Marking a major milestone for high-density computing infrastructure in India, digital infrastructure provider Vertiv successfully delivered the nation’s first direct-to-chip liquid cooling deployment in partnership with CapitaLand India. Unveiled on October 13, 2026, the 30-megawatt liquid-cooled data center deployment is specifically engineered to handle the thermal loads generated by next-generation artificial intelligence workloads and high-performance graphics processor clusters. Traditional air-cooling frameworks face severe efficiency bottlenecks when managing power-dense AI racks exceeding 50 kilowatts, prompting data center operators to transition toward direct-to-chip liquid cooling circuits that circulate fluid directly across processors to absorb heat. Infrastructure engineers and sustainability directors highlighted that the direct-liquid architecture drastically slashes data center energy consumption, lowers power usage effectiveness (PUE) ratios, and reduces water usage across enterprise computing facilities. Technology analysts observe that establishing localized high-density liquid cooling infrastructure empowers regional cloud service providers to support heavy generative AI model training locally while meeting stringent corporate environmental and energy-efficiency targets.

Pioneering High-Density Thermal Management for AI Workloads in India

In a major advancement for South Asia’s AI digital infrastructure ecosystem, Vertiv delivered India’s first direct-to-chip high-density liquid cooling (HDLC) system. Deployed at CapitaLand India Trust’s state-of-the-art facility in Airoli, Navi Mumbai, the architecture provides thermal control for high-density server racks operating heavy artificial intelligence, deep learning, and high-performance computing (HPC) workloads.

The project enables 30 MW of AI-ready capacity for a hyperscale client, marking a critical transition from traditional air-based data center cooling to direct liquid thermal capture.

Overview: Vertiv & CapitaLand Liquid Cooling System Parameters

Technical & Operational ParameterOfficial Facility Specifications
Location / VenueCapitaLand India Trust Data Center • Airoli, Navi Mumbai
Cooling ArchitectureDirect-to-Chip High-Density Liquid Cooling (HDLC)
Enabled Capacity30 MW AI-Ready Compute Capacity
Cooling Network SystemClosed-loop Secondary Fluid Network (SFN) Skids
Validation StandardsSFN-Tab balancing & Level 4 (L4) performance testing

Secondary Fluid Networks and Direct-to-Chip Engineering

Traditional air-cooling methods face severe physical limitations when trying to cool modern graphics processing unit (GPU) clusters operating above 40 to 50 kW per rack. Vertiv’s direct-to-chip liquid cooling technology circulating dielectric coolant directly across processing units removes heat at its source, significantly lowering overall power usage effectiveness (PUE) and energy consumption.

Direct-to-Chip Liquid Cooling Flow Pipeline: ------------------------------------------ High-Density GPU Heat Generation ──> Direct-to-Chip Heat Capture ──> Secondary Fluid Network (SFN) Loop ──> Heat Dissipation & Energy Savings

Before reaching full operational status, Vertiv completed loop-balancing validations using custom SFN-Tab kits, followed by comprehensive Level 4 (L4) commissioning tests to certify system performance under maximum simulated AI processing loads.

Supporting India's Growing AI Infrastructure Blueprint

The project highlights the expanding partnership between Vertiv and CapitaLand Investment as demand for hyperscale AI computing capacity accelerates across India.

With major tech enterprises establishing regional AI factories and data processing hubs in Mumbai and South India, high-density liquid cooling technology is expected to become the baseline standard for green, energy-efficient data center development.