India must develop its own sovereign large language models (LLMs) or risk becoming permanently dependent on foreign AI providers, according to a report by global brokerage Bernstein. The report argues that while India has emerged as a major adopter of AI, it lacks foundational AI models comparable to China's DeepSeek, leaving the country vulnerable to geopolitical disruptions, supply restrictions, and limited control over critical AI technologies.
Bernstein emphasizes that building AI applications alone is not enough. To achieve true technological independence, India must own the "intelligence layer" by developing indigenous foundation models, AI infrastructure, and computing capabilities. The report notes that India's technology ecosystem has traditionally focused on IT services rather than creating large-scale consumer platforms that generate the proprietary datasets needed to train frontier AI models.
The report also points to India's strengths, including its vast engineering talent, expanding digital public infrastructure, and government-backed initiatives such as the IndiaAI Mission. However, it argues that greater investment in research and development, high-performance computing, and long-term funding for deep-tech startups will be essential if India hopes to compete with AI leaders such as the United States and China.
Bernstein concludes that developing a homegrown AI ecosystem is becoming a strategic necessity rather than just an economic opportunity. By investing in sovereign AI models, compute infrastructure, and domestic innovation, India can reduce its dependence on foreign technologies, strengthen national security, and position itself as a globally competitive AI powerhouse in the years ahead.
