Without urgent course correction, increased R&D spending, and bolder innovation strategies, the country risks being left behind in a race that will define the future of economies, geopolitics and national security
A press release issued by the Ministry of Electronics and Information Technology (MeitY) in March 2025 claimed that India was on the cusp of an AI revolution, having allocated Rs 10,300 crore, or US$ 1.2 billion, over the next five years to develop a high-end common computing facility equipped with 18,693 GPUs, making it one of the most extensive AI compute infrastructures globally.
This can give India substantial computing power, given that the Chinese open-source AI model DeepSeek uses only 2000 GPUs and operates on 25000 GPUs. However, the promise to make India a global AI giant is unlikely to be fulfilled anytime soon.
A reality check might help the mandarins at MeitY to wake up from their make-believe world of AI greatness and realign their policies with a wartime urgency, because the fact remains that India is a laggard in the global AI race in which the US and China are lightyears ahead of us.It is not that India lacks talent — it has a 1.2 million strong talent pool of AI/ML professionals — its educational ecosystem turns out over 22,000 AI specialists every year. It boasts a vibrant tech ecosystem that facilitates AI applications.
AI systems require massive amounts of data; indeed, AI is only as good as the data that fuels it, and India has one of the biggest data ecosystems in the world — its 900 million internet users generate huge amounts of data. But our AI innovations are still stuck at a nascent stage with huge resource constraints and lag way behind the frontrunners in scale, innovation and research. With just 33 supercomputers, none of them dedicated to AI, India holds only 2 per cent of the global AI computing power, while the USA and China have more than a hundred supercomputers each, many dedicated to AI use.
Among the fastest 500 supercomputers in the world, India has only 6, with our top-ranked machine, AIRAWAT, placed at only 136th position in terms of speed and performance.The result is that OpenAI’s GPT-4, Anthropic’s Claude, and Google’s Gemini dominate global benchmarks, and US companies continue to release large-scale, general-purpose foundation models trained on vast amounts of data that can be adapted to a wide range of downstream tasks like translation, summarisation, question answering, image generation, etc.
They have access to unmatched compute, massive datasets, and cutting-edge chip designs. Even Chinese models like Baidu’s Ernie 4.0 or Alibaba’s Tongyi Qianwen rival Western large language models (LLMs) in Chinese language performance.
China’s DeepSeek, offering cheaper, open-source, and quickly iterated variations, has jolted the US tech giants and greatly accelerated the race to the top, while in India, our efforts remain limited to tackling diverse local challenges in diagnostics, logistics, finance and governance, with AI startups building mostly language and domain-specific solutions.
India’s LLMs like Sarvam, Krutrim, and BharatGPT focus on low-resource language inclusion, digital public infrastructure like Aadhaar, ONDC, UPI, and lower-cost compute and localisation, lagging in scale, benchmarks, and ecosystem maturity compared to the USA or China.
While the USA or China can leverage their AI advantages for economic expansion, global influence, space race and military superiority, we will remain trapped in a low-level equilibrium at great peril to our economy and national security.Lack of resources remains the major constraint.
According to the 2025 AI-Index published by Stanford University, global private venture capital funding for AI startups reached $132 billion in 2024, of which the USA attracted $109 billion of that, followed by China at $9.3 billion and the UK at $4.5 billion. India stood at the 12th spot with only $1.2 billion in private AI investments. In total, from 2013 to 2024, India’s cumulative private AI investment was $11.29 billion — still less than what the US invested in a single year in 2024.
One reason behind the low investment is the risk-averse start-up mindset of domestic VC funds, which prefers quick solutions and returns over original innovation, which may carry a high risk of failure but also delivers much higher long-range returns. The other is the government’s inability or unwillingness to increase R&D expenditure, which remains among the lowest in the world, at only 0.6 per cent of GDP, compared to 2.7 per cent for China and OECD countries, and over 3.4 per cent for the USA. Government Initiatives like IndiaAI Mission (2024) or our AI Strategy lack the boldness, scale and determination backed by resource commitment needed to compete with the frontrunners, while we continue to pour billions of dollars’ worth of doles into highly wasteful and poorly targeted welfare schemes like food security. AI has already created a huge digital divide in the world, in which only a few countries today are appropriating all the benefits.
The divide is influencing not only geopolitics but also creating huge dependencies. We run the risk of getting excluded from a technology race that is shaping and defining our world if we do not urgently change our approach to become a contender rather than remaining a laggard forever.
(The writer is former Director General at the Office of CAG of India and is currently a Professor at the Arun Jaitley National Institute of Financial Management)