Tata Comm to invest $450M in high-speed fibre network to connect AI data centres

Tata Communications is planning an India-wide fibre network designed as an “AI super highway.” The network will provide high-speed, low-latency links between major AI data-centre corridors. The company plans to invest about $450 million, or roughly Rs 4,280 crore, in the programme.

The first phase targets the Mumbai-Pune-Hyderabad-Chennai corridor. These four cities form one of India’s important routes for AI and hyperscale data-centre growth, according to the company. The planned network will cover around 10,000 kilometers of new long-haul fibre across India.

Later phases will extend connectivity into emerging East Coast and Western India data-centre ecosystems. Locations named by Tata Communications include Vizag, Bhubaneswar, Jamnagar, and GIFT City, along with other strategic sites. The network will also connect with the company’s global backbone, supporting links between Indian facilities and major global AI hubs.

A long-haul fibre network is a high-capacity optical-fibre system linking distant cities, regions, or data centres. It forms the backbone between locations rather than serving only nearby buildings. Tata Communications is building such a network to connect India’s main AI data-centre corridors.

For example, a data centre in Mumbai could exchange information with facilities in Hyderabad or Chennai over the planned network. Fibre carries data as light, allowing large volumes of information to travel quickly. Low latency reduces the delay before data reaches its destination, while high capacity helps handle many connections and demanding workloads.

This matters because AI workloads can require data centres to work together. Tata Communications says its network is intended to support the performance, resilience, and scale needed for next-generation AI workloads. It will also connect with the company’s global backbone and major international AI hubs.

Tata Communications plans to invest close to $450 million, equivalent to about Rs 4,280 crore, in the new network. The company expects to build around 10,000 kilometers of long-haul fibre across India. This scale is intended to create one of the country’s largest AI-ready connectivity networks.

The first phase focuses on four cities: Mumbai, Pune, Hyderabad, and Chennai. Subsequent phases will reach two broader areas described as emerging East Coast and Western India data-centre ecosystems. The company specifically names Vizag, Bhubaneswar, Jamnagar, and GIFT City, as well as other strategic locations.

The article does not give a final number of regions or cities. It says the network will expand in phases and connect India with major global AI hubs. Tata Communications also sees India’s data-centre-to-data-centre connectivity market potentially exceeding $1 billion over the medium term.

When AI data centres connect through faster, higher-capacity, lower-latency links, they can exchange data more efficiently. Lower latency means less delay between sending and receiving information. Higher capacity allows more data and more simultaneous traffic to move across the network.

For example, Tata Communications’ planned links could connect AI facilities along the Mumbai-Pune-Hyderabad-Chennai corridor. The company describes this as a network fabric for data-centre-to-data-centre connectivity. It is intended to support workloads that may need coordinated access to computing resources and data across multiple facilities.

The article does not provide measured performance gains or specific AI applications. It says the network aims to support the performance, resilience, and scale required for next-generation AI workloads. Connecting the system to Tata Communications’ global backbone should also help provide end-to-end links between Indian data centres and major global AI hubs.

Mumbai, Pune, Hyderabad, and Chennai are becoming important because Tata Communications describes their route as one of India’s key AI and hyperscale growth corridors. “Hyperscale” refers here to the large-scale data-centre ecosystem mentioned by the company. The article presents the corridor as a priority for new connectivity.

The planned first phase will connect these four cities with high-speed, low-latency long-haul fibre. This creates a direct foundation for data-centre-to-data-centre connectivity across the corridor. Tata Communications says the wider network will support the performance, resilience, and scale needed by next-generation AI workloads.

The article does not provide separate growth figures for each city or explain the individual reasons each location is expanding. It does identify the corridor as a major AI and hyperscale route. Later phases will broaden the network toward East Coast and Western India ecosystems, including Vizag, Bhubaneswar, Jamnagar, and GIFT City.

The new terrestrial network is an added layer in Tata Communications’ existing infrastructure. It will provide high-speed, low-latency long-haul connections between India’s AI data-centre corridors. At the same time, the company says it will continue serving existing enterprise and data-centre needs through its national long-distance and metro network footprint.

The planned AI network will also be integrated with Tata Communications’ global network backbone. Its wider infrastructure includes more than 500,000 kilometres of subsea optical fibre and more than 200,000 kilometres of terrestrial fibre. These networks can work together across local, national, and international paths.

The article also notes new subsea plans. Tata Communications is planning a Mumbai-Singapore cable and has invested as a consortium member in a Chennai-Singapore system expected to be ready in the last quarter of 2029. The combined footprint is intended to support India-focused and global data-centre connectivity.

Optical fibre carries digital information by converting data into pulses of light that travel through thin glass strands. This basic mechanism is established technology; the article itself focuses on Tata Communications’ network plans rather than explaining fibre physics. Fibre is suited to long-distance, high-capacity connections.

Bandwidth is the amount of data a connection can carry over time. Higher bandwidth helps move large datasets and support many simultaneous connections. Latency is the delay between sending and receiving information. Lower latency helps connected systems respond and exchange information more quickly.

These qualities matter for AI systems because Tata Communications is building links between data centres, not just connecting nearby offices. The company says its network will support the performance, resilience, and scale required for next-generation AI workloads. Its global integration is also intended to connect India with major international AI hubs.