Nxera's DC Tuas sets sustainability benchmark for data centers in Singapore

With 58MW capacity, 1.25 PUE, and Singapore's largest direct-to-chip liquid cooling deployment, Nxera's facility shows how to balance AI's insatiable power demands with environmental responsibility

Singapore's data centre moratorium has created an unusual problem: how do you support exploding AI infrastructure demands in a market where new capacity is severely constrained? Nxera's answer, unveiled this week with the opening of DC Tuas, demonstrates that the solution lies not in building more facilities, but in building smarter ones.

DC Tuas represents a fundamental shift in how data centres approach the sustainability challenge. With a power usage effectiveness (PUE) of 1.25, the facility achieves what many operators in tropical climates consider nearly impossible—running AI-ready infrastructure at enterprise-grade efficiency.

For context, the global average PUE for data centres hovers around 1.5-1.6, while many facilities in Southeast Asia struggle to break below 1.4 due to cooling demands in humid, hot climates.

The 58MW facility, which increases Nxera's Singapore capacity to 120MW, was more than 90% committed before launch—a testament to how desperately enterprises and hyperscalers need AI-capable infrastructure that doesn't compromise on environmental standards.

Direct-to-chip cooling: The game-changer for AI workloads

What sets DC Tuas apart is its deployment of direct-to-chip liquid cooling at scale. As Singapore's largest multi-tenanted facility with this capability, DC Tuas addresses the elephant in the room: next-generation AI workloads running on GPUs like NVIDIA's H100 or B200 chips can consume 700W to over 1000W per GPU.

Traditional air cooling simply cannot handle these densities without massive energy penalties.

Direct-to-chip liquid cooling delivers coolant directly to heat-generating components, extracting heat far more efficiently than air-based systems.

This approach slashes both energy consumption and water usage—critical in a city-state like Singapore where both resources are precious. The technology enables DC Tuas to support rack densities that would be impossible with conventional cooling, allowing customers to deploy cutting-edge AI infrastructure without the typical environmental trade-offs.

Bill Chang, CEO of Nxera and Singtel's Digital InfraCo unit, emphasized this point: "The ability to deploy higher-density, compute-intensive AI workloads sustainably is increasingly critical for a strategic market where data centre capacity is constrained."

Smart design for tropical efficiency

Beyond cooling technology, DC Tuas incorporates multiple sustainability features specifically engineered for Singapore's tropical climate. The 120,000-square-foot, eight-storey facility combines smart thermal features with high-efficiency electrical systems, supplemented by solar power generation, rainwater harvesting, condensate reuse, and blowdown recovery systems.

These aren't marketing additions—they're operational necessities that directly impact the facility's ability to maintain its 1.25 PUE rating. The facility's Green Mark Platinum certification from Singapore's Building and Construction Authority and Infocomm Media Development Authority validates this integrated approach to sustainability.

Hyperconnectivity as competitive advantage

DC Tuas also stands out as Singapore's only hyperconnected data centre integrated with a cable landing station. This gives customers direct access to both international and domestic networks, reducing latency and improving reliability—crucial factors for AI training and inference workloads where milliseconds matter.

As a carrier-neutral facility, DC Tuas offers flexibility that many hyperscaler-focused facilities cannot match, making it particularly attractive for enterprises navigating multi-cloud strategies or requiring specific connectivity arrangements for AI deployments.

Regional expansion ahead

With DC Tuas operational, Nxera is pushing ahead with additional AI-ready capacity in Batam and Johor, scheduled for the second half of 2026. The company expects its operational and pipeline capacity to more than double from 200MW in 2026 to over 400MW mid-term.

This regional strategy acknowledges a reality: Singapore's data centre constraints aren't going away, but demand for AI infrastructure across Southeast Asia continues to accelerate. By developing facilities with similar sustainability standards across the region, Nxera is positioning itself as a partner for organizations that need AI capability without sacrificing environmental commitments.

DC Tuas proves that AI infrastructure and sustainability aren't mutually exclusive—but achieving both requires sophisticated engineering, significant capital investment, and a willingness to deploy advanced technologies at scale. For organizations betting on AI while facing pressure to reduce their environmental footprint, that's exactly what they need.