Yotta says it is investing $2B to deploy Nvidia's Blackwell B300 GPUs at its data center campus in Noida, India to create one of Asia's largest AI superclusters
Himanshi Lohchab /The Economic Times:
Context & Ripple Effects
Yotta’s Noida plan extends its earlier effort to acquire roughly 20,000 Nvidia H100 chips for AI computing services in India, shifting the company’s stated ambition from an initial GPU buildout to a much larger Blackwell-era campus commitment. Yotta’s planned H100 procurement established the company as an early local buyer of Nvidia capacity.
The announcement lands as India’s AI infrastructure market is being built on both supply and demand: Nvidia is also working with Indian venture firms to support AI startups, while Reliance has separately outlined a major Nvidia-powered data-center buildout. Nvidia’s startup-funding partnerships in India make local compute availability more consequential.
First-order effects
- Yotta commits capital toward deploying Nvidia Blackwell B300 GPUs at Noida, making Nvidia the named accelerator supplier for a substantial new Indian AI-compute installation.
- Indian companies seeking GPU capacity gain a prospective domestic large-scale supplier, while Yotta takes on the execution task of turning hardware procurement into usable data-center capacity.
Second-order effects
- The buildout raises competitive pressure on other Indian data-center operators and GPU-cloud providers to secure accelerators, power, sites, and enterprise AI customers; Reliance’s planned Nvidia-backed data-center expansion is a directly relevant competing buildout.
- A larger installed base of local compute can reinforce Nvidia’s ties to India’s AI startup ecosystem, though actual utilization will depend on Yotta’s delivery timetable and customer demand.
Third-order effects
- If comparable projects are completed, India’s AI market could shift from importing cloud capacity toward competing domestic AI-infrastructure platforms, with access to power, financing, and GPU supply becoming core differentiators.
- The project is another example of AI infrastructure being financed and planned as long-lived regional capacity rather than only as on-demand cloud inventory; whether that model holds depends on sustained demand for the resulting compute.
The trend: AI infrastructure is becoming a regional capacity race, as operators pair Nvidia’s newest GPUs with large campus investments to serve local AI demand.