IndiaAI Mission Expands Compute Power, Foundation Model Development In Major Push Towards AI Self-Reliance

New Delhi: The IndiaAI Mission, led by the Ministry of Electronics and Information Technology (MeitY), empanelled yet another 16,000 Graphics Processing Units (GPUs) which was the second round of GPU onboarding. This latest round of onboarding a total of GPUs in the mission amounts to a remarkable total of 34,333 — making its current GPU capability more than three times the mission's initial target.

The IndiaAI Mission was launched in March 2024 with an outlay budget of ₹10,372 crore and described as a sophisticated, multi-scope mission designed to create equitable access to AI technology and advance Indian innovators, startups, and industries. The second GPU onboarding represents a giant leap towards achieving the mission’s pursuit of developing a robust, equitable, and self-sufficient AI ecosystem for the country.

Transforming the Compute Backbone

The IndiaAI Compute Platform, the mission’s technological backbone, has expanded as more Cloud Service Providers (CSPs) have joined its list of providers. The latest round of GPUs was derived from seven new CSPs, Cyfuture India, Ishan Infotech, Netmagic, Sify, Vensysco, Locuz, and Yotta. Each of these firms has provided multiple advanced GPUs as well as B200s. These firms are providing a range of advanced GPUs and AI hardware, from Nvidia’s H100s, H200s, and B200s to AMD’s MI300X and MI325X series and Intel’s Gaudi accelerators.

These additions augment the existing consortium of providers such as Jio Platforms, NxtGen Data Centre, CtrlS Datacenters, Tata Communications, and E2E Networks. Through this federated, cloud-based structure, compute power is being made accessible to startups, academic researchers, government entities, and enterprises at scale via the IndiaAI Compute Portal.

“As we complete this second round of GPU empanelment, the third round is also in the pipeline, and will soon add more to the pool,” Union Minister for Electronics and IT, Ashwini Vaishnaw, announced during the “IndiaAI – Make AI in India, Make AI Work for India” event held in New Delhi. “What we once thought was an ambitious goal, 10,000 GPUs, has now grown to over 34,000. This is the essence of our commitment to making India a global AI powerhouse.”

The Philosophy: Technology for All

Minister Vaishnaw underlined the foundational philosophy behind the mission: the democratization of technology. “Technology should not be in the hands of a few. A large section of society should have access, should be able to develop new solutions and gain better opportunities,” he said.

The minister likened the holistic approach of the IndiaAI Mission to the government’s strategy in the semiconductor space, covering the entire value chain, from fabrication and OSAT to design and research. He emphasized the importance of creating a shared infrastructure that allows equitable access to compute resources and ensures that innovation is not stifled by barriers to entry.

“We are building what I call the ‘common compute layer’—a shared national resource that can be accessed by anyone in the ecosystem,” Vaishnaw explained. “Whether it is a PhD student in a Tier-2 city or a startup founder working on the next big AI breakthrough, they should not be constrained by lack of infrastructure.”

Accelerating India’s Own Foundation Models

Beyond compute, the IndiaAI Mission is focused on creating indigenous foundation models, large AI models trained on India-specific datasets and tailored for local contexts and languages.

The government has now shortlisted three more startups to develop India’s own foundation models. These include Soket AI, which is building India’s first open-source 120-billion parameter model optimized for linguistic diversity with applications in defence, healthcare, and education; and Gnani AI, which is creating a 14-billion parameter multilingual Voice AI model capable of real-time speech processing and reasoning. These join Sarvam AI, announced in April as the first startup under this initiative.

“These teams have a clear mandate,” said Vaishnaw. “They must strive to become among the top five globally in their respective domains. One of them has even set a goal of being among the top two. That’s the spirit we want to foster. These are not just technical milestones — they are strategic, nation-building goals.”

The effort is already drawing global Indian talent back to the country. “What was once brain drain should now become reverse brain drain,” Vaishnaw added, noting that many young engineers working with these teams are considering relocating to India to be part of this mission.

AI Kosh and the Data Revolution

A foundational component of building capable AI systems is access to quality data. Under IndiaAI, the government has launched “AI Kosh,” a centralized repository of datasets for AI development. As of now, 367 datasets have been uploaded and are accessible to developers and researchers.

This represents a critical step toward building ethical, context-aware models that reflect the country’s linguistic, cultural, and geographic diversity. “The pace at which we’ve been able to load data, while ensuring transparency and quality, speaks volumes about our readiness,” said Vaishnaw.

Talent Development and Responsible AI

Another crucial pillar of the mission is skilling. To this end, the government is rolling out AI education and training initiatives across 240 universities. Plans are underway to fund PhD research in core AI disciplines and to establish a layered skilling ecosystem, from basic AI literacy to advanced research programs.

Equally important is the development of safeguards for responsible AI. Ethical standards, regulatory frameworks, and AI risk mitigation tools are being integrated into the mission to ensure that technological progress aligns with societal values and constitutional principles.

An Ecosystem Under Construction

The IndiaAI Mission is not a piecemeal effort. It is a coordinated attempt to build a resilient AI value chain that includes compute infrastructure, foundational models, datasets, application development, talent pipelines, startup funding, and governance.

More challenges and competitions are in the pipeline to foster innovation across domains. Winners of existing AI innovation challenges will be announced in the coming weeks, giving further impetus to AI-driven problem-solving in agriculture, climate, education, and public health.

“The entire ecosystem is now getting built,” Vaishnaw said. “We are not just reacting to global AI trends, we are creating our own narrative, one that is inclusive, competitive, and rooted in Indian realities.”

For a country of 1.4 billion people, this mission is not just about technological progress. It is about nation-building in the digital era. As Minister Vaishnaw aptly summarized, “This technological transition will either disrupt or empower. With the right ecosystem in place, we can ensure that India is empowered, and leads.