Table of Contents
Rajesh Natarajan, chief technology officer at Gorilla Technology, told RCR Wireless News that the company will deploy infrastructure at Yotta’s NM1 facility in Navi Mumbai.
In sum – what to know:
5,000+ GPUs – Gorilla and Yotta plan to install large-scale AI compute infrastructure to support enterprise and public sector workloads in India.
Infra model – Gorilla supplies hardware while Yotta operates and delivers AI services from its Navi Mumbai data center.
AI push – The project supports India’s broader efforts to expand sovereign AI infrastructure and domestic compute capacity.
U.K.-based company Gorilla Technology Group has signed binding agreements with Indian firm Yotta Data Services to deploy AI compute infrastructure in India, including around 640 Nvidia HGX B200 servers equipped with more than 5,000 GPUs. The systems will support AI workloads across enterprise and government use cases.
Under the terms of the agreement, Gorilla will supply the GPU infrastructure through a long-term commercial arrangement, while Yotta will operate the systems at its NM1 data center in Navi Mumbai. The deployment will follow Nvidia reference architecture and support services such as GPU clusters, virtual machines and AI model endpoints.
Gorilla estimates the project could generate more than $500 million in revenue over five years, based on current assumptions. The partnership positions the company as an infrastructure provider within Yotta’s broader platform, which includes hyperscale data centers and AI compute services.
The deployment comes as India expands its domestic AI infrastructure, supported by public investment and national initiatives. Yotta is participating in the IndiaAI Mission and has outlined plans to scale GPU capacity significantly over the coming years. According to data from the Indian government, the Indian AI infra market is expected to reach $17 billion by 2027, with 25% to 35% annual growth, backed by more than $1.1 billion of IndiaAI Mission funding and the deployment of 38,000 GPUs.
“When you move beyond 5,000 GPUs, power architecture becomes a first-order engineering issue, not just a facilities issue. You are dealing with very high rack densities, transient load behavior, power quality, redundancy design, and the ability to sustain those conditions consistently over time,” Rajesh Natarajan, chief technology officer at Gorilla Technology, told RCR Wireless News.
“That is why the choice of operating environment matters. In this case, we are deploying into Yotta’s NM1 facility in Navi Mumbai, which is positioned as a Tier IV data center environment and already built for high-availability infrastructure at serious scale. That gives you the right starting point, but the real work is still in detailed integration, load balancing, failure-domain isolation, and operational discipline,” he added.
