India is stumbling in the global AI steeplechase. It risks relegation to the margins, while others claim the prizes. (Representative image)
Credit: iStock Photo
Not having a large listed artificial-intelligence (AI) firm is partly why foreign institutional investors (FIIs) have removed net Rs 1.7 lakh-crore from the Indian stock market since the start of 2025, even as they pour that money into AI firms like Nvidia. Strangely, this is hardly a topic for discussion on either primetime news channels or Parliament.
This exodus underscores a deeper malaise: despite grand visions, India is stumbling in the global AI steeplechase. It risks relegation to the margins, while others claim the prizes.
India’s AI paradox is stark. Home to 16% of the world’s AI talent, it has almost nothing to show for the effort. The $10 billion India AI Mission aims to bridge computing and infrastructure gaps, but cannot obscure a fundamental truth: India lacks a world-class large language model to rival America’s GPT or China’s DeepSeek. Local ventures like Sarvam AI and Krutrim offer narrow applications, confined to regional languages or niche tasks. Indeed, Krutrim founder Ola is lagging in its core businesses. Without a proprietary LLM capturing global imagination, Indian firms remain dependent on foreign tech giants.
This dependency extends to business models. No Indian private company boasts a proprietary AI framework dominating any sector, unlike OpenAI’s generative tools or Anthropic’s ethical AI. Instead, TCS and Infosys excel in services — implementing others’ AI solutions abroad. Their offerings are enablers, not originators. This service approach, while profitable in the short term, cedes high ground in a ‘winner-takes-all’ arena.
Consider past technological disruptions. Google monopolised search by perfecting algorithms and scaling data dominance, leaving Yahoo behind. Netflix revolutionised streaming, and Facebook mastered the social network. AI exhibits identical dynamics: network effects amplify early leaders. US big tech and China’s Baidu control vast datasets, computing resources, and talent pools, creating barriers that latecomers struggle to breach. India risks becoming the equivalent of a regional video rental shop in the Netflix era — useful but irrelevant to core innovation.
Funding woes exacerbate the lag. Indian AI startups attracted merely $780 million in 2024, dwarfed by the US’ $109 billion. This disparity persists as global capital flows to clearer AI leaders. India’s research spending, which is about 0.6% of its GDP, pales against China’s investments and US’ venture machine.
Infrastructure tells a sobering tale. India hosts 20% of global data but commands just 3% of data-centre capacity. Startups complain of GPU shortages and exorbitant cloud costs. Data sovereignty remains elusive — local firms rely on foreign datasets as quality Indian data sits siloed in government vaults.
Yet hope exists, albeit precariously: that AI proves overhyped like dot-com promises. Recent IDC and McKinsey reports suggest that the $1 trillion in global AI investments yield limited productivity returns. McKinsey notes 80% of businesses using generative AI report ‘no significant bottom-line impact’, with pilots abandoned due to inaccuracies and costs. The IDC finds modest ROI — potentially adding just 0.5 percentage points to productivity growth. If AI’s transformative potential fizzles, India’s service role could prove rewarding, leveraging vast talent, and cost advantages in implementation, mirroring IT services success.
But this is a high-stakes gamble. Should AI reshape industries as profoundly as the Internet, India’s lag will prove disastrous. Missed opportunities would compound, sidelining the nation in the intelligence revolution it helped fuel through talent exports.
In Greek mythology, Tantalus faced eternal hunger with a fruit forever dangling beyond reach — a torment of proximity without possession. India risks a similar fate in AI: surrounded by data and ingenuity yet unable to grasp the transformative bounty. To avoid this, policymakers must double down on R&D, foster proprietary innovation, and build infrastructure at scale.
It will cost money, some of which may be wasted or misappropriated, but both the government and the private sector must take that chance. Otherwise, the world’s most populous nation may forever chase shadows in the age of machines.
Ninad D Sheth is a senior journalist. X @ninadsheth
Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.