As Artificial Intelligence (AI) technologies evolve, the demand for high-performance computing continues to surge. Since 2011, AI training computing has increased by 350 million times, placing immense pressure on global and national infrastructure. For India, which is rapidly emerging as a digital and AI-driven economy, the time has come to reimagine its computing and build a resilient and inclusive AI ecosystem, India’s computing policy must evolve with the changing demands of the AI landscape.
This includes supporting the shift from training to inference and edge computing, particularly in high-impact sectors like telecommunications, manufacturing, and healthcare. It must also prioritise strategic sectors that are early adopters of AI while empowering startups—especially those working on deep tech and multilingual solutions—with affordable and seamless access to computing resources.
Crucially, the policy should strike a balance between large-scale centralised infrastructure and flexible, cost-efficient decentralised systems to ensure broad accessibility and innovation across the country. strategy for long-term innovation, sovereignty, and inclusive growth.
Rising AI Compute Demands and Data Centre Pressure
AI’s exponential growth is paralleled by an unprecedented rise in data centre demand. A McKinsey report estimates global data centre capacity could triple by 2030, with 70% of the demand driven by AI. However, these centres require critical inputs like real estate, power, and high-end semiconductors—resources that are often scarce and unevenly distributed.
To meet AI’s growing needs, nations must look beyond traditional centralised systems. India, for instance, is investing in both centralised GPU provisioning and exploring distributed micro data centres for greater flexibility, cost efficiency, and energy savings.
India’s Current AI Compute Landscape
India’s computing strategy is rapidly evolving to meet the growing demands of AI innovation, marked by a 24% CAGR in data centre expansion since 2019, the launch of the India Compute Portal offering subsidised GPU access, and supercomputing capabilities under the National Supercomputing Mission.
Innovations such as IIT Madras’ Kompact AI and Ziroh Labs are making it possible to run AI models on CPUs, reducing reliance on expensive hardware. Additionally, proposals for decentralised micro data centres aim to optimise power and space efficiency. Together, these initiatives underscore India’s recognition that its AI future hinges on computing strategies tailored to real-world needs—from training and inference to edge deployment.
Global Approaches to AI Infrastructure
Different countries are embracing varied models to strengthen their AI computing capabilities. The USA adopts a market-driven approach, with tech giants like NVIDIA, OpenAI, and AWS leading the charge through privately funded AI “factories.” China takes a state-led path, developing “intelligent computing centres” and underwater data centres to address land and energy constraints—though rapid expansion has caused imbalances in infrastructure utilisation.
Key Pillars for India’s Future-Ready Compute Policy
To build a resilient and inclusive AI ecosystem, India’s computing policy must evolve with the changing demands of the AI landscape. This includes supporting the shift from training to inference and edge computing, particularly in high-impact sectors like telecommunications, manufacturing, and healthcare.
It must also prioritise strategic sectors that are early adopters of AI while empowering startups—especially those working on deep tech and multilingual solutions—with affordable and seamless access to computing resources. Crucially, the policy should strike a balance between large-scale centralised infrastructure and flexible, cost-efficient decentralised systems to ensure broad accessibility and innovation across the country.
The Way Forward
India’s AI future depends on its capability to optimise computing approaches with scale, efficiency, and access in mind. If it makes strategic investments in both centralised and decentralised infrastructure, develops nuanced awareness of shifting compute needs, and implements thoughtful policy, India can quickly move ahead in the global race of AI.
As demand shifts from training to real-time inference and edge, India’s next-gen compute infrastructure must change rapidly – enabling innovation in urban and rural, enterprise and startup, public and private spaces.
About The Author
