‘There is a lot of AI demand from enterprises…will start seeing localisation of infrastructure: Lenovo’s top executives on why India is becoming the top AI destination

“There is very strong AI demand globally, driven primarily by hyperscalers from both a hardware and services standpoint. Over the past year to year and a half, demand has also been driven by tier-2 cloud providers, which we call ‘neo clouds’. Between the hyperscalers and the neo clouds, it is true that, from a products and services perspective, they are currently driving the vast majority of demand for Lenovo’s products and solutions,” Bourguet told indianexpress.com in an interview at the Lenovo Tech World India Edition.

Pascal Bourguet, Global Channel Chief and Chief Operating Officer for International Markets at Lenovo (left), with Fan Ho, Executive Director and General Manager of the Asia Pacific Solutions & Services Group (right). (Image Credit: Anuj Bhatia/Indian Express)

To fuel growth in artificial intelligence, major tech companies are building hyperscale campuses across the heartland, turning farmlands and barren factories into data centres — essentially computing farms. Simply put, the AI boom cannot happen without data centres. While data centres are being built across the world, India wants a slice of the opportunity and is wooing major tech companies to develop large-scale AI data centres in the country.

India’s current data centre capacity is roughly 1.2 gigawatts, but the market is set to more than double, crossing 3 gigawatts within the next five years, according to industry estimates. Part of the reason global data centre companies are keen to enter India is the growth the country is witnessing, along with its massive internet penetration. Many of these facilities are being planned at the gigawatt scale, often by “hyperscalers” — companies that consume vast amounts of computing power. For context, one gigawatt equals 1,000 megawatts, and one megawatt equals 1 million watts.

Investments to build these data centres run into billions of dollars, even as the AI boom, which has sparked fears of a potential bubble, with returns on investment still relatively low for now.

But at the heart of the AI boom is the potential demand for artificial intelligence and rising AI workloads from enterprises. According to one estimate, 60 per cent of data centre clients are enterprises, 30 per cent are hyperscalers, and AI users account for only about 10 per cent.

Bourguet estimates that Indian enterprises currently have around 300 to 500 AI factories, driven by partners, signalling the early-stage adoption of AI in the country. He foresees further growth as enterprises move from project assessments to full-scale AI deployment and application adoption.

As many enterprises recognise the need to deploy AI including their own generative AI applications, they are realising the requirement for new hardware infrastructure. In response, companies like Lenovo are launching specialised servers and AI supercomputing cloud services to meet that demand.

“Particularly when we look across certain verticals, we see CxOs highlighting that they are planning to invest five times more in AI. The standout sectors are government, telecom, and healthcare. Beyond the natural demand in India, I think the Indian Government has also been making a strong push,” said Fan Ho, Executive Director & General Manager, Asia Pacific Solutions & Services Group, adding that sovereign AI is a key theme in the Asia-Pacific region, emphasising local data control and compute capacity.

Both Bourguet and Ho agree that India has a large pool of engineers skilled in machine learning and data science, along with the technical support required to scale enterprise AI. Citing International Data Corporation data, Bourguet said system integrators are expected to capture around 50 per cent of the enterprise AI market share.

While it is still early days for AI adoption, most enterprise IT organisations once viewed AI and the deployment of AI infrastructure as part of a long-term strategy. However, generative AI has accelerated that shift, and many are now beginning to see tangible benefits.

“Historically, major data centre hubs have closely aligned and coexisted with the points of presence of telecommunications networks. That was the phenomenon around a decade ago. This is why you typically see so-called mature markets in the Asia-Pacific region emerging as scale centres — Japan, Australia, and broadly China. However, in the AI era, two major things have changed. This is also why India has been moving ahead rapidly over the last one to two years,” Ho said.

Nutrient that powers AI models

For Ho, “data” is the nutrient that powers AI models. As a result, governance, data protection, and privacy considerations have become central issues. “We are therefore seeing increasing localisation of infrastructure to cater to these requirements,” she said, adding that hyperscalers and new cloud providers are stepping in to bridge the demand gap.

“More organisations are experimenting with and pushing the boundaries of AI use cases. This is driving demand because, ultimately, if you are doing advanced AI research or deploying AI-driven solutions, you need massive compute power. Whether it is research, building management systems, or industry-wide AI applications, the need for compute infrastructure is expanding across virtually every sector,” she added.

India offers several advantages for large-scale data centre development, including strong local demand, the rise of e-commerce, and new rules requiring social media data to be stored locally. However, experts warn that India will need large quantities of power and water, even though land is not a major constraint. Once data centres begin to scale, India will also need to carefully balance environmental concerns.

Ho says Lenovo’s Neptune warm water cooling technology reduces power and water usage by using a closed-loop warm water cooling system instead of chilled water and traditional air conditioning units. Additionally, it cuts total power consumption by reducing reliance on heavy air conditioning and helps address resource constraints while supporting high GPU heat dissipation.

Lenovo held its Tech World event in India a few days ahead of the India AI Impact Summit, a keenly followed event being held in Delhi. The summit will see thousands of participants and is drawing leaders from more than 100 countries, including a dozen heads of state, several dozen ministers or country representatives, and chief executive officers of leading global and Indian firms.