India's AI moment: Summit can pave the way for accountable AI at scale

As AI adoption accelerates, India's AI Impact Summit aims to shape global rules, attract capital, and offer a Global South alternative to US- and China-led models

The AI Impact Summit in New Delhi this week comes when artificial intelligence (AI) is being rapidly adopted across the world. The government’s ambition is clear: To position India as a credible voice in the global AI landscape, largely dominated by the United States (US) and China. With participation from over 100 countries, the Summit is meant to amplify the Global South’s voice, attract capital, and shape rules that do not simply mirror western priorities or Chinese state-led models. In this respect, India has strengths to build on. It is among the world’s largest consumer markets for AI tools, with rapid adoption across firms and households. This demand pull is already drawing unprecedented commitments from global technology giants. Amazon has pledged over $35 billion through 2030, Microsoft $17.5 billion over four years, and Google $15 billion for what will be its largest AI and data-centre hub outside the US. India’s digital public infrastructure, low-cost data, and regulatory stability after the Digital Personal Data Protection Act have strengthened investor confidence. 

India’s broader strategy, alongside investment in sovereign compute and domestic large-language models, aims to balance expensive frontier-model development and focus on sectoral use cases. In this regard, the Economic Survey this year also flagged an asymmetry between frontier-model development and application-led deployment, noting that attempting to close the frontier gap could entail prohibitive costs. The tradeoff, therefore, is between chasing frontier-scale models and directing scarce resources towards domain-specific AI systems aligned with domestic priorities. Structural pillars such as expanding access to compute through 38,000 graphics processing units (GPUs), building the AI Kosha dataset repository, establishing an AI safety institute, and creating an AI incidents database are significant institutional moves. But enforceable accountability, robust data governance, and clear redress mechanisms will remain crucial. Voluntary compliance will not suffice if incentives push firms toward opacity. There are other constraints as well. India’s data-centre capacity is still a small fraction of the global total, and the country lacks homegrown AI firms that have achieved meaningful scale. Moreover, hyperscale infrastructure comes with heavy demands, including uninterrupted power, advanced cooling systems, fibre connectivity, and large water use. The design choices made now will determine whether India can expand AI compute without deepening environmental stress. 

At a broader level, the Summit’s success should not be measured only by headline investment announcements. Notably, it is happening against the backdrop of a fragmented global governance landscape. The 2023 Bletchley Park AI Safety Summit produced declarations and commitments, but a consensus on long-term risks remains elusive, and even regulating big tech cannot fully constrain the diffuse misuse of AI by countless developers. International spillover makes harmonised standards difficult, yet global discussion is essential. It is already evident how algorithmic tools can produce wrongful exclusion. Nevertheless, despite the challenges, it is encouraging that a global discussion on the issues will be held, and there is hope that some of the potential ill effects of AI will be contained soon. A serious debate will also be needed on the possible impact on jobs and the issue of AI-skilling to absorb the expanding workforce. There are a number of other issues to be discussed. Given that AI is rapidly evolving, with the potential to reshape the world in many ways, the Summit can be a good starting point for defining the direction and standards.