsider coverage, lenders revise assumptions and future buyers discount an asset whose concentration risk has become visible.
Lock-in deepens the damage. A company may technically be able to switch clouds, models or chip ecosystems, but only after rebuilding operations around a new architecture. Adaptation can cost more than the original disruption, turning a procurement decision into a constraint on financing, partnerships and exit.
Commercial pressure then enters politics. Governments that initially offered tax incentives, energy allocations or regulatory flexibility may respond differently once foreign dependence, resource consumption or weak local employment becomes salient. Localisation demands, ownership scrutiny and national-security review become more likely precisely when the investor has the least room to manoeuvre.
Legitimacy can deteriorate along the same path. AI projects are promoted through national gains—growth, productivity and modern infrastructure—while their costs are felt locally through land, electricity, water, automation and data extraction. When workers or communities conclude that they carry the burden while foreign platforms capture most of the value, social resistance becomes an investment variable.
The cascade reaches its final stage when a project remains financially viable but fails strategically. A sovereign wealth fund earns a return without acquiring technology or decision rights. A government hosts infrastructure without developing suppliers. A company adopts AI quickly yet becomes more dependent on vendors it cannot replace.
What policymakers must consider
ASEAN governments should treat major AI assets as economic infrastructure rather than ordinary technology-sector FDI. Their resilience can affect banking, public administration, telecommunications, manufacturing and logistics. Investment review therefore has to examine the system of dependencies created by a project, not only its capital value and employment promise.
The country cases point to different priorities. Malaysia needs to keep resource allocation and local value creation inside the data-centre bargain. Indonesia needs mechanisms capable of joining licensing, energy, data and local implementation into one exposure assessment. Vietnam needs to preserve investor clarity as data governance and power planning become more central to its technology strategy.
Across the region, incentives should be tied to outcomes that survive construction: skilled employment, supplier upgrading, research capacity, interoperability and recurring domestic value. Policy continuity also needs institutional foundations. Long-term assets cannot depend exclusively on the coalition or bureaucratic sponsor that secured their entry.
Public legitimacy belongs in the same calculation. Governments that cannot explain who benefits, who bears infrastructure costs and what capability remains domestically will eventually face pressure to rewrite the bargain.
What the market must calculate
Companies, institutional investors and sovereign wealth funds should place geopolitical technology intelligence inside the investment thesis rather than attach it as a country-risk appendix after approval.
Investment committees should begin with a harder question: how much of the asset’s value depends on political and technological conditions that the company does not control? Losing access to a critical provider could force an expensive migration after operations have already been organised around one stack. Changes in data policy, power pricing or political leadership could erode the business model, while future buyers may discount an architecture they regard as difficult to modify.
Optionality has a cost, but concentration has a hidden price. Interoperability, distributed infrastructure and credible exit rights may look inefficient until switching is no longer voluntary.
Investors should also assess the political bargain supporting the asset. Access secured through one elite relationship can disappear after a transition. A project embedded in local suppliers, talent and visible public value is more likely to survive changing leadership than one protected only from above.
Sovereign wealth funds face the hardest test. Their mandate should ask what strategic capability remains when the investment period ends. Equity ownership without control over knowledge, technology or decision-making can produce exposure without sovereignty.
This intelligence is also a source of opportunity. Fragmentation raises the value of firms that can operate across vendors, jurisdictions and political systems. ASEAN companies can become trusted integration layers; investors can combine regional scale with local legitimacy; sovereign funds can use capital to negotiate technology transfer and domestic participation.
Before the lesson becomes expensive
ASEAN’s AI boom is real, and its scale will continue to attract capital. The decisive question is whether that capital builds capability or simply places more infrastructure inside dependencies controlled elsewhere.
The answer will emerge from the political bargain surrounding each asset: who controls the stack, who absorbs the costs, who captures the value and whether the arrangement survives a shock.
Policymakers must connect investment ambition with resilience and legitimacy. Market actors must recognise that technology choices can become political constraints long before they appear as losses on a balance sheet.
The Gulf showed how quickly compute can move from commercial promise to strategic target.
ASEAN should not wait for its own crisis to discover what its due diligence failed to see.
