Authored by Dipankar Sarkar, Co-Founder at KiranaPro
The rapid rise of China’s DeepSeek shows that new players can successfully enter the artificial intelligence field without massive budgets like those of Silicon Valley giants. While DeepSeek proved it’s possible to build competitive AI systems more efficiently, India needs to develop its own unique approach rather than copying existing players.
DeepSeek’s story offers valuable insights. Founded by finance expert Liang Wenfeng, the company cleverly used existing computing resources from its investment business to build AI systems. They managed to create their AI model for about $6 million, while OpenAI reportedly spent around $100 million on GPT-4. This shows how smart resource use and creative technical solutions can help newcomers compete with established companies.
However, India shouldn’t just try to build similar AI systems to those that already exist. Instead of using the same technical approach as companies like OpenAI and DeepSeek, India could explore newer methods of building AI systems. For example, some researchers are developing promising new ways to build AI models, like the “Mamba” approach, which works differently from current systems and might be more efficient.
The government’s role should be carefully planned. Rather than directly funding AI development, which private companies can handle better, government resources should help create the right environment for innovation. This could include supporting research into new AI methods, helping develop and use alternative computing chips, and creating public-private partnerships similar to how India successfully launched UPI for digital payments.
India should also look for opportunities to challenge existing tech monopolies. Just as the startup Perplexity is offering a fresh take on internet search, Indian companies could find similar opportunities to innovate in other areas. This requires thoughtful collaboration between government agencies, investors, and tech companies to identify and pursue these opportunities.
The current dependence on NVIDIA’s specialized AI chips is both a problem and an opportunity. India could take the lead in promoting other options, including chips from AMD and newer companies. Government policies and targeted funding could help create demand for these alternatives, potentially making AI development more affordable and innovative.
To build lasting progress, India needs to attract AI experts, particularly Indians working abroad in this field. This talent strategy should focus on specific, practical projects rather than open-ended research. The new India AI mission could be shaped to include more concrete goals and clear ways for private companies to participate.
One approach worth considering is creating non-profit AI research centers with industry participation, similar to how OpenAI started. These centers could serve as shared facilities for AI development while putting any profits back into expanding their capabilities. This could help spread out the high costs of AI development while ensuring more organizations can access these technologies.
The key to success is avoiding the urge to simply catch up with current leaders. Instead, India should focus on areas where it can lead through innovation. This could include developing AI solutions for uniquely Indian challenges, exploring new technical approaches, or creating novel ways to deploy AI that build on India’s strengths in software development.
As AI technology continues to evolve rapidly, India has a chance to become a pioneer rather than a follower. By learning from DeepSeek’s efficient approach while creating its own path, India could establish a unique and lasting position in the global AI landscape. Success depends on smart government support, private sector innovation, and a clear focus on areas where India can truly stand out rather than competing in already crowded spaces.
