SEA Daily Brief: Singapore captured 99 percent of the region's AI infrastructure funding, Microsoft's Indonesia Azure region turns one, AI takes a third of Singapore's shrinking venture pool

Singapore Captured Roughly 99 Percent of Southeast Asia's US$1.2 Billion in AI Infrastructure Funding, Tracxn Finds

Southeast Asia attracted about US$1.2 billion in AI infrastructure funding between 2019 and 2026, and Singapore accounted for roughly 99 percent of it, according to Tracxn data reported by TechNode Global on May 28. Malaysia recorded US$1.5 million of early-stage activity over the period, while Indonesia and Thailand showed no disclosed AI infrastructure funding at all. Tracxn attributes Singapore's dominance to its established financial infrastructure, deep talent pool and pro-innovation regulatory environment, which have made the city-state the default domicile for AI infrastructure companies across the region.

The dataset also shows the market is both early-stage and cooling. Early-stage rounds accounted for over 95 percent of all disclosed capital across 2023 to 2025, with no late-stage transactions recorded in any year. Deployment fell sharply year-on-year: 2024 saw US$608 million of US$614 million total come from early-stage rounds, while 2025 recorded just US$320 million in total — US$312 million of it across four early-stage rounds — a 48 percent decline. Seed funding edged up from US$4 million in 2023 to US$7.64 million in 2025.

Microsoft's Indonesia Central Azure Region Marks Its First Anniversary as AI Compute Localises In-Country

Microsoft's Indonesia Central Azure region passed its first anniversary on May 29, the cornerstone of the company's US$1.7 billion, four-year commitment to cloud and AI infrastructure in Indonesia first announced in 2024. The region was built with three availability zones and clusters of NVIDIA H100 and AMD MI300X GPUs, and for the first time lets Indonesian organisations keep data within national borders on the Microsoft cloud — including in-country residency for Azure OpenAI Service, Microsoft 365 Copilot, GitHub Copilot and Microsoft Fabric.

The operational details point to a deliberate AI-workload design. Liquid cooling brings the region's estimated power usage effectiveness to about 1.12, well below the industry average near 1.5, and a 250-megawatt solar purchase agreement that came online in early 2026 puts the region on a path to 100 percent renewable energy by 2028. Microsoft also says it has trained more than 3 million Indonesians in digital skills, exceeding its 2.5 million target, and cites enterprise migrations such as BNI Finance moving more than 200 terabytes of data and over 100 application services onto the region within two months.

AI Absorbed Nearly a Third of Singapore's Venture Funding in 2025 Even as the Total Pool Shrank 34 Percent

AI startups in Singapore raised approximately S$1.8 billion (US$1.4 billion) in 2025, nearly a third of the country's total venture funding value, according to DealStreetAsia data published on May 28. AI accounted for 42.8 percent of the 472 deals recorded over the year, and the sector's share of total deal value nearly doubled to 30 percent, with AI deal value up 28 percent year-on-year.

That concentration occurred against a contracting backdrop. Total venture funding in Singapore fell to S$5.9 billion (US$4.6 billion) in 2025, a 34 percent decline from the prior year. Singapore nonetheless accounted for 74 percent of all fintech funding raised in the region last year, reinforcing its position as Southeast Asia's dominant capital hub even as the absolute size of that hub shrinks.

Eyes on the Day Ahead

Over the next few days, watch Bank Indonesia's halved US$25,000 FX-purchase documentation threshold take effect at the start of June and whether the rupiah re-tests its record low near 17,706 per US dollar. In Malaysia, the near-term watch items remain follow-through on the Securities Commission's May 20 revised digital asset exchange guidelines and Bank Negara's three Digital Asset Innovation Hub stablecoin and tokenised-deposit pilots. Regionally, expect more first-half data centre and AI-region capacity announcements from hyperscalers in Malaysia and Indonesia as the Singapore-versus-neighbours split between AI capital and AI compute comes into sharper focus.