en.Wedoany.com Reported - At a factory in Mumbai, western India, technicians work around the clock drawing glass into optical fibers thinner than a human hair, destined for data centers that power the global artificial intelligence boom. Indian manufacturers are supplying cables, power systems, and cooling equipment for AI infrastructure and profiting from it, despite concerns that India is lagging in the technology development race.
Nomura analysts describe this role for Indian manufacturers as a classic "selling shovels" play—where selling tools during a gold rush is more profitable than seeking the precious metal. Ankit Agarwal, Managing Director of Sterlite Technologies, says serving the world's largest tech companies marks a major shift for the firm, which was previously seen primarily as a telecom network supplier. "This is a once-in-a-lifetime opportunity," Agarwal told AFP. "This is a completely new network."
India has embraced AI enthusiastically but has yet to produce globally competitive models and still relies on imported advanced semiconductors. Investors also worry that generative AI could disrupt the country's $315 billion IT services industry, automating tasks such as customer service and application maintenance that were previously outsourced to India. These concerns have weighed on the market, with the benchmark Nifty 50 index down about 7% this year, and foreign investors having pulled more than $27 billion from Indian equities.
The data center boom has brought windfall gains to largely overlooked companies. Prateek Nigudkar, fund manager at Shriram AMC, told AFP that India doesn't need to manufacture chips or train models to benefit—every data center requires switchgear, transformers, cables, cooling systems, and grid connections, and Indian manufacturers are already in that supply chain.
Optical fiber cable producers are among the biggest beneficiaries. Sterlite announced earlier this year a multi-year order worth $1.1 billion from a U.S. hyperscale cloud provider, helping boost its share price by 400% this year. Telecom equipment maker HFCL posted its strongest April-to-June quarter ever, with revenue more than doubling on AI demand. Founder Mahendra Nahata says the company is going all out to expand production. "In the past, we might have added capacity every three or five years. Now, it's a routine task," he said, comparing the construction wave to a "tsunami."
The gains are not limited to fiber optic cables. TD Power Systems' shares have risen 68% on expectations of growing demand from AI facilities, while MTAR Technologies benefits from demand for power supply units used in fuel cell systems serving major tech operators.
Analysts warn that valuations have become too high, while opposition to energy-intensive data center projects also poses a risk. AI intelligence firm 10a Labs' tracking arm, Data Center Watch, says at least 75 U.S. projects worth about $130 billion were delayed or blocked in the first quarter of 2026. Indian executives acknowledge these risks could hurt short-term sentiment but argue that domestic demand will ultimately offset any overseas slowdown.
India holds nearly one-fifth of the world's data but accounts for less than 5% of global data center capacity, leaving huge room for growth. A review of financing announcements by AFP shows that over the past year, U.S. hyperscale cloud providers have committed to adding about $57 billion in capacity in India, while conglomerates Adani Group and Reliance Group have pledged $100 billion to digital and infrastructure projects. Unlike the United States and parts of Europe, data center development in India has so far faced limited resistance.
Sharad Agarwal, CEO of data center operator Sify Infinit Spaces, says opposition may still be in its very early stages, and operators have adopted technologies such as closed-loop water cooling systems to ease pressure on local resources. Sify operates 16 data centers, with nearly a dozen more under development, and is considering an IPO, which would make it India's first publicly listed data center operator. "We're talking about massive investments. If I'm going to be a gigawatt-scale company, I need $5 billion," Agarwal said, noting that 1 gigawatt of computing capacity is often called "hyperscale." "Only then can we grow."
