Strategy · August 2026
In Asia’s compute and infrastructure markets, macropolitics is no longer commentary. It is an input to underwriting — as material as power price, land title and offtake.
How are geopolitics shaping Asia’s business scene in 2026?
Three forces now sit inside investment memos that used to live in newspaper op-eds:
- Technology competition — export controls, trusted-compute requirements and stack preference (US, Chinese or hybrid) change who can buy what silicon and under which audit regime.
- Sovereign cloud and AI programmes — Malaysia, Indonesia, Singapore and peers are funding or mandating domestic control over sensitive workloads. Capacity must be ownable, auditable and energised — not merely marketed.
- Energy realism — AI campuses compete with industrial and residential demand for megawatts. Grid connection timelines and generation mix decide which projects clear.
Public reporting across the region points to multi-billion hyperscaler commitments into Singapore, Malaysia, Indonesia and Thailand — alongside sovereign AI cloud budgets and a rising M&A preference for sites that already hold power and permits. The opportunity is real. So is the constraint set.
What does this mean for capital formation?
Evergreen infrastructure capital and deeptech ownership capital answer different questions, but both lose when they ignore macropolitics:
- Infrastructure needs contracted offtake logic, land and power pathways, and capital structures that survive AI narrative cycles.
- Deeptech needs controlling ownership and production architecture when sovereignty and confidentiality are the product — not a feature slide.
Allocators who treat “Southeast Asia AI” as a single theme will misprice jurisdiction risk. Platforms that map jurisdiction, stack and power explicitly will earn diligence.
Where is the bridge opportunity?
Between Chinese and Western technology ecosystems that increasingly coexist in ASEAN; between family-office capital concentrated in Singapore and operating assets in neighbouring markets; and between sovereign policy ambition and private platforms that can deliver compliant capacity.
Redwood’s perspective is practical: underwrite the bottleneck, disclose related parties, design liquidity early, and build an operating spine that compounds across holdings.
Explore Platform II — Data Centres →
Frequently asked questions
Is AI demand enough to underwrite a data-centre project?
No. Demand is the narrative. Underwriting turns on power, land, permits, offtake quality and capital structure.
Why do US and Chinese stacks appear in the same ASEAN markets?
Host governments often want investment and capability from multiple ecosystems while imposing localisation, certification and audit conditions. Coexistence is a policy outcome, not a contradiction.
What is Redwood’s underwriting stance?
Prefer contracted and allocated realities over absolute claims; treat sovereignty and energisation as first-order risks; publish perspectives that can be diligenced.