India is positioning itself to lead the AI wave. At the India AI Impact Summit in February 2026, infrastructure investment pledges were estimated at $200–250 billion, according to the ministry of electronics and IT. What will determine whether this moment becomes a lasting movement is not capital alone, but the infrastructure ecosystem underpinning it, particularly high-capacity, programmable networks.
India is scaling compute rapidly. Jefferies projects Indian data centre capacity reaching 8 GW by 2030, representing approximately $30 billion in investment, with clusters forming across Mumbai, Chennai, Hyderabad, Bangalore, and other key cities. But as AI moves beyond centralised training into distributed, orchestration-intensive workloads, the network is no longer just linking compute resources. It is becoming part of the compute fabric itself.
AI workloads driving new-age network infrastructure demand
AI workloads create fundamentally different network demands. Training clusters involve thousands of processors exchanging data every few milliseconds; any bottleneck in data movement idles GPU clusters worth millions. These connectivity challenges span three layers: subsea links connecting India to the world, mega-scale terrestrial networks linking distributed data centers that must function as single training clusters and distributed low-latency connectivity reaching enterprises and edge locations for AI inference at scale. This third layer is where service providers see a growing opportunity, with many exploring managed optical fiber network models to build dedicated infrastructure for hyperscalers.
Global hyperscalers continue to see India as a priority build market. Google’s $15 billion America-India Connect initiative is establishing Visakhapatnam as a new subsea gateway. Microsoft’s $17.5 billion commitment includes India’s largest hyperscale cloud region in Hyderabad. OpenAI has partnered with Tata Group under its Stargate project to develop AI infrastructure, starting with 100 MW of data centre capacity.
Reliance, Adani and L&T bet big on AI infrastructure
Indian conglomerates are matching this ambition. Reliance Industries has shared plans to commit $110 billion over seven years. Adani Group plans to scale its data centre platform from 2 GW to 5 GW. Larsen & Toubro is building gigawatt-scale AI factories on NVIDIA GPU infrastructure in Chennai and Mumbai.
India’s AI talent base is also significant. The India Skills Report 2026 estimates the country accounts for 16% of the world’s AI talent, and investments like IndiaAI Mission’s support for 500 PhD scholars and 5,000 postgraduates are building depth in model development and systems architecture.
A decisive shift is underway in AI, from large language model training toward monetisable intelligence via inferencing at scale. This transition to always-on production systems in enterprise workflows, and the rise of agentic AI, will generate distributed traffic patterns that require programmable, high-capacity networks. Service providers and enterprises that treat connectivity as a strategic priority are best positioned to capture the full value of what is shaping up to be one of the most significant technology build outs in India’s history.
India’s AI moment is backed by record investment, expanding talent, and growing global confidence. Turning it into a lasting movement will require compute, connectivity, and intelligence working together at scale. The building blocks are in place. The opportunity now is to connect them.
The writer is VP, Solutions Engineering and CTO for Asia Pacific, India and Japan at Ciena.
Disclaimer: The views expressed are the author’s own and do not reflect the official policy or position of Financial Express.