Whalesbook
Building India's Own AI Future
India has taken a major step toward building its own artificial intelligence ecosystem with a five-year Memorandum of Understanding (MoU). Larsen & Toubro Semiconductor Technologies (LTSCT), Larsen & Toubro-Vyoma, and BharatGen Technology Foundation are partnering to create a comprehensive national AI computing platform. This initiative aims to use India-made chips, AI models, and data center technology, supporting the national goal of relying less on foreign tech, improving data security, and fostering independent AI development.
Key Components of the AI System
LTSCT will lead the development of custom AI chips and computing platforms designed for national AI tasks, including large language models (LLMs). This specialized hardware aims for better energy efficiency and faster processing, essential for large-scale AI operations. Larsen & Toubro-Vyoma will contribute its large data center capacity, including its 30 MW data center ready for AI in Kanchipuram, Tamil Nadu, to provide the computing power needed. BharatGen Technology Foundation, supported by the Ministry of Electronics and Information Technology (MeitY) and the IndiaAI Mission, will define key AI tasks and optimize systems for running models. BharatGen has been approved to develop and deploy national AI models for large-scale use.
The Global AI Race and Market Size
The global AI chip market is set for significant growth, expected to reach about $1.1 trillion by 2035, up from $58.2 billion in 2025. This represents a compound annual growth rate of 33.9%. Major companies like NVIDIA, which held over 32.4% of the market in 2025, along with Intel, AMD, Qualcomm, and MediaTek, currently dominate this space. India is investing heavily in AI infrastructure, aiming for collective spending over $200 billion, with major commitments from Reliance and Adani. However, creating complete AI independence faces major difficulties. Achieving full AI sovereignty is often seen as very hard due to key bottlenecks in global supply chains, from raw materials to advanced chips.
Challenges to Achieving AI Independence
The idea of a self-reliant AI future is appealing, but the path forward has many risks. Developing custom AI chips and advanced AI models requires vast investment and years of research and development, where global leaders have a large head start. Seeking complete independence might lead to separate systems, higher costs, and make it harder to compete globally. While India's 30 MW data center capacity is substantial, it must be seen when compared to massive global data center projects. India also has its own aggressive plans for infrastructure expansion, including $110 billion from Reliance.
Risks and L&T's Position
This strategy faces significant challenges. LTSCT, aiming to be a product-focused company, is a new competitor in the highly challenging semiconductor design field. Its parent, Larsen & Toubro (LT), faces ratings issues, and its stock is seen as overvalued by 19%. LTSCT's revenue was about $41,000 as of March 2025, showing its early stage of product development. The semiconductor industry is global, with complex supply chains that are hard to avoid. Dependence on foreign parts and tools is a potential risk. Also, homegrown AI models, though vital for local needs, must eventually match the performance and wide use of leading global models. There is a considerable risk of investments becoming useless or money being spent poorly if domestic models and hardware don't perform well or get adopted widely.
Path Forward for India's AI Ambition
This partnership is a key move in India's long-term plan to build its AI capabilities. The success of this national AI computing platform will depend on steady government support, smooth execution across the entire development chain, ongoing innovation to close the technology gap, and creating affordable solutions that can compete globally. The project aims to make India a major player in AI, but the journey will involve navigating complex technological, economic, and political challenges.