The recent correction in technology stocks should not be mistaken for the end of the artificial intelligence trade. The more important question is not whether AI-linked stocks are falling, but whether the investment in AI infrastructure and applications is slowing. And, by that measure, the cycle remains firmly intact, says Samir Arora, founder of Helios Capital.
Arora argues that investors are conflating two very different things: the performance of AI-related stocks and the spread of AI itself. The latter, he believes, is continuing, with some of the world's largest technology companies not only maintaining their investment plans but stepping them up.
“When Google, Meta and the others say we will invest $200 billion or $250 billion, and their stock prices fall, is that a failure of US AI? Or does that mean the real spenders on AI continue to believe in it, perhaps even more than they did before?” he asks.
Arora cautions against celebrating a fall in US technology stocks on the assumption that the money will automatically flow into Indian equities or that Indian IT services companies will become beneficiaries of a failed AI trade.
That, he says, is a misleading way to read the market. “AI trade, whether the stock price goes up or down as of now, is not failing,” he says. By AI trade, he means the continuing investment and adoption of the technology, not the day-to-day movement in the share prices of companies associated with it.
Big Tech's spending may earn reasonable return
Arora remains reasonably constructive on the large US technology companies, but says investors should look beyond the headline capital expenditure numbers and assess the revenue that this investment can generate over the next two to four years.
The market, in his view, is focusing on the money being spent without adequately considering the future revenues from the computing capacity being created.
He does not expect these investments necessarily to deliver the extraordinary profits that investors have become accustomed to from some of these businesses. Instead, he sees AI infrastructure increasingly resembling a utility business.
“It is more of a utility kind of business. It is like somebody setting up a nuclear plant or a power plant. Basically, you will get rental income out of it. But still, it won't be bad,” he says.
Arora believes the large technology companies can earn a reasonable return on their capital expenditure while also strengthening their existing businesses and participating in a new growth opportunity.
He points to the scale and credibility of the companies making these investments. Unlike the dotcom boom of 1999, when money flowed into a multitude of speculative internet ventures, the current AI investment cycle is being led by established companies with substantial businesses and cash flows.
“This is not like '99, where the internet guys were basically people leaving their regular jobs and starting travel channels and paid food channels and websites, and then getting private equity money. Here, the guys spending the money are Amazon, Google, Meta and Microsoft…,” he says.
For these companies, the hurdle is not necessarily to generate outsized profits from AI immediately. They need to earn a reasonable return on the incremental capital invested and use the technology to support their broader businesses.
Big Tech Vs pure-play AI stocks
Arora makes an important distinction between the large technology platforms and companies whose valuations depend almost entirely on AI.
Amazon, Meta and Microsoft, he points out, have substantial non-AI businesses. They should therefore not be clubbed together as pure-play AI stocks.
The more speculative end of the market includes semiconductor-related plays and new AI ventures. On the latter, Arora is decidedly cautious.
He does not see a compelling reason for public-market investors to buy into start-ups such as Anthropic or OpenAI at the valuations being discussed.
“For them to prove that AI is making money at that starting public-market valuation, that is an extremely difficult thing,” he says.
While established technology companies can afford to invest in AI because the technology can reinforce existing revenue streams and create additional business opportunities, a pure-play AI company, by contrast, has to justify its valuation largely through its ability to monetise AI itself.
Arora believes that is a much more difficult proposition, particularly at elevated valuations.
Google is ‘confusing'
Among the large technology companies, Google presents a more complicated investment case. It is the biggest individual stock holding in Arora's global fund, but he acknowledges that the disruption posed by AI to its core search business creates uncertainty.
The concern is straightforward: if users increasingly turn to AI tools such as Gemini for answers rather than using conventional search, Google's established search business could come under pressure.
Arora says Google will have to find ways to carry advertising into AI-generated answers. How it does so remains uncertain.
“Google is the only one a bit confusing, actually, although in our fund, it's the number one holding in terms of individual stocks,” he says. At its current valuation, he believes Google could find new ways to monetise the answers generated by AI.
“At the end, everybody has to monetise the answers,” he says.
The memory shortage opportunity
Arora also sees opportunities in semiconductor-related companies. He specifically mentions owning Samsung, pointing to the continuing shortage in memory chips.
His investment case rests on the expectation that the shortage could persist for another four to five years. Even if earnings eventually moderate after that period, he argues, investors should not assess the business as though those lower earnings will arrive immediately.
The company would have had several years of growth or stable earnings before that moderation. That changes how the eventual decline should be valued.
His broader point is that investors should distinguish between a normalisation of earnings after an exceptional period and a collapse in the underlying business.
The semiconductor cycle may not sustain abnormal growth indefinitely, but that alone does not invalidate the investment case.
Super bearish on Indian IT
Arora is particularly dismissive of the argument that a fall in US AI stocks automatically benefits Indian IT services companies, and market movements would make one believe.
When semiconductor and AI-linked stocks declined, Indian IT shares rally, as traders interpret the move as evidence that AI investment was losing momentum and traditional technology services could regain relevance.
Arora calls this a flawed interpretation. The correction in a stock price does not mean that the underlying technology is losing momentum. If the companies investing in AI continue to increase their capital expenditure, the threat and opportunity for traditional IT services remain very much alive.
He is, in fact, more bearish on Indian IT after seeing the market celebrate the decline of other technology companies. “We are not bearish anymore on India IT; we are super bearish,” Arora said.
According to Arora, the AI investment cycle is still unfolding. The market may be reassessing valuations, the scale of capital expenditure and the returns that technology companies can earn. But the technology itself is not losing momentum and there are pockets attractively priced after the recent fall.
Source:Network18
