Entrepreneurship studio logo

Entrepreneurship Studio

AI•

Startup•

Cloud

NeevCloud Launches with 40,000 GPUs, Revolutionizing AI Cloud Services for SMEs in India

This image features a graphic design banner for "neevcloud". The design includes a central logo that depicts a stylized cloud with circuit-like connections, emphasizing digital or cloud-based technology. The name "neevcloud" appears beside the logo. The background shows a visual representation of a network topology with servers connected by lines, suggesting data interchange or a cloud network infrastructure. A human figure is interacting with one of the servers, indicating user engagement with the technology. The overall color scheme is blue and white, reinforcing a tech-oriented, clean aesthetic.
By
Anurag Trivedi

Narendra Sen, CEO of NeevCloud, aims to build an AI cloud infrastructure customized for Indian users, with a plan to incorporate 40,000 GPUs by 2026. This infrastructure is intended to aid Indian businesses with AI training, inference, and other related tasks.

Yotta, a notable AI cloud provider, recently became the first NVIDIA Partner Network (NPN) cloud partner in India and has achieved Elite Partner status globally. Yotta plans to establish an AI cloud infrastructure featuring 32,768 GPUs by the end of 2025. NeevCloud is determined to surpass this.

Additionally, NeevCloud is preparing to introduce an AI inferencing platform that will support open-source models such as the Llama 3 series, Mistral models, and DBRX by Databricks.

“Later this month, we aim to launch the DBRX due to high demand for this model in India. Afterward, we will release text-to-image models from Stability AI. Our approach is methodical to ensure smooth operations. We prioritize user experience and may expand our offerings if all goes as planned. Concurrently, we are enhancing our capabilities,” Sen explained to AIM.

Inferencing involves using a trained model on new data to generate predictions or decisions. To date, about 3,500 developers have registered to use NeevCloud’s inferencing platform.

The company will soon launch the beta version of this platform and will offer free tokens to attract developers to test it. The platform will utilize around 100 NVIDIA GPUs, including the NVIDIA H100s, A100, and L40 models. NeevCloud also plans to introduce AMD’s M1300X.

In terms of AI infrastructure, Sen revealed that NeevCloud has ordered 8,000 NVIDIA GPUs from HP Enterprise (HPE), expected later this year.

NeevCloud is set to directly compete with Yotta in the AI-as-a-service sector. Meanwhile, E2E Network, another player in the Indian market, provides NVIDIA’s H100 GPU and NVIDIA A100 Tensor Core GPUs at competitive rates.
Sen founded NeevCloud in 2023, building on his experience with Rackbank Datacenters, a company he has been running for many years in Indore. Located in Crystal IT Park, Indore, this 35,000 sq ft data center can accommodate over 32,000 servers.

The company has developed a liquid immersion cooling technology, named Varuna, to efficiently cool high-density computing hardware used in AI, machine learning, and high-performance computing (HPC). This technology involves submerging servers and other IT hardware in a dielectric, non-conductive coolant liquid, which directly transfers heat from the components to the liquid.