Taiwan's National Center for High-performance Computing (NCHC), part of the National Institutes of Applied Research (NIAR), has received a Silver rating under the Green Building Label program administered by the country's Ministry of the Interior, a designation the center earned for the energy-efficient design of its Cloud Compute Center at the Tainan Science Park.

The facility, officially launched in December 2025, operates as a large-scale AI and high-performance computing hub and an international telecommunications node. It expands Taiwan's national computing, storage, and network service capabilities at a time when AI workloads are pushing power supply, cooling, and energy efficiency to the forefront of computing infrastructure planning.

The certification, issued under Taiwan's EEWH green building evaluation system, validates design choices NCHC made from the outset of the project. The organization incorporated energy conservation, seismic resilience, and high-standard infrastructure requirements into the overall planning and construction of the facility rather than retrofitting efficiency measures after the fact.

The most consequential number for operators watching data center economics is the facility's annual average Power Usage Effectiveness (PUE), which the center reports as below 1.3. PUE measures data center energy efficiency by comparing total facility power consumption to the power delivered to computing equipment. A value closer to 1 means supporting systems such as cooling and power distribution consume proportionally less energy, and a greater share of the electricity purchased actually reaches compute workloads.

Where the Efficiency Comes From

NCHC deployed efficiency measures across five areas of the facility: the building envelope, cooling systems, electrical systems, lighting, and renewable energy generation.

On the building side, the design uses exterior wall insulation and reduced window areas to limit heat gain from Tainan's outdoor environment, reducing the load on cooling plant from the start.

The cooling infrastructure combines magnetic-bearing chillers, hot- and cold-aisle containment configurations, and EC (electronically commutated) fans to improve cooling efficiency. Each of these technologies targets a specific loss point: magnetic bearings reduce mechanical friction in chillers, aisle containment prevents the mixing of supply and return air, and EC fans modulate speed efficiently compared with conventional fan motors.

The electrical systems carry the same design logic. The facility uses high-efficiency transformers, an energy management system, and high-efficiency uninterruptible power supply (UPS) systems to cut conversion and distribution losses between the grid and the machine room.

Lighting is fully LED, controlled by a two-wire lighting control system, and an on-site solar photovoltaic system supplies renewable energy directly to the facility.

Capacity Growth Against an Energy Budget

The Silver rating reflects a positioning choice that matters for national computing infrastructure: NCHC is balancing the expansion of Taiwan's computing capacity against energy efficiency and sustainability commitments. As AI model training and HPC workloads scale, the power and cooling envelope of a facility increasingly determines how much usable compute it can host. A PUE below 1.3 means that for every megawatt delivered to servers, less than 0.3 megawatts of additional power goes to overhead systems — a meaningful margin for a facility that also functions as an international telecommunications node.

The building-level certification and the operational PUE figure together cover two distinct layers of efficiency. The EEWH Silver rating assesses the building and site design as constructed; the PUE figure reflects ongoing operational performance across the facility's mechanical and electrical systems.

NCHC states that it will continue to improve building energy efficiency, optimize PUE, and expand renewable energy use going forward, with the goal of enhancing overall energy performance and supporting more efficient and sustainable data center operations as the center's workloads grow.