Given India’s diverse nature, the chipmaker has developed an updated growth strategy that extends beyond solely relying on GPUs.
<div class="media-with-label__label">
Credit: ChatGPT </div>
</figure>
</div>
</div>
</div>
</div>For several years, Nvidia has championed sovereign AI initiatives, yet it recognizes that India’s significant cultural and economic variations necessitate a distinct approach.
The expansion of the chipmaker in India will not be driven by large shipments of GPUs, unlike in the United States. Instead, the company intends to prioritize software development initially, with computing power considerations to follow. It is committed to a data-first philosophy, implementing a localized AI strategy tailored to India’s diverse demographics, cultures, and linguistic landscape.
Nvidia is also appealing to India’s vast developer community by emphasizing its open-source offerings and by formulating a local AI strategy that incorporates smaller language models, optimized for the region’s energy and data center infrastructure.
Vishal Dhupar, Nvidia’s managing director in South Asia, stated, “When the data generated here is integrated into an Indian stack, it creates population-scale solutions that citizens can leverage to address numerous significant challenges India faces, whether in education, healthcare, or mobility.”
The Indian government has explicitly noted its unique approach to AI consumption compared to nations like the U.S. For instance, farmers do not require extensive large language models (LLMs) and can effectively resolve issues with specialized, smaller models operating on low-power chips.
Nvidia’s strategy for India’s sovereign IT initiatives includes extending open models such as Nemotrol and various development tools to Indian developers. Dhupar commented, “Open models have the potential to significantly enhance sovereign AI, enabling our developers to embed their distinct language, traditions, and culture into all their endeavors.”
Last year, Nvidia enabled developers to create CUDA programs using Python, a language widely favored in India. This coding capability indirectly boosts the sales of Nvidia GPUs, as developers utilizing CUDA code require these graphics chips to process AI workloads.
Blackwell-class GPUs from the company are being deployed in Indian data centers, despite the nation’s infrastructure not yet matching the maturity of the U.S. However, India is well-positioned for an AI infrastructure environment that moves beyond traditional GPUs, especially considering power and resource limitations, which is why it is proactively preparing for a future with low-power chips and edge processing.
During the World Economic Forum last month, India’s IT minister, Ashwini Vaishnaw, put forth a proposal to directly adopt lower power consumption chips capable of running smaller language models.
Vaishnaw’s reasoning is that these more compact models can address 95% of Indian users’ needs at a significantly reduced cost. Furthermore, the nation aims to avoid the pitfalls of an AI bubble, where the collapse of an overvalued AI enterprise could negatively impact the economy.
Nvidia’s open-source Nemotron models have contributed to the creation of several localized models, including the 17-billion-parameter BharatGen, which supports applications in public services, agriculture, security, and cultural preservation, according to the company.
Nvidia has also provided AI contributions to India’s central digital payment system, the United Payments Interface (UPI), which has garnered accolades for its speed and efficiency. In a statement, Nvidia noted that NPCI, the entity managing UPI, “is exploring training FiMi, a financial model for India, using the Nvidia Nemotron 3 Nano model and its own datasets.”
Furthermore, Nvidia’s AI technology is linked to the 8-billion parameter Chariot model, a multilingual communication platform, and Sarvam.ai, an AI platform designed for multimodal applications specific to India.
Beyond these efforts, the company collaborates closely with India to cultivate the next wave of developers and to foster emerging startups. For example, it is partnering with the Indian government’s Anusandhan National Research Foundation (ANRF) to accelerate advanced AI research across the nation’s leading academic institutions.
The company plans to grant ANRF recipient institutions complimentary access to Nvidia AI Enterprise software and provide specialized technical guidance through the Nvidia AI Technology Center. An Nvidia spokeswoman affirmed, “The collaboration will also encompass AI bootcamps, workshops, and hackathons aimed at bolstering India’s AI research ecosystem.”
Yotta is integrating 20,000 Nvidia Blackwell Ultra GPUs into its Shakti cloud, aligning with India’s national sovereign AI infrastructure objectives. Larsen & Toubro and E2E Networks have also declared intentions to establish new data centers utilizing Nvidia GPUs.
Nvidia has previously discussed its sovereign AI initiatives in Europe, which involve establishing GPU data centers and forging alliances with telecommunications, software, and industrial firms. Gartner recently projected that European investments in sovereign IaaS could reach $12.6 billion by 2026, a significant increase from $6.9 billion in 2025.
