By Vipul Prakash
The global technology circle has a popular narrative regarding India and AI that speaks about the country’s large talent funnel. It is almost reflexive, and the narrative further goes towards how India executes well, even scales unprecedentedly; however, the core of the belief remains that the country does not innovate in the AI aspect. However, the narrative is absolutely outdated, and in some cases, it is becoming critically flawed at best.
The global industry constituents and other stakeholders, such as VCs, often talk about DeepSeek and its breakthrough in early 2025, highlighting that China is moving fast. They believe that the US is moving even faster in this regard, but in India, the ball is yet to get rolling. This belief remains fundamentally wrong, and questioning whether India is working towards developing innovative AI technologies misses the real picture of what is happening behind the scenes.
The Shift No One’s Talking About
What distinguishes India in the AI race is the country’s ability to think differently. Unlike the popular belief that India is trailing in AI development, the reality remains that India is leading the change, however, in ways that people do not expect. By taking a look at the fundamental level, we soon understand that India has 12 lakhs of AI professionals already working today. They are not researchers working inside a laboratory, but practitioners who are building real-life products. Prominent research reports show that India has the world’s highest AI skills penetration rate, while our universities are graduating 22,000+ AI specialists every year. What this means is that by 2027, India will have more than 20 lakh highly equipped AI professionals.
What matters more is that 74% of Indian GenAI startups launched in 2021-22 are AI-native. These are not companies that added AI to existing products, but firms that have built from the ground up to solve problems with intelligence. There are firms out there in India that are solving the multilingual problem at volume, where the AI models are able to handle 10 million queries every month across 12 Indian languages. What distinguishes this is that no company from anywhere else in the world has cracked the linguistic hurdle. There are firms out there that are building causal AI for healthcare and have already garnered accuracy in the top 25%, in contrast to doctors. India also houses firms that are innovating in semiconductor design with the help of AI.
The originality and innovation in these companies are not incremental improvements, as numerous frontier breakthroughs are taking place in India, by Indians, who are more interested in solving global problems than in following a pre-determined route of development.
The Compute Story: From Deficit to Leadership
While India is doing its fair share of developments, not everything is solvable. India has a computing problem, as we control only 2% of the world’s AI compute, despite having 18% of the global population. However, an unprecedented change has been witnessed in recent months, as the Government of India has made it a national priority.
With the IndiaAI mission, 8 major organisations have come together, like Tech Mahindra, IIT Bombay, Fractal Analytics and others, with a shared vision of developing indigenous foundation models. For instance, IIT Bombay is building a trillion-parameter model, whereas the PARAM Rudra initiative is deploying a 25-petaflop supercomputer. This also highlights the distinguishing factor in India’s approach that is not working towards matching the US or China in raw scale, but building smart compute infrastructure that is integrated with strong points like renewable energy, human talent and problem diversity. The end goal of this is to create a national AI compute grid where researchers, startups and enterprises are able to access world-class infrastructure at subsidised rates.
Causal AI: The Frontier Where India Leads
Where India’s real advantage lies is the world’s profound shift from predictive AI to casual AI. This is because understanding what will happen is no longer enough, as users are more interested in the why and what should be done about it. This notion is also agreed by the key industry constituents, such as Judea Pearl, the Turing Award-winning father of causal inference, who calls causality “the biggest roadblock to human-level intelligence.” Yoshua Bengio, a pioneer of artificial neural networks and deep learning, agrees, as he feels the next frontier of AI progress runs through causality.
Additionally, global tech giants are investing heavily in Casual AI as they have already identified the market gap. What sets India apart with an advantage is that the country has the talent to lead here and the unique problems that make causal AI essential. For instance, users may ask about why Tuberculosis is a major concern in a region, but not in another. Users may also want to understand why crops fail in one season but thrive in another, despite similar conditions, or why some policies work at scale while others do not.
The mutual point in these queries is that all of these are concerns for Indians, and solving them casually with explainable, trustworthy intelligence will create solutions that matter to the global population.
The Sovereignty Play: Building for India, Scaling to the World
India is not competing on the USA’s or China’s terms, but building on its own. The focus in India is not to create the second Silicon Valley or imitate Beijing, but to build the country’s name itself with a tech ecosystem that focuses on Indian problems, is built with Indian talents and is scaled with the country’s ambition to export globally. Industry estimates suggest that a significant 79% of Indian GenAI startups prefer building products in-house, and are not waiting for permission from global tech giants.
The Indian government is introducing numerous initiatives to support this aspect, such as the National Deep Tech Startup Policy, which has increased funding for frontier tech. This highlights signalling that the strategy has shifted from execution to invention.
Why This Matters for the Next Decade
A decade from now, companies that will matter are being built now, and they are not focusing on what works in Silicon Valley. In contrast, they are emphasising solving problems that Silicon Valley doesn’t have the context to see.
For instance, a healthcare AI that has been trained using Indian medical data is capable of understanding Indian genetics, Indian disease patterns, and Indian healthcare infrastructure. Similarly, an agricultural AI that understands monsoon variability, soil diversity, and farmer economics across 700,000 villages, or a governance intelligence that optimises for scale, diversity, and resource constraints at a level no Western system ever needed to. These systems essentially represent global problems, and India has the talent, problems and increasingly, the infrastructure to lead in this aspect. This is changing the narrative from the grassroots, and the question is whether Indians will have the conviction to stick with their frontier innovation.
The real shift is coming in terms of the gradually changing narrative about India in AI, as it leads to a more comprehensive way. By the next decade, India will become a model case in showcasing frontier AI to the world, especially when it’s built for scale, diversity, and for problems that actually matter.
(The author is the Founder & CEO of FireAI)
Disclaimer: The opinions, beliefs, and views expressed by the various authors and forum participants on this website are personal and do not reflect the opinions, beliefs, and views of ABP Network Pvt. Ltd.
