India is entering a pivotal phase in its Artificial Intelligence (AI) journey. Earlier this year, Bengaluru-based startup Sarvam AI unveiled two indigenous foundation models. Sarvam's achievement is significant not only for one company, but for what it reveals about how AI innovation occurs in India. The company did not build in isolation. Its multilingual offerings reflected a combination of India-specific datasets, public compute support, cloud and chip infrastructure, developer tools, and research ecosystems from across the global technology landscape. The result was not merely technological adoption, but global-scale innovation tailored to India's linguistic and market realities.
This pattern extends well beyond a single company. According to NASSCOM, India has become the world's second-largest hub for generative AI startups, with more than 890 GenAI startups and a 3.7-fold increase in cumulative startup formation. These figures point to an ecosystem that is still expanding rapidly and attracting new entrants, investment, and experimentation.
This is precisely why competition policy has entered the AI conversations. As AI systems grow more capable, while core concerns on infrastructural, talent, distribution, and access to foundation models are relevant — the policy question of how to preserve contestability without undermining collaboration is equally timely.
It is in this context that the Competition Commission (CCI) of India's market study on AI and competition released in October 2025 assumes significance. The study identifies issues that warrant close attention, including critical inputs such as cloud infrastructure, compute resources, data, and foundation models. These concerns are legitimate. Access to these inputs will play an important role in determining how competitive AI markets evolve over time.
At the same time, the CCI's emphasis on market studies, stakeholder engagement, advocacy, and ex-post enforcement reflects an important recognition: AI markets remain dynamic, technologically fluid, and far from settled. High market shares held by a few large players should therefore trigger careful monitoring and targeted scrutiny, but not a spree of enforcement activities or ex-ante obligations absent evidence of foreclosure, exclusionary conduct, or likely consumer harm
{{/usCountry}}At the same time, the CCI's emphasis on market studies, stakeholder engagement, advocacy, and ex-post enforcement reflects an important recognition: AI markets remain dynamic, technologically fluid, and far from settled. High market shares held by a few large players should therefore trigger careful monitoring and targeted scrutiny, but not a spree of enforcement activities or ex-ante obligations absent evidence of foreclosure, exclusionary conduct, or likely consumer harm
{{/usCountry}}The European Union's Digital Markets Act (DMA) was introduced to address concerns about the power of large digital platforms. While its long-term effects remain uncertain, early evidence suggests that broad ex-ante obligations can produce unintended consequences. The same is now visible in AI markets which are citing regulatory uncertainty under the DMA and related EU rules. A 2025 peer-reviewed study analysing more than 1,000 startup and investment records across 27 EU member states found that the DMA was associated with lower levels of market entry and investment in regulated sectors.
The experience carries a clear lesson for India: As a CUTS submission to the OECD Global Forum on Competition cautioned, importing a stringent, DMA-style ex-ante regime could stifle innovation and slow the growth of a still-expanding digital economy. The same concern applies to disproportionate enforcement: penalising scale or integration without evidence of exclusionary conduct or consumer harm could deter the investment and partnerships that domestic AI firms need to scale, at a critical stage in the development of India’s AI ecosystem.
India's AI ecosystem is still building scale and impact. Domestic firms are competing to develop products suited to India's economic, linguistic, and social realities, while investors continue to deploy capital into a rapidly evolving market. The CCI has projected that India's AI market could grow from $ 11.17 billion in 2025 to $ 131.31 billion by 2032, reflecting the scale of the opportunity still ahead.
Preserving space for India’s AI ecosystem to scale, however, does not mean leaving anti-competitive conduct unchecked. Existing competition law already provides tools to address anti-competitive agreements, exclusionary conduct and abuse of dominance where evidence warrants action. This framework could be reinforced by strengthening capacity for vigilant monitoring, evidence-based enforcement and structured engagement with the firms shaping India’s AI ecosystem, including cloud providers, chip suppliers, model developers and startups. Continued investment in technical and economic expertise would help the CCI distinguish pro-competitive scale from genuinely exclusionary conduct, ensuring that intervention addresses demonstrable harm without stifling the firms driving the country’s growth.
India's AI ambitions depend on many factors: entrepreneurial talent, investment, research capability, public infrastructure, government support, and access to global innovation networks. Sarvam's journey — from a two-person founding team in 2023 to unveiling sovereign language models on a national stage in under three years — reflects the convergence of all these elements. It happened because the ecosystem allowed it to. The next generation of Indian AI companies will emerge from similar combinations of domestic ingenuity and international collaboration.
The shape competition policy takes will also materially influence how India's AI ecosystem evolves. Recent evidence suggests that the ecosystem remains dynamic and contestable: a 2026 survey of 227 Indian AI firms found widespread multi-cloud adoption, limited use of exclusivity clauses, and falling compute costs, indicating that firms continue to enjoy meaningful choice across key layers of the AI ecosystem. In this environment, a proportionate competition regime can preserve the investment, partnerships and access to critical inputs that underpin innovation. A calibrated approach would protect the conditions that allow such breakthroughs to happen; a premature one risks foreclosing them before they begin.
(The views expressed are personal)
This article is authored by Anuj Gupta, MD, BowerGroupAsia and Abhinav Jindal, senior economist, Power Management Institute, NTPC Ltd.