While it may not carry the tag of ‘Silicon Valley of India’ like Bangalore, Chennai, located just 300 kilometres away, has emerged as a dominant force in the country’s growing AI ecosystem. With 19 data centres, the city ranks second in India for operational data centre capacity, behind Mumbai, which has 38 data centres.
As of June last year, the total operational data centre capacity in Chennai was estimated to be 88 megawatts (MW). Since then, major new establishments from CtrlS (72 MW data centre) and Sify have further increased this capacity.
The Coastal Advantage for Data Centres
Mumbai and Chennai dominate India’s data centre landscape due to their coastal geography, which makes them best suited for undersea cable landing stations. These facilities connect the extensive ocean‑spanning fibre-optic lines that carry international internet and phone traffic directly to terrestrial networks.
On Thursday, at the inauguration of Sify’s Chennai 02 ‘AI-ready’ data centre, which has a capacity of over 130 MW, CFO Vijay Kumar said that the two cities function like “international airports for data”.
“For the internet, the two [major] gateways in the country are Mumbai and Chennai, where all the subsea cable systems come and land on both sides,” said Kumar, indicating the importance of undersea cable systems in data transfer. “Chennai, being closer to the cable landing systems, offers low latency and high operational effectiveness,” he told AIM in an exclusive interaction.
Major players, such as NTT Data, Bharti Airtel, Tata Communications, and Sify, have equipped their Chennai data centres with on‑site cable landing stations. The Chennai 02 facility’s landing station can handle data transfers at rates of up to 4 petabytes per second.
There are seven active undersea cables in Chennai, with three more under construction. These connect to nineteen countries worldwide, as well as other cities in India. Connections to major data hubs include Europe (France and Italy) and Southeast Asian countries such as Indonesia, Malaysia, Singapore, and Thailand.
While numerous data centre projects are underway, research from JLL estimates that the city’s operational capacity will double by next year. Mordor Intelligence estimates that the overall capacity may exceed 551.15 MW by 2030.
Though India has numerous coastal cities where data centres could thrive, Chennai stands out alongside Mumbai for its exceptional growth in this sector. This dominance also stems from Chennai’s robust business infrastructure, extensive talent pool, thriving IT ecosystem, and growing number of global capability centres (GCCS).
Chennai accounts for 15% of India’s total IT workforce, and Tamil Nadu ranks third in the country in terms of software exports. Notable Indian IT giants such as Infosys, Wipro, and Tech Mahindra have established offices in the city, besides global firms like Accenture, Amazon, Microsoft, IBM, Oracle, and others.
Furthermore, the state hosts more than 250 GCCs. Chennai is also widely recognised as the ‘SaaS capital of India’.
Kevin Imboden, global director of market research at EdgeConnex, that has set up a data centre in the city in collaboration with Adani Enterprises, said, “While other cities across India received initial interest as data center hubs, whether for ties to large headquarters (Mumbai), the IT industry (Bengaluru), or the national government (Delhi and the key Noida area), Chennai embodies parts of all these cities, with a strong and multifaceted array of industries calling the area home [for data centre establishments].”
Besides, the state government actively supports data centre development through its 2021 policy. The Tamil Nadu Data Centre Policy offers various financial benefits for companies establishing data centres in the state. These include tax concessions, electricity subsidies, land cost reductions, and additional financial assistance to investors.
Moreover, there is also a booming ecosystem of infrastructure that can handle AI-related workloads in Chennai. Last year, Sify was the first in the country to achieve NVIDIA’s DGX-ready data centre certification for liquid cooling.
This certification validates that data centre components, including infrastructure, software platforms, and managed services, are fully optimised for NVIDIA DGX AI systems. These turnkey solutions integrate high-performance GPUs, storage, and software into a cohesive package for AI workloads.
Soon after, the company also announced the Sify Cloudinfinit +AI, a platform that provides GPU-as-a-service for enterprise users to support various AI, deep learning and compute-intensive workloads. The company provides services for India’s leading banks, manufacturing hubs, and multiple government organisations.
Recently, E2E Cloud unveiled India’s most extensive GPU clusters, which include two high-performance clusters, each featuring 1,024 NVIDIA H200 GPUs, placed in Delhi NCR and Chennai. Furthermore, institutions like IIT Madras and the Institute of Mathematical Sciences (IMSc) in Chennai operate advanced GPU clusters to support AI research and workloads.
“Going forward, any new infrastructure will be an AI-ready data centre to host high workloads,” said Kumar. “It is important that infrastructure should precede the development of applications,” he added, indicating how data centers are ready to host the demands of upcoming AI workloads and applications.
IIT-M is the X Factor
Besides hardware infrastructure, the presence of the Indian Institute of Technology Madras (IIT-M) in the city not only provides a talented pool of engineers but has also undertaken several research projects and initiatives in AI.
For example, the Wadhwani School of AI (WSAI) at IITM brings together industry leaders, researchers, and students to collaborate on addressing various challenges in different areas of artificial intelligence. Additionally, it offers undergraduate and postgraduate programs in AI and data science.
Most of these research initiatives arising from the institution are focused towards building AI for India. In a conversation with AIM, Balaraman Ravindran, head of WSAI, said the school is focusing on developing protocols to test the consistency of language models’ reasoning explanations across multiple attempts, and creating datasets to evaluate stereotypical content specifically in Indian languages rather than relying on English translations.
Recently, IIT Madras and the IITM Pravartak Technologies Foundation partnered with California-based Ziroh Labs to establish the Centre of AI Research (CoAIR) to solve India’s compute accessibility challenges using AI models optimised for CPUs and edge devices.
Ziroh Labs introduced Kompact AI at IIT Madras – a platform that runs foundational AI models on CPUs instead of GPUs. Developed by Indian engineers, it already supports 17 optimised models (including DeepSeek, Qwen, and Llama) and aligns with India’s ‘AI for All’ initiative.
Furthermore, AI4Bharat, the AI lab at IIT Madras, recently introduced IndicTrans3-beta, a state-of-the-art (SOTA) multilingual translation model designed to support translations across 22 Indic languages.
Last year, the AI lab also launched BhasaAnuvaad, a speech translation dataset tailored for Indian languages, boasting coverage across 13 languages and approximately 44,400 hours of audio. This was the largest publicly accessible speech translation resource of its kind for Indian linguistic diversity.
All things considered, Chennai’s research and development initiatives, along with powerful AI-ready infrastructure hubs, are set to play a key role in the IndiaAI mission, which focuses on building a frontier AI model in the country.
The post India’s AI Future Will Ride on The Chennai Express appeared first on Analytics India Magazine.