Global technology giants account for over 60% of NTT’s India data centre demand

Avinash Joshi, Executive Managing Director - India, NTT DATA.

Global technology giants have taken up nearly 60–65 percent of NTT Data’s data centre capacity added in India over the past three years, said Avinash Joshi, Executive Managing Director–India, NTT DATA, reflecting how the global artificial intelligence (AI) boom has turned the country into both a major consumer market and a key technology talent hub for them.

For most US-headquartered AI giants, India provides a significant growth opportunity as a consumer region. India is OpenAI’s second-largest user market outside the US, while it is the largest user base region for Perplexity AI.

Speaking to Moneycontrol, Joshi said, “About 60-65 per cent of our capacity is being done for global tech giants. This was not the case a few years back, but in the last three years or so, a majority of what we have built is for the global tech giants based on the build-to-suit model.”

“Those are huge mega projects, which we will continue to do,” he added. NTT has so far built operational data centre capacity of about 340 MW, and another 160 MW is in the works, which Joshi expects will be ready over the next two year.

Japan-headquartered NTT DATA is the third-largest data centre service provider globally, and has been growing its presence in India over the last decade.

Since 2011, NTT has invested over $3 billion in India and has its second-largest workforce outside Japan in the country. NTT is currently deploying an additional $1.5 billion across three years until FY27.

“This comes to about $500 million of investments each year, and again, a portion of that will go into the data centre business. You would have probably read some time back, we acquired a big land parcel in Mumbai,” Joshi said.

Earlier this month, NTT committed more than Rs 2,400 crore to build a new multi-data center campus in Devanahalli near the Bengaluru airport. Once ready, the full campus will house three data centres with a combined capacity of more than 60 MW.

According to Joshi, India is on the cusp of the data centre market expanding as demand is expected to explode with the new Digital Personal Data Protection (DPDP) Rules getting implemented, need for sovereign cloud, and states competing with incentives to attract data centre investments. He expects newer competitors to enter the space, such as Tata Consultancy Services’ (TCS) recent entry into building data centres.

“We remain very bullish on the potential opportunity offered by the Indian market, and to add to that the whole momentum around GCC is also growing.  For us, India continues to remain and become important not only from a data centre standpoint but also from an IT services standpoint,” he said.

He added, “We announced the opening of six new AI-powered cyber defence centres, four of which are operational in India. We see huge demand, especially in priority areas like cyber.”

Joshi sees that data sovereignty demand is not always driven by the need to keep information within the country, but also depends on enterprises looking for just better performance and latencies based on their workloads.

“More regulated industries, especially if they have capital, would prefer the sovereign data centre, the sovereign infrastructure. In this case, the data centre will be ours, but it will be their infrastructure in our data centre. That's where I think the demand will come,” he said.

“The advent of AI is an important factor in why demand is coming, and everybody is bullish about us as a country. The use of AI and LLM models, combined with the government’s push under the India AI mission, development of domain-specific language models, all of these are contributing to the demand,” he added.

Joshi noted that there’s an expectation gap with customers in the industry as they hope to get high double-digit efficiencies by deploying AI, but that’s not possible if the underlying data architecture is not modernised.

“Some of the mature clients, after doing a couple of POCs (proof of concepts), realize that they cannot get the full value of AI unless they fix the foundation of the data…It is a learning curve, and people are realizing it. The success ratio of all the PoCs for AI is not more than 17 per cent or 20 per cent. Sometimes clients may experience a bit of POC fatigue,” he said.

Joshi added, “Our advice to clients is rather than trying to do AI everywhere, pick a domain and go end-to-end. It could be customer experience or financial operations or delivery, or distribution.”