India’s Hidden AI Story: The Companies Building the Backbone

Jestin James Principal Officer And Analyst, Axczel Research

Published September 21, 2026 • 9 min read

Where this post comes from. Major reference for this article: “Goldman finds a hidden AI trade beneath India’s lagging market” by the team at Press Insider (17 September 2026). Press Insider wrote about a Goldman Sachs research report. We added data centre and government programme facts from other sources, named where they appear and listed at the end.

If you look at India’s stock market this year, it seems like India is missing the AI boom. The Nifty, India’s main stock index, is down about 10% in 2026. India also has no giant chip maker or Big Tech company like the ones in Taiwan, South Korea or the US. So where are the India AI stocks that investors can actually buy?

But according to Press Insider, Goldman Sachs, one of the world’s biggest investment banks, says that is only half the picture. In a report dated 17 September, Goldman points to 42 Indian companies, a group of India AI stocks that help build what artificial intelligence (AI) needs to run. It calls them “AI Enablers.” Together they are up about 60% this year. The next best group Goldman compared, healthcare, is up about 10%.

Here is what that means in plain words.

Think of the gold rush

During a gold rush, not everyone found gold. But the people who sold shovels and tents did very well.

AI works the same way. Chatbots get the headlines. But behind them are huge buildings full of computers, called data centres. These need electricity, power cables, transformers and chips. The companies that supply these are the “shovel sellers” of AI. That is the group Goldman is looking at.

Why India looks “left out”

Press Insider reports that Goldman sees India as an “anti-AI” market. Only about 16% of the value of MSCI India, a popular index of Indian stocks, is linked to AI. In South Korea and Taiwan it is 70% to 80%. In China and Japan it is 30% to 50%.

That is a big gap, and it is why many investors have written India off for AI. Goldman’s point is that they may be looking in the wrong place.

Who are the 42 India AI stocks?

Goldman started with about 1,800 listed Indian companies, worth around $5 trillion in total. It then read company earnings calls and news to see who was really making money, spending money or signing deals linked to AI. That left 42 companies worth about $670 billion.

Half of these India AI stocks make equipment and machinery. Not all are giants: 12 are large companies, 9 are mid-sized, 13 are small and 8 are very small. Power equipment makers are worth about $95 billion, power producers about $47 billion and power transmission companies about $45 billion.

Press Insider names 18 of them: Reliance Industries, Bharti Airtel, Larsen & Toubro, Tata Power, Power Grid Corp. of India, Adani Green Energy, Adani Energy Solutions, ABB India, Siemens, Hitachi Energy India, CG Power and Industrial Solutions, Polycab India, Netweb Technologies, Kaynes Technology, Sterlite Technologies, Gujarat Fluorochemicals, Navin Fluorine International and Himadri Speciality Chemical. The full list of 42 is in Goldman’s report.

Are India AI stocks just hype?

Many AI rallies happen because people get excited and pay higher prices for the same profits. Press Insider says this one is different. These India AI stocks are actually earning more.

Since the start of 2025, the group has returned about 53%. Goldman worked out that growing profits added 65 points to that, while lower valuations took away about 12. Analysts have also raised their 2027 profit estimates for the group by about 22% this year, while estimates for the wider Nifty 500 (India’s 500 biggest companies) were cut by about 2%.

These companies are spending more on building, too. Goldman expects their spending to grow about 65% in 2026, after 18% last year. On its own, the group could add about 6 points to the 16% growth in spending expected across the Nifty 500.

Data centres are growing fast

Data centres are the “factories” of AI, so how fast they grow matters. Research firms count a little differently, but they all point the same way. Size is counted in the electricity a data centre can use: one gigawatt (GW) is 1,000 megawatts (MW).

Press Insider quotes CBRE, a property advisory firm, saying India had about 1.75 GW of working data centre capacity by mid-2026, with $38 billion of new investment promised in the first six months of the year. It also quotes JLL, another property firm, expecting capacity to rise from about 1.6 GW to 6 GW by 2029.

We read JLL’s mid-2026 report directly. It says data centre operators have grown about 60% a year since 2021. It also says 82% of the new space taken up in the first half of 2026 was booked in advance by giant cloud companies, who have promised over $50 billion.

Savills India says 258 MW was added in the first half of 2026, up 59% from 162 MW a year earlier, and expects India to pass 7 GW by 2030. Rubix Data Sciences, quoted by Business Standard, says building a data centre here costs 30% to 40% less than in China or the US.

Every one of these data centres needs power, cables and equipment. That is why so many of the 42 India AI stocks are power and equipment businesses.

