The country’s AI ecosystem has not yet shown major signs of anti-competitive conduct despite concentration in computing infrastructure and frontier AI models, a new study by The Dialogue has found. Based on a survey of 308 businesses, AI developers and consumers, it said key challenges faced by developers today include data standardisation, talent availability, interoperability and privacy rather than a lack of service-provider choice.
The findings has come around a time when competition concerns around AI markets receive greater attention, particularly over the role of large technology companies in cloud services, AI models and applications. In October 2025, an AI-focussed market study from the Competition Commission of India (CCI) had suggested that companies should conduct self-audits of their AI systems to identify and address potential competition concerns.
Among AI developers, 74.5% said there was sufficient choice among cloud providers, while about 85% relied predominantly on public cloud infrastructure. For those that had switched or attempted to switch cloud providers, technical difficulties were a bigger hurdle than the availability of alternatives, the study found. These constraints include data transfer, compatibility issues and the risk of service disruption, while relatively few (14.6%) cited a lack of alternative providers.
The study said these switching challenges were largely technical and may not be addressed through competition interventions by the CCI. Instead, it has suggested targeted technical or regulatory measures. “One of the survey’s key findings is that switching costs stem primarily from the technical complexity of migrating workloads rather than a lack of competitive alternatives,” the report said.
Meanwhile, open-source AI models such as Meta’s Llama are giving start-ups and developers more competitive alternatives, the study said. Some 96.2% of AI developers surveyed said they rely on open-source models and tools, while over 79% see them as a partial substitute for proprietary frontier models such as OpenAI’s GPT, Google’s Gemini and Anthropic’s Claude. The report said open-source AI is lowering entry barriers, particularly for model fine-tuning and application development.
The findings come as India weighs the need for ex-ante competition rules. The report noted that the draft Digital Competition Bill (DCB), which was released in 2024, proposes ex-ante obligations for systemically significant digital enterprises (SSDEs). However, it said competition regulators globally have generally taken a cautious approach to AI markets, favouring market monitoring and targeted intervention over broad ex-ante obligations as the sector continues to evolve. The current version of DCB lists nine core digital services that include cloud services but does not specifically include AI services.
According to the study, the AI adoption is generating business benefits with 85% of businesses rating its impact positively. The Dialogue has further recommended expanding access to computing, supporting open-source AI and interoperability, and monitoring major AI partnerships and investments.
“India’s AI ecosystem is showing encouraging signs of competition, innovation and adoption. The evidence in this study is important because it helps distinguish between potential competition risks and demonstrated market harm. India should continue creating the conditions for competition, experimentation and innovation to flourish through an evidence-led approach,” said Kazim Rizvi, founding director at The Dialogue.