A Dual Strategy for AI Self-Reliance
To build a world-class artificial intelligence ecosystem, a country needs two critical ingredients: immense computing power and significant capital. Recognizing this, the Indian government is doubling down on its national AI strategy. Already in motion
through the IndiaAI Mission, which has an outlay of over ₹10,372 crore, the country has been focused on democratizing access to the building blocks of AI. Now, a new proposal is being discussed to create a large-scale fund specifically for AI, which could inject between ₹15,000 and ₹20,000 crore into the ecosystem. This two-pronged approach—providing subsidised access to infrastructure while also offering long-term risk capital—is designed to tackle the biggest hurdles faced by Indian AI startups and researchers. The goal is to move beyond simply using AI to building foundational technologies in India, for India and the world.
The Power of Shared Compute
Modern AI, especially large language models (LLMs), requires an enormous amount of processing power for training and operation. This power comes from Graphics Processing Units (GPUs), expensive hardware that is often beyond the reach of early-stage startups and academic institutions. The IndiaAI Mission is directly addressing this by creating a shared public compute infrastructure. Under the mission, tens of thousands of GPUs have been made available at affordable rates through a common portal. As of mid-2026, the shared capacity had already crossed 45,000 GPUs, with hundreds of projects from startups, government entities, and researchers leveraging this subsidised resource. By pooling these high-performance resources, the government is lowering the barrier to entry, allowing innovators to experiment, build, and test sophisticated AI models without bearing the full, prohibitive cost of the underlying hardware.
A Dedicated Fund for Frontier AI
While shared compute solves the infrastructure problem, building a sustainable AI business requires substantial and patient capital. This is where the proposed National Frontier AI & Compute Fund (NFAICF) comes in. The fund, which is reportedly still under deliberation, would act as an anchor investor, providing long-term capital for developing advanced 'frontier' AI models, as well as for building out the necessary data centres and other strategic infrastructure. The cost of training a single cutting-edge AI model can run into hundreds of millions, if not billions, of dollars, a scale of investment that traditional venture capital may be hesitant to make. A government-backed fund could bridge this gap, ensuring that promising Indian AI companies have the financial runway to compete with global tech giants who possess massive balance sheets.
Fueling a Homegrown Ecosystem
The combination of accessible compute and dedicated funding is designed to create a vibrant, self-sustaining AI ecosystem. The IndiaAI Mission has already shortlisted multiple organisations to develop indigenous foundational models trained on Indian datasets and languages, ensuring that AI tools are culturally and contextually relevant for the population. This includes models like BharatGen, a government-funded large multimodal model that supports 22 Indian languages. By providing both the tools (compute) and the fuel (capital), the government aims to spur a new wave of innovation. This will enable startups not only to build applications on top of existing platforms but also to create their own foundational models, fostering sovereign capability and global competitiveness in the generative AI race. The strategy is a clear move from being an AI consumer to an AI creator.
The Road Ahead and Potential Hurdles
While the vision is ambitious, the proposed fund is still in the consultation phase. Key details regarding its final size, structure, and governance model are yet to be finalised. Establishing a clear and transparent framework for investment will be crucial to ensure that public money effectively catalyses private investment and supports the most promising ventures. Balancing strategic national interests with commercial viability will be a key challenge for the fund's managers. Furthermore, the success of the broader mission will depend not just on counting available GPUs, but on their utilisation, the demand generated, and the ability of Indian AI companies to translate powerful models into sustainable, revenue-generating businesses. The journey from policy to a thriving AI economy is complex, but the direction of intent is clear.
