The Truth About Indian I.T. & The AI Bubble Nobody Is Talking About | Pankaj Murarka’s Bold View

Hinzugefügt: 2026-05-10

[00:00:00]So if you have to look at midcaps, small caps and large caps for FI27, there is a distinct gap coming in between value and growth and I think that's what we should be looking at as well because so far your the fourth quarter has given you a good beat in terms of numbers overall and we should talk about it as well but what happens in the first half on the back of the impact is crucial. Yeah, >> first half >> let me intervene here >> maybe let me make two points first on the markets uh it's very true >> markets a reflection of the economy >> it's economy if you know if you zoom out and just uh put let's say the last five years in perspective last six years because this recovery started postco uh it's a k-shaped economy there are pockets of economy which are doing pretty well but there are pockets of economy which are not doing well >> and My view is that you know as far as this econom and this is not an India phenomena this is a global phenomena this is something very similar how the economies are playing out in Europe US everywhere else in the world so it's a global phenomena uh and I think this is how the economy probably in this economic cycle one uh conclusion I was looking for a mean reversion but I've come to a view that this is how economies will be over the next two to three years whatever time we have in this business cycle markets are reflection of that market's extremely polarized So on one end of markets you have stocks which are trading at 70 80 100 times multiples and on their other end of the market values become uh even cheaper or what do we call in our uh you know analyst uh language devalues. So uh values become value trap because every time you see a value and you try to uh get into those names it one or two quarters down the stock is down 20%. So uh I think this is how uh and because money changes performance. Uh so this is also leading to you know more and money flowing towards those more expensive stocks because eventually those stocks are delivering returns for you or outcomes for you. So markets are polarized. The index is somewhere in the middle.

[00:02:00]>> No. So the polarization this time also being seen between the smids and the index. Right.

[00:02:04]>> Absolutely. So that's one trend. The second trend which is clearly playing out is uh this is the first time I'm seeing uh just to put in context CY25 India's underperformance versus emerging markets is the sharpest we've had after 1995 >> okay India is not underper India is a lagard market uh when it comes to global markets if you see US markets in the last 3 weeks they've made all all of these markets are now trading at new 52- week highs and we're still away from that And it's the worst actually it's the worst performing market.

[00:02:38]>> Yeah.

[00:02:39]>> So here sorry please finish your point.

[00:02:41]>> Uh I want to go back and relate this one of the biggest driver of the growth and this has uh it's uh I think it has underpinnings in the what's happening in the world. The biggest driver of the growth today is all of us know the elephant in the room is AI.

[00:02:55]>> Yeah. India had similar underperformance in 1995 and that's primarily because whenever you have a new foundational technology which will change the world the first 3 to four years or 5 years uh uh uh a significant amount of investment goes into uh building infrastructure for making that technology accessible.

[00:03:17]>> Mhm.

[00:03:18]>> Internet happened in 93.

[00:03:20]>> