New Delhi: India’s Artificial Intelligence industry is entering a new phase. Investors are putting hundreds of millions of dollars into homegrown AI companies, and startups that were barely known a few years ago are now reaching billion-dollar valuations.
But behind the excitement is a difficult question: Can these companies turn huge investments and valuations into real profits?
The numbers show how quickly the market is moving. According to a June 2026 analysis by CRISIL, Indian AI companies attracted more capital in the first half of 2026 than during the whole of 2025. Two deals played a major role: a $600 million investment in Neysa Networks and a $234 million first close for Sarvam AI’s Series B round.
Sarvam is perhaps the clearest example of the new AI wave.
The Bengaluru-based company raised $234 million in June at a post-money valuation of $1.5 billion, making it one of India’s newest AI unicorns. HCLTech alone invested $150 million in the round. Sarvam plans to use the money to develop frontier AI models and build the computing infrastructure needed to train and operate them.
Krutrim and Neysa are also part of the growing group of Indian AI unicorns.
The attraction is obvious. India has hundreds of millions of internet users, a large technology workforce and a growing demand for AI tools in areas such as banking, healthcare, government services and enterprise software.
But AI is an unusually expensive business.
Building advanced models requires huge amounts of computing power, specialised chips and engineering talent. Even after a model is built, running it for millions of users can be costly.
This creates a major difference between traditional software startups and AI companies. A software company can often add another customer at relatively low cost. AI companies may have to spend heavily on computing every time customers use their products.
CRISIL has warned that Indian AI valuations are running ahead of established commercial revenue. It also says the biggest challenge for India’s AI companies will be commercial viability, turning access to capital and computing power into sustainable business revenue.
That does not mean the valuations are necessarily wrong.
India could have a major advantage if its companies build AI products specifically for Indian languages, businesses and government needs. Sarvam, for example, is focusing on models designed for India’s scale, including voice and document understanding.
