Yotta Commits $2 Billion to Nvidia Blackwell Supercluster in India

The landscape of global AI infrastructure is shifting rapidly toward the Asia-Pacific region, with India emerging as a primary battleground for high-performance computing. Yotta Data Services’ decision to commit $2 billion (approximately Rs 16,600 crore) to Nvidia’s Blackwell Ultra GPUs represents one of the single largest private investments in AI hardware outside of the United States and China. This move is not merely a procurement exercise; it is a strategic alignment that integrates India into Nvidia’s global supply chain for AI training and inference services. By deploying 10,300 Blackwell B300 GPUs at its Noida hyperscale campus, Yotta is effectively leapfrogging current generation hardware to provide the most advanced compute capabilities available on the market.

Central to this partnership is the establishment of Asia’s first DGX Cloud supercluster. Under a four-year agreement valued at $1 billion, Nvidia will utilize nearly half of Yotta’s new GPU capacity to serve its own global customers. This co-location strategy allows Nvidia to expand its cloud-based AI training services while providing Yotta with a guaranteed revenue stream and the prestige of hosting a primary node for the world’s most valuable semiconductor company. For Yotta, the remaining capacity will be split between serving domestic APAC customers and supporting India’s National AI Mission, which seeks to democratize access to compute for local startups and research institutions.

“

Yotta Data Services’ decision to commit $2 billion (approximately Rs 16,600 crore) to Nvidia’s Blackwell Ultra GPUs represents one of the single largest private investments in AI hardware outside of the United States and China.

This investment comes at a time of intensifying competition within the Indian technology sector. While Yotta is doubling down on Nvidia architecture, other domestic giants are diversifying their hardware stacks. Tata Consultancy Services (TCS) and AMD recently expanded their partnership to challenge Nvidia’s dominance in the region, signaling a broader market push for multi-vendor AI environments. Furthermore, the Adani Group’s massive $100 billion commitment to energy and digital infrastructure suggests that the race to build the 'backbone' of the AI economy is only just beginning. Yotta’s CEO, Sunil Gupta, has positioned the company as an early mover, aiming to have the Noida facility operational by August 2026 to capture the first wave of Blackwell-driven demand.

However, the rapid expansion of AI infrastructure in India is meeting friction at the regulatory level. While the government is eager to support hardware deployment, its stance on digital safety and content moderation is creating operational hurdles for global platforms. Meta’s Vice President of Policy, Rob Sherman, recently flagged significant challenges regarding India’s proposed 'three-hour' deadline for removing offensive content. This regulatory tension highlights a growing divide: India is becoming a world-class destination for AI compute and data centers, yet it remains a complex environment for the software and social platforms that utilize that very infrastructure. For companies like Meta, the cost of compliance may soon rival the cost of the hardware itself.

Looking forward, Yotta’s $2 billion bet will serve as a litmus test for the viability of high-end compute exports from India. If the Noida supercluster successfully serves global APAC customers as planned, it will validate India’s potential to move beyond software services and into the realm of high-value infrastructure hosting. The success of this venture will depend on Yotta’s ability to maintain the rigorous power and cooling requirements of the Blackwell architecture while navigating a domestic landscape where policy and technology are evolving at different speeds. As the August 2026 go-live date approaches, the industry will be watching closely to see if India can truly become the 'AI factory' for the Eastern Hemisphere.