The Indian data centre industry’s capacity¹ is set to more than double to 2-2.3 GW by fiscal 2027, led by increasing digitalisation of the economy as enterprises increase their investments in cloud storage and consumer demand for data surges. Further, rising penetration of Generative Artificial Intelligence (GenAI) will drive the demand over the medium term.

Incremental capital expenditure (capex) to support the strong demand would see a higher proportion of debt funding, resulting in a moderate increase in debt levels. That said, capacity additions will lag demand growth, keeping offtake risks low. As a result, the industry can expect healthy and stable cash flows, which will keep credit profile of players steady.

A Crisil Ratings analysis of industry players, representing ~85% of the market share by operational capacity, indicates as much.

Data centres cater to the computing and storage infrastructure demand, which is driven by two primary drivers. One, enterprises are rapidly shifting their businesses to digital platforms, including cloud, a trend that has accelerated post Covid-19 pandemic. Two, increased accessibility of high-speed data has led to a surge in internet usage, including social media, over-the-top (OTT) platforms and digital payments. Notably, mobile data traffic logged a compound annual growth rate (CAGR) of 25% over the last five fiscals. It stood at 24 GB per month at end-fiscal 2024 and is expected to rise to 33-35 GB by fiscal 2026.

In addition to the ongoing demand, rapid advancement of GenAI, which requires higher computational power and low latency than traditional cloud computing functions², will also provide tailwind to the data centre demand in India.

Says Manish Gupta, Senior Director and Deputy Chief Ratings Officer, Crisil Ratings Ltd, “To meet the growing data centre demand, an investment of Rs 55,000-65,000 crore is required over the next three fiscals, primarily towards land and building, power equipment and cooling solutions. Data centre operators typically build infrastructure - land and building, which account for 25-30% of overall capex – with the expectation of future tie-ups. While this approach may expose incremental capacities to utilisation risks, strong demand is expected to support capacity utilisation to reach 80-90% within a year or two.”

The capacity additions are driven by expansion plans of the existing players as well as entry of new players. These are on the back of significant demand from hyperscalers³. As hyperscalers typically wield high bargaining power due to large capacity requirements in a data center, they are able to secure competitive pricing. Typically, pricing of hyperscalers is likely to be 10-20% lower than other customers. Hence, balancing the ramp-up in capacity utilisation with pricing remains key for returns on data centre investments.

Says Anand Kulkarni, Director, Crisil Ratings Ltd, “Once capacities are tied up, data centres benefit from predictable cash flows backed by a stable client base resulting in low churn rates. This is due to high switching cost for customers on account of their investments and possible business disruptions when switching. Amid significant capex plans for expansion, the debt-to-earnings before interest, tax, depreciation and amortisation (Ebitda) ratio of data centre operators is expected to increase to ~5.4x this fiscal from ~5x last fiscal, before improving from next fiscal as capacity utilisation ramps up.”

In the milieu, data centre players’ timely capacity commissioning and tie-ups with customers, along with their ability to sustain pricing, will bear watching.

¹ Data centre capacity in India stood at ~950 MW as of March 2024
² For instance, processing an average ChatGPT query requires approximately 10 times as much electricity than a Google search.
³ Hyperscalers data centre facilities typically demand 5 MW-30 MW of IT power capacity. These data centers are generally built to suit the customer requirement and tenanted by fewer co-locating enterprises that commit to leasing large operating space.