Veteran investor Manish Chokhani, Director at Enam Holdings, has issued a sharp warning on the global AI rally, saying valuations of major AI and semiconductor companies resemble the 1999-2000 dot-com bubble, even as the technology itself is transformational.
Speaking to CNBC-TV18, Chokhani said the market is behaving exactly as it did during the build-up to the internet boom, when capital chased infrastructure creators rather than long-term winners. “It reminds me of the time when your channel started, during the 1999-2000 euphoric period. Telecom companies built the backbone, and at some point the market realised the capex didn’t justify the returns,” he said.
Chokhani pointed out that while AI will fundamentally reshape businesses, stock prices may not reflect that reality. “Someone paying a $5 trillion valuation for what is essentially a capital goods stock is odd. Chips degrade in 3-5 years. This capex cycle cannot continue at this pace,” he said, referring to runaway valuations of semiconductor companies.
He added that history offers a clear warning: the NASDAQ took 12 years to reclaim its dot-com peak after the bubble burst. “Excesses in valuation often take decades to correct. Even Infosys didn’t cross its 2001 peak for almost a decade.”
The bubble, he believes, is already visible in the frenzy surrounding companies linked to AI infrastructure. “The craziness around OpenAI valuations and new entrants like Google pushing aggressively into chips show we are in bubble-like territory. Typically, demand falls suddenly—everyone wakes up at the same time and stops spending.”
But Chokhani stressed that while stock prices may crack, the technology wave itself will be durable. “AI is here—and it’s transformational. But not necessarily for the stock prices of many AI companies.”
Instead, he believes the real beneficiaries will be sectors that use AI at scale. “Just like the internet era—where financial services, healthcare, and retail became winners—the second derivative plays will gain the most. Jio invested massively in broadband, but companies like Zerodha, Groww and Zomato made the real returns.”
Chokhani said investors should avoid chasing hype and instead identify businesses that can harness AI productivity to drive profitability. “The way to play AI is through the users—those serving consumers—not necessarily the capital providers.”
Below is the verbatim transcript of the interview.
Q: CNBC-TV18 is celebrating its 26 years today, but the markets are a little bit subdued. What are you making of the current market setup? You know, it's a little bit worrying, because we had a pretty good GDP print. The nominal number, obviously, was still in high single digits, which is a bit of a worry. But the RBI has pivoted; growth focus is what they're talking about. For the time being, the market is ignoring any kind of good news. How are you feeling about the Indian markets?
Chokhani: It's not a sort of T20 game in a sense. And I sometimes laugh and joke with my colleagues that maybe I belong to the earlier generation which played Test matches where victory was decided over five days, and not as you have in T20, where if you lose five overs you are out of the game. So that's not the case when you're investing. You have to take a longer-term view of businesses that you're owning and people that you're backing. And occasionally, if you get ahead of yourself, like we did, I think, in September 2024 when the valuations were completely off the charts, we spent a year kind of consolidating and still hoping for this growth to come back. But when you see a single-digit GDP print and the RBI has to pivot towards growth, it's telling you that there's a challenge.
And while all of us participants remain optimistic and hopeful about a earnings recovery, it's also a fact that if inflation remains at 2–3% and nominal growth remains at 2–3%, then even getting to a 15% type headline earnings growth is not in the bag, and the market is still trading north of 20 times. And again, the challenge that I know you'll ask me is: at what point do foreign investors come back, and so on and so forth? The reality is, look, everyone is here chasing growth and a higher rate of return. So, people will happily buy Zomato or a lot of other QSR-type companies at 100 times forward on FY27 earnings, but they don't want to buy a 5x P/E PSU bank, for example. So, it's always growth in the context of valuation that one has to think about. And our market needs to digest that a bit—that we are ahead of ourselves in many, many pockets of the market.
Q: We'll come to the pockets where you see some overvaluation still, because it's been a long period of consolidation—14 months for our markets to get to their all-time high levels. But before that, your own verdict and view of the global markets. Do you think this AI-tech momentum continues into 2026? And if that is the case, will India continue to underperform emerging markets like Korea, Taiwan, China, which are seen as AI plays? What is your top-down view for the global markets?
Chokhani: I think the AI reality is a bit like what we experienced, in fact, when your channel started—in the 1999–2000 period, when there was complete euphoria on the emergence of the internet. And it was the telecom companies that were building the backbone. At a certain point, the market just pivoted and said the capex doesn't justify the returns that are going to come out of this space.
And in fact—I may be off by a few months—but I don't think the NASDAQ came back to its 2001 peak until 2013. So, it's not a 3–4-month phenomenon. It often takes decades for excesses of valuation to get corrected. Even our most famous company, Infosys, which peaked in 2001, I don't think took out that previous peak for the next decade.
