One might think that the spark that ignited this revolution was ChatGPT in November 2022, but OpenAI’s runaway hit just provided the final kindling to a slow fire that was burning since 1955 when AI (artificial intelligence) was envisaged first in the now famous Dartmouth Conference. The pace of innovation in the last two years since ChatGPT is unmatched—with many powerful LLMs (Large Language Models) from Big Tech firms and start-ups beating humans in some faculties, a Cambrian Explosion of applications running on top of these foundation models bringing AI to workplaces and businesses, and recently the Chinese model DeepSeek upending the costs and resources narrative and promising to democratise AI.
The Age of AI is well upon us, and the question is what India should do to leverage it towards its goals of a developed and equitable economy by 2047. India is well positioned to capitalise on this AI boom given its army of software developers and engineers, a thriving start-up ecosystem, global IT companies, a huge domestic market, its population-scale Digital Public Infrastructure, and a proactive government with an innovation agenda. It consistently ranks among the Top 5 AI countries in most global rankings, and receives a throng of AI grandees like Sam Altman, Jensen Huang, Yann LeCun ever so often, adding heft to its AI ambitions. Having said that, the two clear AI super powers who tower above every other country are the US and China, with most of the innovation, infrastructure, capital and research centred around them. In this scenario, while India would continue to build on the basics, it would need some out-of-the-box thinking and execution to leapfrog to become another AI leader.
The comforting fact is that India has done this before.
Even before we leapfrog, India will need to execute on the basics, and to be a global leader, that means investing and building across the AI stack, from GPUs to applications to research. This implies building AI infra in the country, thus building our own AI chips and data centres. While creating TSMC-level fabs is extremely expensive and difficult, building AI does not require only the latest 1.2nm chips and DeepSeek R1 has shown how it is possible to build foundation models on older GPUs and lesser of them. India will need to build a level of self-reliance, or atmanirbharta, on the basic infrastructure of AI. We will also need to build our own LLM, and the economics and possibilities of doing this have opened up greatly in the last two years, especially with plummeting infrastructure costs and clever algorithmic and systems tweaks, demonstrated again by DeepSeek, making it easier to build foundation models. It is heartening that now MeitY (Ministry of Electronics and Information Technology) has created a national mission around this. True value will be generated and delivered through AI apps built on top of them, and it is here that the animal spirits of our start-ups and IT companies need to be released to unleash a blizzard of AI apps.
What India should also focus on is to fund deep research in AI, where we are severely behind, since the bedrock of AI innovation is research. I do hope that fundamental research also becomes part of a National Mission with our universities and tech companies leading the charge.
However, even as we build the basics, the differentiator would be how we deliver AI to 1.4 billion Indians, and then the world, the same way we delivered digital like UPI and Aadhaar to everyone: by creating DPI and treating it as a free and widely accessible Digital Public Good. This is the ‘third path’ that India has walked before and where we can innovate once again. If we can build our LLMs and applications as part of the fabled ‘India Stack’ and deliver the amazing tools and apps that AI enables, we can create a hundred million creators, deliver information and services to the masses, and empower every device and citizen with AI.
I call this audacious initiative ‘Jan AI’, a play on GenAI, but for the ‘Janata’, and I do hope that this becomes the core of the National AI Mission, much like DPI was that years ago. This is the ‘AI for All’ road, less travelled by other countries, which will make every Indian citizen ‘AI Literate’ and catapult the country to AI glory and leadership and make it a world AI power sitting alongside, and perhaps ahead of, the existing superpowers.
The author is Jaspreet Bindra, Founder, AI&beyond, India, and Tech Whisperer Limited UK. Views are personal.