Bessemer Venture launches $350-million India fund - Industry News
US-based Bessemer Venture Partners has closed its second India-dedicated, early-stage focused fund with a corpus of $350 million. “With our second fund, we will continue backing early-stage founders building across AI-enabled services and SaaS, fintech, digital health, consumer brands, and cybersecurity,” the firm said in a statement. Partners Vishal Gupta and Anant Vidur Puri will lead the investment strategy.
Investing in India
Bessemer Venture began investing in the country nearly two decades ago. Its first India-dedicated fund of $221 million was launched four years ago. The company, which currently has $1.5 billion worth of assets under management (AUM) in India, will begin deployment of the new fund in the next 4-6 months. So far, it has backed 80 startups in India including Big Basket, Urban Company, Perfios, and Livspace. In the last five years, more than 80% of its investments in India have been Series A or earlier–companies.
Focus on AI and Software
While the fund invests in early-stage companies across several themes, there is a strong focus on AI and software, said Puri in an interaction with FE. “AI is one of the biggest and ground-breaking technology shifts during the course of our lifetime so far, and this yields a variety of different opportunities,” he said.
Both Puri and chief operating officer Nithin Kaimal see a strong opportunity in building across areas within AI, such as foundational models, infrastructure, applications, and AI-enabled services. Globally, companies such as Anthropic, Open AI, and Mistral AI, among others, have successfully built large language models, while in India, Ola’s Krutrim and Sarvam AI have ventured into the space.
Announcing Voxel51's Series B Led By Bessemer Venture Partners ...
“The fact about the foundational model opportunity is that it is capital-intensive and also needs large datasets,” noted Puri. He added that now is the right time to build foundational models as over the last few months a lot more GPUs have been made available for different research fields, and relevant Indian datasets have been made public. “India is not the economy to spend $10 billion to build foundational models. The emergence of DeepSeek has proved that you can build these models without having to spend hundreds of millions of dollars,” Puri said, adding that he expects more foundational model companies coming out of India, including those building small language models.
