India has seen a 3X increase in number of GenAI startups in 2025: Economic Survey

India is setting a mark in the artificial intelligence landscape, as the country records a 3X increase in the number of Generative AI (GenAI) startups in the country, the Economic Survey on Thursday noted. 

According to the report, India had 890 GenAI startups in the first half of 2025 compared to 240 startups in the year-ago period. Amidst the rising number of startups coming up in the sector, VCs have also turned their attention to these companies. 

In H1 CY2025, GenAI startups attracted cumulative funding of $990 million—almost 30% higher than the capital that flowed into these startups in H1 CY2024. 

Firms working on GenAI infrastructure and models were a clear darling for investment firms, with these companies seeing around 63% of the total funding coming in. Startups building GenAI applications attracted 30% of the funding, while companies offering GenAI services captured a smaller piece of the pie with just 7%. 

According to the survey, Karnataka is the prime hub for GenAI startups, with 39 of the companies setting up in the state, followed by Maharashtra and Delhi. A handful of companies are also headquartered in Telangana, Haryana, Tamil Nadu, and Uttar Pradesh. 

Rise in funding in these startups also mirrors wider funding trends in deeptech companies. According to Nasscom, deeptech firms saw a 78% rise in funding in CY2024, the report noted. 

Also Read

Millions of Indian gig workers are stuck in low-pay jobs, Economic Survey finds

Challenges persist

Despite progress made in funding and setting up a nurturing startups ecosystem, India’s progress in AI continues to face challenges, the survey noted. “The AI ecosystem is characterised by pronounced asymmetries across countries, firms, capabilities, and stages of the value chain. These asymmetries are not incidental; they are structural outcomes of how AI is financed, developed, and deployed.”

One of the key pain points highlighted in the report is that the capability to design and train large foundational models remains highly concentrated in the hands of a few large firms that exercise control over the market and put pressure on necessary resources for AI. 

Adding to this, export restrictions imposed on most advanced processors required for scaling up frontier model development have posed a significant challenge to companies in India. 

The report also noted challenges related to transparency, the risks associated with data centres’ demand for large quantities of electricity, export controls imposed on advanced chips, and technology transfer.

For context, over 70% of all data centres are located in high-income countries, as of June 2025. Around 11% of data centres have been set up in China, and only a mere 3% have been set up in India, the survey noted.