Reshaping India’s AI infrastructure for high-temperature data centres and large-scale enterprise workloads, NxtGen’s diamond-cooled NVIDIA servers promise 15% efficiency gains.
NxtGen has announced the deployment of the world’s first NVIDIA H200 GPU servers integrated with Akash Systems’ patented ‘diamond cooling’ technology, marking a significant development in AI data centre infrastructure.
The Bengaluru-based company said the innovation will improve compute density and energy efficiency, while reducing long-term costs for enterprise AI workloads.
Diamond cooling, originally developed for satellite systems, uses diamond’s high thermal conductivity to dissipate heat up to five times faster than copper. In data centres operating in warmer climates, such as India, the technology enables GPUs to sustain peak performance without thermal throttling, even at ambient temperatures of 50°C.
NxtGen reported that the system delivers around 15% higher compute output and improved FLOPs per watt compared with conventional cooling methods.
Performance highlights include a reduction of GPU hotspot temperatures by approximately 5°C, throttle-free operation under heavy AI training loads, and a 15% improvement in energy economics.
The company said these gains translate into enhanced power usage effectiveness (PUE) and greater capital efficiency per rack, critical for large-scale AI deployments.
A.S. Rajgopal, Managing Director and Chief Executive of NxtGen, stated that the deployment strengthens India’s sovereign AI infrastructure by increasing usable compute per watt and per rack. He added that the technology supports structural efficiency, allowing higher performance delivery while lowering infrastructure costs.
Akash Systems’ co-founder and chief executive, Dr Felix Ejeckam, described the partnership as addressing two pressing challenges in AI infrastructure: energy and capital efficiency. He said the improvement in effective GPU compute was “transformative”.
NxtGen’s AI platform integrates GPU architectures from NVIDIA, AMD and Intel, offering enterprises optimised hosting environments with sovereign data residency and compliance.
The company said the deployment reinforces its commitment to building scalable, energy-efficient AI infrastructure as adoption accelerates across sectors, including government, healthcare, manufacturing and financial services.
