Tata Communications’ plan to invest Rs 4,350 crore in a 10,000-km AI-ready connectivity network is more than a telecom expansion. It is a signal that India’s data-centre economy is beginning to organise itself around dedicated digital corridors linking computing clusters, enterprise demand and international cable infrastructure.
The company announced the investment on Monday, saying the project would be developed in three phases between FY30 and FY36. The proposed network will connect major data-centre corridors and add about 6 petabits per second (Pbps) of capacity to Tata Communications’ existing network, which currently has 1.25 Pbps of capacity and is utilised at around 70-80%.
The first phase will focus on the Mumbai-Pune-Hyderabad-Chennai corridor. Later phases are expected to extend connectivity to emerging data-centre ecosystems in eastern and western India, including Visakhapatnam, Bhubaneswar, Jamnagar and GIFT City. The company said the investment would be financed through operating cash flows.
That geography is important because the project is not being framed as a single-city data-centre buildout. Instead, it treats data centres, AI hubs, cable landing stations and enterprise networks as parts of a connected national system. The company’s stated objective is to provide high-capacity, resilient and low-latency links between locations where computing, cloud services and business demand are concentrated.
The urban infrastructure behind AI
AI infrastructure is often discussed through the visible assets of the digital economy: data-centre buildings, server halls, cooling systems and power connections. Tata Communications’ announcement highlights a less visible layer that determines whether those assets can operate as a network: the fibre routes connecting them.
A data centre can host computing capacity, but AI workloads also require rapid movement of large volumes of data between facilities. The company said the planned network is designed for data-centre-to-data-centre connectivity, with the ability to connect Indian AI hubs to global markets through international cable landing stations and Tata Communications’ global network backbone.
This makes the project relevant to urban planning even though much of the infrastructure will be underground or outside city centres. Mumbai, Pune, Hyderabad and Chennai are being treated as a linked growth route rather than isolated technology markets. The same logic applies to the proposed extensions towards Vizag, Bhubaneswar, Jamnagar and GIFT City, where the availability of connectivity can support the development of emerging data-centre ecosystems.
The supplied announcement does not establish how individual routes will be built, which agencies will approve them, or how much of the network will use new versus existing rights of way. It does, however, make clear that Tata Communications intends to build on its national long-distance and metro networks while continuing to expand them for enterprise and data-centre requirements.
Capacity is the central number
The most significant numerical detail is the proposed addition of 6 Pbps of capacity against an existing 1.25 Pbps. If delivered as described, the new capacity would be several times the company’s current network capacity. The existing network is already being used at approximately 70-80%, according to the company, creating a stated basis for further capital deployment.
These figures also show why the project is being planned over a long period rather than as a one-time capacity addition. The three-phase schedule runs from FY30 to FY36, indicating that the network is intended to expand alongside demand from cloud adoption, AI deployment and enterprise connectivity.
Tata Communications’ managing director and chief executive, Ganesh Lakshminarayanan, said the data-centre-to-data-centre connectivity market in India could exceed $1 billion over the medium term. The company also said a growing proportion of its expansion was being driven by anchor customer commitments and contracted demand.
Those statements point to a capital model based partly on identifiable demand rather than only on an open-ended bet on future AI growth. The company has not disclosed the value of the anchor customer commitments in the supplied material, nor has it provided a phase-wise breakdown of the Rs 4,350-crore investment. The precise allocation between corridors, network upgrades and new infrastructure therefore remains unclear.
The network is being designed for global hyperscalers, AI platforms, neocloud providers and enterprises. This customer mix matters because each group may require different combinations of route diversity, speed, resilience and access to international networks. The announcement does not specify service pricing, contracted bandwidth or the expected capacity allocation among these customer categories.
From metro networks to international cables
Tata Communications said it operates a global network spanning more than 500,000 km of subsea optical fibre and over 200,000 km of terrestrial fibre. The proposed AI network will build on that infrastructure, linking domestic data-centre clusters with international cable landing stations.
This arrangement places the project at the intersection of three network layers. The first is the metro network within and around major cities. The second is the national long-distance network connecting urban and regional data-centre locations. The third is the international backbone that links Indian facilities to global markets.
For cities, the implication is that data-centre competitiveness is not determined only by land, electricity or building permissions. Network connectivity is another infrastructure dependency. A location may attract computing facilities only if it can connect reliably to other centres, customers and international routes.
The announcement does not provide information on electricity consumption, water requirements, land use or the physical environmental impact of the proposed network. It therefore cannot establish the wider sustainability profile of the planned infrastructure. It does show, however, that the digital expansion associated with AI will require investment beyond the data-centre campus itself.
A changing map of India’s digital economy
The first-phase corridor runs through Mumbai, Pune, Hyderabad and Chennai, four established urban and technology markets identified by Tata Communications as a key route for AI and hyperscale data centres. The later inclusion of Visakhapatnam, Bhubaneswar, Jamnagar and GIFT City suggests an effort to connect newer locations to the country’s more established digital infrastructure.
That expansion could reduce the dependence of India’s data-centre ecosystem on a small number of established clusters, but the supplied material does not claim that the project will decentralise data-centre development or change regional market shares. At this stage, the company has only identified the corridors and ecosystems it intends to connect.
The distinction between connecting a location and creating a data-centre market is important. Fibre infrastructure can enable low-latency links and improve access to national and international networks, but the growth of a data-centre ecosystem also depends on other conditions that are not detailed in the announcement. These include customer demand, power availability, land, cooling and regulatory approvals. The company’s statement focuses on connectivity and does not provide evidence on those wider requirements.
The policy and institutional layer
The project will be financed through Tata Communications’ operating cash flows, placing the immediate investment responsibility with the company rather than identifying a public funding package in the supplied material. Implementation will nevertheless depend on the permissions and coordination required for national long-distance, metro and international network infrastructure.
The announcement does not identify the government departments, municipal bodies, utilities or agencies that will be involved in route approvals. It also does not state whether the project will require new public-right-of-way arrangements or rely primarily on existing network corridors. These details will determine how quickly the three phases can move from investment planning to physical deployment.
Tata Communications’ statement comes as India’s demand for cloud and AI infrastructure increases. The company’s network plan is therefore both a commercial expansion and a response to the infrastructure requirements of a more distributed digital economy. The central institutional question is how quickly network capacity can be expanded across multiple jurisdictions while maintaining the resilience required by data-centre operators and enterprise customers.
What the announcement confirms—and what it does not
The company has confirmed a planned investment of Rs 4,350 crore, a target of around 10,000 km of AI-ready connectivity infrastructure, an additional 6 Pbps of capacity and a three-phase development period between FY30 and FY36. It has also identified the first corridor and the locations it expects to connect in subsequent phases.
The announcement does not establish the final route map, construction schedule for each segment, expected completion date for individual corridors, number of connected data centres or the financial contribution expected from each customer category. It also does not provide a project-level assessment of land, energy, water or traffic impacts associated with network construction.
The next stage will therefore be defined by implementation detail. Information on phase-wise capital expenditure, route approvals, anchor customer commitments, deployment milestones and the balance between existing and new infrastructure will show how the proposed AI network translates into physical connectivity.
For now, Tata Communications’ plan provides a clear view of one structural shift: India’s AI buildout is being planned not only around computing facilities, but also around the long-distance, metropolitan and international networks that allow those facilities to function as a connected system.
