India is steadily strengthening its position in the global artificial intelligence (AI) value chain, but its role today is less about building cutting-edge AI models and more about laying the hardware foundation that will power the industry’s next phase of growth, according to a report by Crisil Intelligence.
The report describes AI as a five-layer, or “penta-stack”, ecosystem in which semiconductors occupy one of the most critical positions. While India has made significant progress in chip assembly, packaging and design, it still has some distance to cover before emerging as a full-fledged semiconductor manufacturing hub capable of supporting the global AI race.
AI runs on chips before it runs on software
According to the report, AI is built on five interconnected layers: electricity, semiconductors, infrastructure, foundation models and applications.
Semiconductors form the second layer and are the computing backbone of AI. They include graphics processing units (GPUs) and specialised AI accelerators that train and run large language models such as OpenAI’s ChatGPT, Google’s Gemini, Meta’s Llama, Anthropic’s Claude and China’s DeepSeek.
Without advanced chips, AI infrastructure—including hyperscale data centres, cloud platforms and inference systems—cannot function.
Where India stands today
According to Crisil, India’s semiconductor ecosystem is expanding rapidly but remains concentrated in assembly, testing, packaging and chip design rather than advanced fabrication. One of the biggest developments is Micron Technology’s assembly and testing facility in Sanand, Gujarat, representing an investment of more than $800 million.
Currently, India’s only semiconductor fabrication project that is under execution is the Tata Electronics-Powerchip Semiconductor Manufacturing Corporation joint venture at Dholera, Gujarat, which aims to manufacture 28-nanometre chips.
The report also noted that while several India-based companies are active in semiconductor design and supporting infrastructure, domestic manufacturing capabilities are still evolving.
Why semiconductors matter for AI
Every AI workload, from training foundation models to processing user queries, requires high-performance chips capable of handling massive computing tasks. As the adoption of AI accelerates globally, demand for advanced semiconductors is expected to rise sharply.
However, Crisil points out that semiconductor companies should not automatically be viewed as AI companies.
“A chip company is not an AI startup simply because GPUs run large language models,” the report said.
India’s AI opportunity extends beyond chips
While semiconductors are a critical enabler, the report says AI-specific investments are currently concentrated in infrastructure and applications. Infrastructure, including GPU cloud providers, hyperscale data centres and inference platforms, along with AI applications accounted for 85% of tracked AI investment capital during the period covered by the report.
Foundation models attracted the remaining 15%, reflecting their strategic importance despite a relatively smaller share of investment.
Strong foundations, but AI-native ecosystem still developing
According to Crisil, India already has significant strengths in the foundational layers of the AI economy. Besides semiconductors, the country also has major energy companies such as Adani Green Energy, Tata Power Solar and ReNew Power, which are expected to benefit from rising electricity demand from AI data centres.
However, the report cautions that energy and semiconductor companies are horizontal industries serving multiple sectors and should not be classified as pure AI investments.