NxtGen has announced the deployment of NVIDIA H200 GPU servers integrated with Akash Systems’ Diamond Cooling® technology, marking what the companies describe as a first-of-its-kind implementation aimed at improving performance and energy efficiency in AI data centre environments.
The deployment is designed to address thermal constraints associated with large-scale AI workloads, particularly in high ambient temperature regions such as India. According to the companies, the diamond-based cooling approach complements existing air and liquid cooling architectures, helping reduce thermal throttling and enabling sustained peak GPU performance. The technology is expected to improve effective compute output and energy efficiency by approximately 15%, translating into higher performance per watt and improved capital utilisation for AI infrastructure.
Originally developed for space-grade satellite systems, the cooling technology leverages the thermal conductivity of diamond materials to dissipate heat more efficiently than traditional copper-based cooling mechanisms. This allows servers to operate at higher ambient temperatures while maintaining consistent performance — a capability increasingly relevant as enterprises expand AI training and inference workloads at scale.
NxtGen said the deployment aligns with its broader strategy to build sovereign AI infrastructure in India, supporting sectors such as government, financial services, healthcare and manufacturing. By increasing compute density and improving power efficiency at the rack level, the company aims to reduce long-term infrastructure costs while delivering higher performance for enterprise AI applications.
Akash Systems noted that advancements in cooling technologies are becoming critical as AI infrastructure scales globally, with energy consumption and thermal management emerging as key constraints for next-generation data centres.
The deployment forms part of NxtGen’s ongoing efforts to expand GPU-based infrastructure across multiple architectures, enabling enterprises to host and scale AI models within sovereign cloud environments that meet local compliance and data residency requirements.