CCI study flags steep barriers for Indian AI startups, calls for open data and compute access to level playing field

The CCI found that control of a few large firms across the AI stack — from cloud infrastructure to data and foundation models — is creating a tilted playing field.

India’s competition watchdog has sounded a cautionary note on how the country’s AI ecosystem is evolving.

The Competition Commission of India (CCI), in its Market Study on Artificial Intelligence and Competition, released on October 6, 2025, noted that while Indian startups are driving much of the innovation in applied AI, they remain constrained by steep entry barriers and structural dependencies on global tech giants.

What stakeholders  said

According to the study, nearly 67 percent  of Indian AI startups work in the “AI application model layer”, with only 3 percent developing foundation models.

The respondents in the study included 106 stakeholders, including 50 startups and researchers, 30 customers/major deployers of AI and other experts.

Most rely heavily on open-source technologies — “76 percent of the startup respondents build their application solutions using open-source technologies," the study noted.

In terms of tools, “88 percent of the respondents use ML to build their AI solutions, 66 per cent use generative AI (LLMs), 78 percent use NLP and 27 percent work in the field of CV.”

Despite the rapid spread of generative AI tools, startups remain cash-strapped. The study showed that most startups were either self-funded or backed by angel investors, and only “16 percent of respondents reported access to next-level funding.”

What about barriers to fair competition?

The CCI found that control of a few large firms across the AI stack — from cloud infrastructure to data and foundation models — is creating a tilted playing field.

“Control of a few large firms across the AI stack may create barriers to entry for smaller players. The survey identifies availability of data, talent, compute facilities and cost of cloud services amongst potential entry barriers," the CCI study.

It adds that high training costs, limited compute access, and reliance on foreign hyperscalers like AWS, Google Cloud, and Azure make it difficult for Indian firms to compete.

“Established entities own vast high-quality datasets, which may not be accessible to small firms… Removing these barriers is imperative to create a level playing field, encourage the entry of new players and stimulate competition in the AI marketplace," the CCI recommended.

Are there concerns over pricing and collusion?

Yes. Responding to CCI’s perception survey, 37 percent of startups acknowledged a “possibility of AI facilitated collusion,” while 32 percent feared price discrimination and 22 percent predatory pricing driven by algorithmic systems.

The report elaborates that “self-learning algorithms may reach collusive outcomes on their own depending upon the market conditions… guided by the objective of profit maximisation.”

What CCI recommends

To build a “competitive, dynamic and innovative AI ecosystem,” the CCI recommends several measures:

• Access to infrastructure and data: “Continued emphasis on measures and policy initiatives by the government to facilitate access to AI infrastructure, and to enhance AI capabilities is recommended.”
• Self-audit of AI systems: “Enterprises are urged to include self-audits of AI systems for competition compliance that will allow businesses to proactively identify and address potential competition concerns.”
• Transparency framework: “The enterprises are encouraged to adopt transparency measures to reduce information asymmetry.”
• Advocacy and regulatory coordination: The CCI will hold conferences on AI and regulatory issues, “focus on strengthening its technical capabilities,” and “undertake steps to promote inter-regulatory coordination and international cooperation.”