The government is helping too

The IndiaAI Mission is a government programme, and one big part of it is computing power. AI needs special chips called GPUs, which are very expensive. According to the IndiaAI website, eligible users can rent them at up to 40% lower cost. Ten companies were picked to offer the service for 36 months, with an option to extend by 12 months. Together they offer more than 18,000 GPUs, including chips from NVIDIA, AMD and Intel.

The number keeps growing. Press Insider reports that in March, the electronics and IT ministry said 38,231 GPUs had been added through 14 service providers.

India AI stocks are not cheap, and not risk-free

These India AI stocks are not bargains. Press Insider reports they trade at about 36 times their expected yearly profit, roughly 85% more than the MSCI India index, and close to the top of their own five-year range of 28 to 39 times. In simple terms, buyers are paying a lot today for profits they expect tomorrow.

To be fair, growth matters here. Once you factor in how fast these companies are expected to grow, they are priced about the same as the wider index: 1.3 versus 1.4 on a common measure called the PEG ratio.

Not every part of this basket of India AI stocks is winning. Estimates for data centre hardware makers were raised about 95%, but they were cut by 15% for data centre developers and 14% for power producers. That suggests investors are picky about who they back.

Our idea: an index to follow India AI stocks and the country’s AI progress

If India’s AI story sits underneath the main index, it is hard for regular people to see it. So we think there is a case for a simple index that tracks the AI enablers in one place: power, grid, data centres, hardware and chips. Over time it could also add companies that use AI in their business. That would make it easier to follow India’s AI build-out next to the data centre and computing numbers above. This is only an idea for discussion, not a product.

This post is based on the reports and public data listed below, for information only. It is not investment advice. Figures are as reported by these sources and may change. Any mistakes in summarising them are ours.

Key takeaways

The claim: According to Press Insider’s report on a Goldman Sachs report dated 17 September 2026, 42 Indian companies (India AI stocks) that Goldman calls “AI Enablers”, which help build what AI needs to run, are up about 60% this year, while the Nifty is down about 10%. The author did not read the Goldman report directly.
Why India looks left out: Goldman sees India as an “anti-AI” market, with only about 16% of the value of MSCI India linked to AI, against 70% to 80% in South Korea and Taiwan. Data centre capacity is growing fast, with about 1.75 GW of working capacity by mid-2026 according to CBRE.
Not cheap, not risk-free: Press Insider reports the group trades at about 36 times expected yearly profit, roughly 85% more than the MSCI India index. Not every part of it is winning: estimates were raised about 95% for data centre hardware makers but cut by 15% for data centre developers and 14% for power producers.

Frequently asked questions

What are “AI Enablers”?

Goldman Sachs uses the term for 42 Indian companies that help build what AI needs to run, such as power equipment, power cables, transformers, chips and data centres. The article compares them to the people who sold shovels and tents during a gold rush: the chatbots get the headlines, but data centres behind them need electricity, cables and equipment.

How did Goldman pick the 42 companies?

According to Press Insider, Goldman started with about 1,800 listed Indian companies worth around $5 trillion. It then read company earnings calls and news to see who was making money, spending money or signing deals linked to AI. That left 42 companies worth about $670 billion. The full list is in Goldman’s paid report, which the author did not read.

Why does India look “left out” of the AI boom?

Press Insider reports that Goldman sees India as an “anti-AI” market. Only about 16% of the value of MSCI India is linked to AI, compared with 70% to 80% in South Korea and Taiwan and 30% to 50% in China and Japan. India also has no giant chip maker or Big Tech company.

How fast are data centres growing in India?

Research firms count differently but point the same way. CBRE says India had about 1.75 GW of working data centre capacity by mid-2026. JLL expects capacity to rise from about 1.6 GW to 6 GW by 2029. Savills India says 258 MW was added in the first half of 2026, up 59% from a year earlier, and expects India to pass 7 GW by 2030.

What does the article say about valuation and risk?

Press Insider reports the group trades at about 36 times its expected yearly profit, roughly 85% more than the MSCI India index and close to the top of its own five-year range of 28 to 39 times. Once expected growth is factored in, it is priced about the same as the wider index, at a PEG ratio of 1.3 versus 1.4. Estimates were cut for data centre developers and power producers, which suggests investors are picky.

Sources

Main source (most stock market facts in this post)

Press Insider, “Goldman finds a hidden AI trade beneath India’s lagging market,” by Press Insider Staff, 17 September 2026. It reports on a Goldman Sachs Global Investment Research report dated 17 September, which we did not read directly. It also quotes CBRE, JLL, Wood Mackenzie and the electronics and IT ministry, as noted above: https://pressinsider.com/news/goldman-finds-a-hidden-ai-trade-beneath-indias-lagging-market/