Hitesh Tailor, Director, Sales Engineering, Ciena India
Over the past decade, India’s digital landscape has witnessed exponential growth, driven by increasing internet penetration, affordable data and widespread smartphone adoption. Network traffic has steadily expanded, further accelerated by cloud services, streaming platforms and enterprise digital transformation. Network traffic is set to grow further with the growing adoption of artificial intelligence (AI).
India is poised to become a global AI hub. Under the IndiaAI Mission, announced in 2024, the government allocated Rs 103 billion ($1.2 billion) over a five-year period to strengthen its AI capabilities. Combined with initiatives like Digital India, Make in India and data localisation policies, there is significant interest in developing high-end domestic compute capacity, potentially building one of the most extensive AI compute infrastructures globally. In parallel, hyperscale cloud providers, service providers and colocation players are taking strategic steps to scale their data centre footprints to support AI workloads.
The AI workload tsunami is coming
AI workloads, particularly training large language models (LLMs), require vastly different infrastructure compared to traditional cloud or analytics applications. To synchronise training across these geographically separated nodes or data centres, massive amounts of data must be exchanged at ultra-high speeds. This requires robust and extremely high-capacity data centre interconnect (DCI) networks.
In a recent study conducted by Ciena, an average of 52 per cent of new data centre facilities are expected to be dedicated to AI workloads. This is expected to put more strain on DCI infrastructure over the next two to three years.
India’s data centre infrastructure must scale
India is experiencing a surge in new data centre developments, with its power consumption growing by over four times in the past six to seven years to reach 1,263 MW as of April 2025. It is likely to cross 4,500 MW by 2030 and attract investments of $20 billion-$25 billion in the next six years.
Tier 1 service p roviders across India have taken new initiatives to address the country’s growing AI and data centre demand. For instance, NVIDIA and Reliance Industries have collaborated to build supercomputing infrastructure to support the demands of AI. India is also witnessing investments by hyperscalers like AWS and Microsoft in the area.
To support LLM training and inference workloads, data centres must communicate at extremely high speeds within campuses, across metros and even on the long haul. Traditional DCI solutions may fall short of the scale and requirements of massive AI data centres. Indian data centre operators have thus begun investing in intelligent, adaptive and high-capacity network architectures designed to dynamically respond to changing traffic patterns.
What does it mean for optical networks?
While there is growing recognition of the need to accelerate optical fibre deployment, Indian service providers continue to face challenges such as implementation delays, lengthy permitting processes and difficulties in obtaining right of way (RoW) for laying new cables. However, the administration is committed to resolving these issues. It has recently simplified the RoW rules and set up a dedicated portal to streamline the process. The high cost of laying fibre and lease utility poles, ducts and optical ground wire can also remain a deterrent to optical infrastructure expansion. Network operators will need to get the most from every fibre pair.
To help service providers maximise the potential of their fibre infrastructure, both long-haul and metro, Ciena continues to challenge the status quo with its record-breaking coherent optical technology. WaveLogic 6 provides massive 1.6 Tb per second single-carrier wavelengths for metro deployments and delivers ultra-high speed connectivity, including a mix of 400G and 800G client traffic across the longest links. In addition, new integrated C&L-band Wavelength Selective Switch modules allow service providers to fully utilise both the C- and L-bands, using only one-third of a single rack.
Within and around data centres, traditional Intensity Modulation Direct Detection (IMDD) technologies may not be sufficient to address the growing demand. Innovative products such as Ciena’s 1.6T Coherent-Lite pluggables provide a significant boost in performance and capacity to support new realities.
There is no universal architecture or strategy for AI-ready DCI networks. Some providers will build AI-optimised greenfield campuses with dedicated interconnect backbones, while others will retrofit existing facilities with higher-capacity links and intelligent routing layers. Hybrid cloud and edge AI use cases will further complicate network design.
A robust and scalable DCI infrastructure will be fundamental as India enters a new era of AI-led growth. Whether enabling AI research, powering next-gen chatbots, or automating enterprises, AI workloads will transform how Indian data centres are built and networked. Investments in fibre infrastructure, intelligent automation and high-speed interconnects will not just be a technical necessity, they will be a strategic differentiator in India’s digital economy of the future.