There are a number of places in the US as well where these five or six companies have taken out the bulk of the valuation for that market. They've, in fact, been underperforming emerging markets as well as precious metals. And it's showing some sort of regime change—that there is inflationary expectation in the developed Western world, and that inflation will show up in terms of higher prices for commodities, and increasingly for things like gold, silver, uranium, copper, and so on.
Plus, emerging markets excluding India have already started outperforming last year, and that trend is going to continue. So, India needs to correct a bit for its excesses on valuation, and also discover its mojo and what's our next big profit opportunity. Because in the absence of creating our own tech, our own brands, and our own way to create the next burst of external businesses, we're sort of trapped in a low-growth, only-market-share-internal-grabbing kind of exercise. So, we will probably tread water for a while.
Q: But just a simple answer: do you think we are in an AI bubble? The setup is similar to 1999–2000 because what followed was the dot-com bust. Are we there yet or not?
Chokhani: So, like I said, it's similar to that period where internet companies, of course, continued to do well. It was the valuations that suffered. And again, it seems to me very strange that someone would pay a $5 trillion valuation for what is essentially—don't hold me to the exact term—a “capital goods stock.” If you're buying chips that will effectively become degraded in 3–5 years, that capex cycle can't continue at this pace. Plus, others like Google have come into the game as well, and the craziness going on with the OpenAI valuations. So valuations are really bubble-like. And typically what happens is that demand suddenly falls—everyone wakes up at the same time and says, “we're spending too much.” So, what looks justified on current earnings is often predicated on those earnings actually coming through. And typically what happens is that these earnings then don't come through, and stocks correct as much as 50–60%, and from there it takes a very long period to then come back to previous highs.
So, I do think while AI is here—and it's transformational for all of us, and we’ll all be living in that reality in the next 3–5 years, very much like the way the internet transformed our lives, businesses, and so on—the same will hold true for AI, but not necessarily for the stock prices of many of these AI companies.
And again, to take that analogy forward: it will be the users of the internet—financial companies like ours, healthcare companies, retail companies—that will be big beneficiaries of this AI boom. The way to play AI is therefore eventually to go to the businesses that benefit from serving consumers, rather than necessarily those who are the providers of capital.
Think of it in the Indian context—maybe Jio made a lot of investments to build out the broadband network for India, and it kept India going during COVID. But the people who benefited from it were not Jio in terms of ROI, but companies like Zerodha, Groww, Zomato, and others. You need to think in terms of the second derivative: where is this ball going?
Q: If you had to give us a couple of themes—you’re highlighting that there has to be some corrective phase in the near term because of valuations, because growth is not catching up, and maybe earnings, as you mentioned the last time, the street is a little too optimistic. But if you have to name a couple of themes that still look well-placed—maybe post the recent correction in a lot of those sectors, whether it's EMS, renewables, defence, all of which saw a lot of interest in the last 12–18 months but have now fallen off the cliff—what would they be?
Chokhani: I buy things bottom-up, and it's not really thematic. When things become popular, they tend to go to excess. And like Dr. Marc Faber used to very famously say: at the beginning of a cycle, the entrepreneur has the vision and the investor has the money; by the end of the cycle, the entrepreneur has taken away the money and the investor is left with the vision.
A lot of that is happening now—as you see companies from overseas selling down stakes, Indian promoters selling down stakes, private equity selling down stakes. And Indian retail has been lapping it up by way of the massive inflows we’ve had. Even professional fund managers have been struggling to justify what they're buying at these valuations, taking cover in the fact that they have no choice but to deploy capital.
So the unloved segments—when you go and invest there—could offer opportunities. For example, I could mention PSU banks as one example. The finance minister herself has said her attention will be on privatisation. We also know that this whole trade shifting away from tech to hard assets is a decade-long trade. You've already seen the performance last year of gold and silver, it's happened with uranium, and copper is zooming. That's a trade that will take place. And starting from—not exactly depressed valuations—but relative to the rest of the market, they look very cheap.
And I’ve spoken in the past about winning when you get asymmetrical bets. If you think in terms of: is this a 5x P/E stock five years out? If you can't look out to, say, 2030 and you can't predict earnings of the company, you have no business buying it. But if it looks like it could be a 5x P/E stock by 2030 based on the price you're paying today, the odds are that you reach a 15–25 multiple by that time—which means you make 3–5 times your money, which is a reasonable rate of return over a five-year timespan.
So, there are enough things to do. I don't mean to say the entire market is overpriced. But when you're buying stocks at 50x or 100x P/E, your execution has to go completely right, and your macro has to stay completely right for you to make money. Being a value guy, it's been harder for us to hold on to stocks which we bought early, but we can't hold on until they go to 100x. So, we are rotating more towards what I call the value side of the market.