'The day that the market realises that they've overspent (on AI) and there's a sudden collapse in the capex, then India can start outperforming again.'

The artificial intelligence (AI)-driven trade, which has dominated the stock markets, especially in the United States, could come under the lens this year, Christopher Wood, global head of equity strategy at Jefferies, said on Tuesday.
However, for India, the 'AI story' is still unfolding, said Nilesh Shah, managing director at Kotak Mutual Fund.
From a stock market standpoint, so long as the AI capex is surging, Indian markets will continue to underperform, said Wood.
AI Capex Driving Global Markets
Key Points
- Surging AI capex globally may keep Indian markets underperforming compared to semiconductor-focused economies like Taiwan and South Korea.
- Hyperscalers plan $620 billion AI spending, but markets are questioning returns and potential excess data centre capacity risks.
- India is seen as a reverse AI trade, benefiting only if global AI capex slows or investment collapse occurs.
- Indian IT stock selling appears excessive, with insiders holding positions and long-term growth outlook still under debate.
- India unlikely to dominate LLMs but can leverage AI through applications, SLMs, and AI-driven investment research innovations.
They were speaking at the Business Standard Manthan Summit in New Delhi on the topic of role of AI & its impact on markets, moderated by Puneet Wadhwa.
Wood said the AI cycle began around three years ago, with Microsoft investing in OpenAI in 2023, as hyperscalers responded to the competitive threat posed by OpenAI.
"In my view, this is the year when the US stock market would start to question whether these hyperscalers are going to get any return on their investment (RoI).
"We started that process in the first quarter (Q1CY26) earnings season, where the markets questioned whether the hyperscalers will earn adequate returns because their business models have gone from asset light to asset heavy," Wood said.
India Reverse AI Trade
The four hyperscalers have projected to spend $620 billion as AI capex this year, which is fuelling the rally in AI stocks, especially in South Korea and Taiwan.
That said, Wood believes that the more the returns on AI capex are questioned, there will be more related concerns about the potential looming excess capacity in data centres.
This will, in turn, raise concerns regarding those funding such data centre construction and related infrastructure, he said.
Hyperscalers $620 Billion Spending
"But the day that the market realises that they've overspent (on AI) and there's a sudden collapse in the capex, then India can start outperforming again... For now, India is a reverse AI trade globally," Wood said.
"When we listen to the global AI companies, they say software is dead as AI can replicate what programmers took years to do. But when we come back to Indian IT companies, they say programming is just about one-fifth of their work.
"There is client interaction, designing, and communication among other things," Shah said.
India is unlikely to be a major player in Large Language Models (LLM), according to Shah.
"Our best bet remains in leveraging LLMs to create applications or small language models (SLMs), which provide solutions to the world.
"In India, not just large listed companies, but small startups are also doing a wonderful job in the area of AI," he said.
IT Stocks Selling Overdone
IT selling overdone?
The massive selling in Indian information technology (IT) stocks is unfathomable, as the tools which the industry has to manage uncertainty are not working, Shah said.
Despite the sharp fall, IT owners, or even senior employees with Employee Stock Option Plan, have not liquidated their position, he pointed out.
"Normally, when insiders sell, you know something is wrong.
"However, not a single IT owner or IT senior employee has liquidated their position.
"In markets' terminology, the stock prices reflect the net present value of future cash flow, which is dependent upon the terminal growth rate.
"As terminal growth rate goes from mid-single digit to low single digit or double digit to mid-single digit, the value changes by 30, 40, 50 per cent," Shah explained.
While the near-term business is unlikely to be impacted, Shah believes it will not materially alter share prices.
"What we need to handle is the long-term growth rate. And there, both business and the fund managers are going wrong," he said.
AI Redirects Global Fund Flows
India may see some "very interesting" small-caps in the AI sphere, but in the big picture, the key relevance for investors is whether IT firms will play a positive role in enabling corporates to move to these SLMs, Shah said.
That will be a viable role for IT companies, he added.
In terms of India's performance in the world markets, a monitorable will be whether the country will do the best relative to other markets, if the AI story blows up, according to Wood.
Fund management
Going ahead, fund management will be led by people who can leverage AI to deliver solutions, said Shah.
"But the reality is we are at risk, undoubtedly.
"Today, the computational power which AI brings on table is unbelievable.
"We have joined, hence, with Pascal.ai, the first asset management company in India to leverage AI for our investment research, to use AI for finding out patterns, doing better analysis of the past," he said.
For foreign investors, India was the stock market that "worked" in the emerging markets context, Wood said.
But they left the market as the 'AI story' started to develop across other markets, he said.
The AI boom has redirected capital toward semiconductor-heavy markets, such as Taiwan and South Korea, where companies are benefiting directly from surging AI-related capital expenditure, Wood said.
"Despite DeepSeek related concerns in early 2025, the continued rise in AI spending led to a sharp rally in semiconductor stocks, prompting global fund managers to raise allocations to these markets.
"As a result, India -- seen as less directly linked to the AI hardware supply chain -- saw selling," he said.
Feature Presentation: Ashish Narsale/Rediff








