India’s ambitious AI initiative is set to bolster its computational infrastructure further, with the government planning to add approximately 25,000 graphics processing units (GPUs) to its current commitments of 40,535 GPUs. According to a senior government official, the fourth round of GPU procurement under the IndiaAI Mission is nearing completion.

Subsidies Driving AI Growth

In a bid to make AI research and development more accessible, the government has been subsidizing GPU resources for researchers and startups. This effort has significantly reduced the cost of GPU usage to under $1 per hour, a rate described as the lowest in the world.

sbb-itb-f9e5962

Technical Evaluations Underway

As part of the fourth GPU procurement, technical bid evaluations are currently ongoing. Among the 14 cloud service providers already empanelled, four companies – Yotta Data Services, E2E Networks, NxtGen Datacenter & Cloud Technologies, and Cyfuture India – have confirmed their participation.

One provider has reportedly offered at least 17,000 Nvidia B300 liquid-cooled GPUs, which are highly sought after for AI applications.

Breakdown of Commitments

• E2E Networks: Deploying 1,024 Nvidia B200 GPUs.
• NxtGen Datacenter & Cloud Technologies: Committing 4,096 Nvidia B200 GPUs.
• Cyfuture India: Offering 1,024 GPUs, though the specific model remains unspecified.
• Yotta Data Services: Deploying 8,192 Nvidia B200 GPUs from commitments made in earlier rounds.

Currently, out of the 40,535 GPUs committed under the IndiaAI Mission, 22,787 units (56.21%) have been allocated, according to the IndiaAI Compute Portal.

Deployment Timelines

E2E Networks’ GPUs are expected to go live for end users this month, while Yotta’s deployment is anticipated by March. NxtGen might take a few more months, and Cyfuture’s timeline remains unclear.

Addressing Challenges

Despite the ambitious scale of the project, challenges such as a global memory shortage have impacted deployment timelines. One anonymous source highlighted that only 15,114 GPUs are operational so far, attributing delays to a 650% increase in memory prices and a severe shortage of NAND flash memory. Additionally, the cost of NVMe disks has reportedly surged fivefold.

Perspectives on India’s AI Ambitions

Kesava Reddy, Chief Revenue Officer at E2E Networks, stated, "The IndiaAI Mission’s decision to significantly increase common compute capacity is a powerful signal of India’s commitment to becoming a global AI leader." He also noted that E2E is commissioning India’s first B200 mega cluster with Nvidia Blackwell GPUs, aimed at advancing AI training and inference capabilities for enterprises and startups.

Cyfuture India’s CEO, Anuj Bairathi, described the GPU procurement efforts as "quite aggressive" and highlighted the growing demand for GPUs under the IndiaAI initiative: "There is a great demand for GPUs coming from various users via India AI."

AS Rajagopal, CEO of NxtGen Datacenter & Cloud Technologies, reaffirmed the company’s commitment to deploying 4,094 B200 GPUs.

Scaling Sovereign AI Infrastructure

These developments underscore India’s drive to strengthen its sovereign AI infrastructure and reduce reliance on foreign hyperscalers. As companies fulfill their deployment commitments, the government’s initiative is expected to support India’s aspirations of becoming a global hub for AI innovation.

While the challenges of global supply shortages persist, the IndiaAI Mission continues to push forward, with significant GPU acquisitions reinforcing its commitment to making AI accessible and affordable for researchers and startups.

Read the source