Thu Jan 23 12:00:00 UTC 2025: ## Perplexity AI CEO Challenges Infosys Co-founder on India’s AI Strategy

**NEW DELHI** – Aravind Srinivas, CEO of Perplexity AI, has publicly challenged the AI development strategy proposed by Infosys co-founder Nandan Nilekani, arguing that India should prioritize building its own large language models (LLMs) rather than solely focusing on applications built on existing models.

Srinivas, while praising Nilekani’s contributions to India’s technological landscape, disagreed with Nilekani’s advice to Indian startups at the Meta AI Summit in October. Nilekani had suggested that Indian startups should avoid the expensive task of developing LLMs and instead concentrate on creating cost-effective, practical AI solutions tailored to India’s specific needs.

Srinivas countered this viewpoint on X (formerly Twitter), stating that India risks missing a crucial opportunity by neglecting LLM development. He argued that the perceived high cost of training these models is a misconception and pledged $1 million and five hours per week to support a team dedicated to developing India’s LLM capabilities.

“I feel like India is falling into the same trap I did…assuming that training models will cost an exorbitant amount of money,” Srinivas said, referencing his own experience at Perplexity AI. He emphasized the need for India to develop its own proprietary AI capabilities, particularly for Indic languages, to achieve global competitiveness. He drew a parallel to Elon Musk’s admiration for ISRO, highlighting the potential for achieving great things with efficient resource management.

Srinivas’s challenge presents a contrasting vision for India’s AI future, pitting a focus on indigenous LLM development against a strategy prioritizing application-centric solutions. The debate highlights the strategic choices facing India as it aims to become a major player in the global AI landscape.

Read More