Last week, Finance Minister Nirmala Sitharaman announced during her Budget speech that the Centre would allocate Rs. 500 crore to establish a Centre for Excellence in AI and Education. This is a positive step towards fostering homegrown AI innovation. But there is still more progress to be made. India has the intellectual strength and capability to create AI models, supported by a growing, globally competitive talent pool in fields like software development, engineering, and entrepreneurship. This positions India to become a significant player in the AI market.
DeepSeek, a Chinese AI firm providing an open-source Large Language Model (LLM), has surprised the world by creating a model as powerful as ChatGPT-4 at a fraction of the cost. According to the API documentation on its website, DeepSeek offers services at $0.14 per million input tokens and $0.28 per million output tokens. In contrast, OpenAI's ChatGPT Plus subscription and Perplexity AI's Pro plan are priced at $20 per month.
India is no stranger to technological innovation—consider Mangalyaan, the Mars orbiter launched by ISRO to study the surface and atmosphere of Mars, which cost just $74 million. If a small Chinese company can build an AI model that competes with giants like OpenAI in a short span of time and at a fraction of the cost, India can too.
India can learn much about agility and innovation from the circumstances under which DeepSeek was developed—with very limited funding and resources. DeepSeek's R1 model costs 20-50 times less to operate than OpenAI's GPT models while outperforming them in reasoning tasks. India has also been competing in this space with efforts like Ola's homegrown AI chat assistant, Krutrim, and BharatGPT, a generative AI platform that can generate text, voice, and video content in Indic languages.
The emergence of DeepSeek highlights a major difference between India and China in their approaches to research and development (R&D). China spends 2.65% of its GDP on R&D, while India invests just 0.7%. The gap widens in the private sector, with Chinese industries contributing 2.06% of GDP to R&D, compared to 0.23% in India. In 2024, China's total R&D spending grew by 8.3% nominally compared to the previous year, reaching around 3.6 trillion yuan.
The Government of India must step up and take action to regulate the sector while facilitating both foreign and domestic investments. It is also crucial to ensure that patents are protected and intellectual property rights upheld, allowing startups to safeguard their innovations and expand globally.
Fundamentally, the government needs to prioritise AI as a strategic sector. This means allocating resources, incentivising private investment, and developing the necessary infrastructure to support AI growth. Additionally, India should closely examine China's policy approach, particularly how it enabled a company like DeepSeek to thrive, and implement similar strategies to foster homegrown innovation.
(Venk is a venture investor with NuVentures and CEO of US fintech focused IT company NuWare)
Disclaimer: These are the personal opinions of the author
