Data Centre Week: Consolidation, Power and Green Rules — 03 July 2026
This week in APAC data centres
- Singtel is reportedly closing five legacy Singapore data centres and shifting to AI-ready capacity (Data Center Dynamics).
- Keppel begins construction of Keppel DC Singapore 9, a ~659,612 sq ft AI-ready campus building, in mid-2026 (W.Media).
- APAC's green data centre market is forecast to grow from US$11.14bn (2024) to over US$63bn by 2030, a 33% CAGR (GlobeNewswire).
- The data centre PDU market is projected to reach US$14.23bn by 2032 at a 19.5% CAGR (MarketsandMarkets).
This week’s theme is consolidation under pressure: Singapore and Johor operators are retiring old halls, breaking ground on denser AI-ready capacity and racing to meet tighter green-power rules. For data centre managers, every trend points back to measuring and controlling power and access at the rack.
Why is Singtel shutting five legacy data centres in Singapore?
Data Center Dynamics
Singtel is reportedly retiring five older Singapore data centres as it shifts workloads to newer, higher-density AI-ready capacity across its Nxera platform.
What this means for your operations: Consolidating onto fewer, denser sites raises the cost of any downtime, so remote out-of-band access and per-rack power visibility become non-negotiable for the racks that remain. Intelligent PDUs and IP-KVM let a leaner team run more critical load without more site visits.
When does construction start on Keppel DC Singapore 9?
W.Media
Keppel will break ground on Keppel DC Singapore 9, the third AI-ready building at its Genting Lane campus, in mid-2026, completing a three-building estate of roughly 659,612 sq ft.
What this means for your operations: New AI-ready halls in land-scarce Singapore are engineered for 100kW-plus racks, which is where legacy 15-40kW power distribution breaks down. Operators fitting out these halls need rack PDUs and DCIM rated for high-density, real-time power and thermal monitoring from day one.
What does the TM Nxera Johor top-out mean for the Singapore-Johor corridor?
Data Center Dynamics
Telekom Malaysia and Singtel's Nxera have topped out the first building of their Iskandar Puteri campus, a 64MW first phase scalable to 200MW on a roughly US$2.28bn investment, targeting commercial launch in 2H 2026.
What this means for your operations: As overflow demand moves across the Causeway, Singapore teams increasingly manage estates that straddle two countries. Consistent out-of-band management and centralised DCIM across both sites keep a single team in control without constant travel.
How fast is the APAC green data centre market growing?
GlobeNewswire
New research puts the APAC green data centre market on course to grow from US$11.14bn in 2024 to over US$63bn by 2030, a 33% CAGR, led by hyperscale build-out and tighter efficiency mandates.
What this means for your operations: Green targets such as Singapore's 1.25 PUE requirement are only auditable if you can measure energy at the rack and outlet. Metered intelligent PDUs and DCIM are the instrumentation that turns a sustainability pledge into reportable numbers.
How big is the data centre PDU market becoming in the AI era?
MarketsandMarkets
Analysts project the data centre PDU market to climb from US$4.89bn in 2026 to US$14.23bn by 2032, a 19.5% CAGR, with intelligent, per-outlet metered units the fastest-growing segment.
What this means for your operations: AI racks pushing 100-140kW leave no headroom for guesswork on power draw, so per-outlet metering and remote switching move from nice-to-have to baseline. Choosing intelligent rack PDUs now avoids a costly retrofit when densities climb again.
Denser racks, green-power mandates and cross-border estates all demand rack-level power metering, DCIM and out-of-band access you can trust. eNOVA supplies Raritan intelligent PDUs and DCIM and enterprise KVM to keep your Singapore and APAC sites measurable and remotely manageable, talk to our team about your next high-density build.
Frequently Asked Questions
Why are Singapore operators consolidating older data centres?
Singapore's land and power constraints, combined with AI workloads that need far higher rack density, make older low-density halls uneconomic. Operators like Singtel are retiring legacy sites and moving load onto newer AI-ready capacity, which concentrates critical workloads onto fewer, denser racks.
What rack power density do new AI-ready data centres in Singapore support?
New AI-ready halls such as Keppel DC Singapore 9 are engineered for racks well above 100kW, compared with 15-40kW in legacy facilities. Supporting that density requires intelligent rack PDUs, high-capacity busways and DCIM that tracks power and temperature per rack in real time.
What is Singapore's PUE requirement for new data centres?
Under the DC-CFA2 allocation, new Singapore data centres must target a Power Usage Effectiveness of 1.25 at full IT load and draw at least 50% of power from green sources. Meeting and proving those targets depends on outlet-level metering and DCIM to capture auditable energy data.
How does intelligent PDU monitoring help with AI workloads?
Intelligent PDUs provide per-outlet power metering, remote switching and environmental sensors, so operators can see exactly how much each high-density AI rack draws. This prevents overloaded circuits, supports capacity planning and enables remote power cycling without a site visit.
Why does out-of-band management matter for the Singapore-Johor corridor?
Workloads increasingly span both Singapore and Johor, so a single team often manages equipment in two countries. Out-of-band IP-KVM and serial console access lets engineers reach servers and switches even when the network is down, avoiding cross-border travel for routine fixes.
Published by
eNOVA Technologies is Singapore's specialist distributor for data centre IT management solutions, representing Adder, Guntermann & Drunck, Raritan, Sunbird, ZPE Systems, and VuWall across Singapore and Southeast Asia. Our technical content is produced with AI assistance and reviewed by our in-house team before publication.
This article was produced with AI assistance and reviewed by the eNOVA Technologies team. All technical claims are verified against manufacturer documentation.
