Singapore’s Chip and Data Centre Industries Ramp Up Water Efficiency to Ease Strain on Scarce Supply

Singapore’s Chip and Data Centre Industries Ramp Up Water Efficiency to Ease Strain on Scarce Supply

June 21, 2026

Singapore’s Chip and Data Centre Industries Ramp Up Water Efficiency to Ease Strain on Scarce Supply

Singapore’s semiconductor and data centre sectors, among the country’s largest industrial water users, are accelerating efforts to reduce consumption as demand for chips, cloud computing and artificial intelligence strains limited water resources. The push comes as both industries face mounting pressure to operate sustainably in a city-state where water is a strategic national asset.

Chip manufacturing requires vast quantities of ultra-pure water to prevent contamination across roughly 1,000 production steps, while data centres rely on constant cooling to keep servers running safely. At United Microelectronics Corporation (UMC)’s wafer fabrication plant in Pasir Ris, around 10,000 cubic metres of NEWater are used daily. The water is first treated into ultra-pure quality for chipmaking, then collected, treated again and reused. Thomas Tey, Senior Fab Director at UMC Singapore, said the company focuses on improving efficiency at every stage of production because even small reductions can yield significant water savings.

UMC’s key strategy involves separating wastewater into nearly 20 different streams rather than treating it uniformly. This allows water to be cleaned to varying standards and reused for specific purposes, such as returning to manufacturing processes or supporting cooling towers. The approach enables the plant to recover nearly 70 per cent of the water it consumes, amounting to about 3.9 million cubic metres each year — enough to fill roughly 1,560 Olympic-sized swimming pools. The company is also studying advanced treatment methods, including reverse osmosis and electrodeionisation, to further improve water recovery and produce higher-quality recycled water.

Data centres face a different set of challenges. These facilities generate large amounts of heat from services ranging from video streaming to AI applications, requiring continuous removal. Cloud computing firm OVHcloud has adopted a targeted cooling method that sends chilled water directly to processors rather than cooling entire server rooms with traditional air-conditioning systems. The company said this approach uses less water than conventional methods and improves overall efficiency. Thiru Prakassh, Regional Data Centre Lead for Asia Pacific at OVHcloud, noted that the company uses sensors throughout its facilities to track server activity, equipment loads and cooling performance. The data is analysed using AI tools to predict demand and adjust water use more precisely throughout the day.

OVHcloud estimates that these systems could reduce water consumption by up to 30 per cent and power use by up to 50 per cent. The firm is also exploring the use of local weather data to fine-tune cooling requirements in real time. As more AI systems, cloud services and semiconductor facilities come online, reducing water use is becoming more than an environmental goal — it is increasingly a business necessity. Where every drop counts, finding ways to produce more while using less may prove just as important as the technologies being built for Singapore.

Source: theindependent.sg

Read Also
Singapore’s Chip and Data Centre Industries Ramp Up Water Efficiency to Ease Strain on Scarce Supply
ASEAN Data Centre Investment Surge Nears $30 Billion as Regional Competition Heats Up
Clark County Approves Major Las Vegas Data Center Expansion, Eyes Stricter Future Regulations
Tesla Files Trademark for ‘Megapod’ Modular AI Data Center Hardware
AI Data Center Developers Build On-Site Gas Plants to Bypass the Grid
Union Biometrics Expands Biometric Security Solutions to AI Data Centers in North America and Europe
Google Cloud Launches Local Data Residency for Security Operations in Seoul
Aston Power Secures $20 Million to Scale Private Power Platform for Data Centers
Verse Raises $54M in Series B to Launch AI-Powered Platform That Accelerates Data Center Grid Connections
California Startup Orbital Raises $5M to Build AI Data Centers in Space

Research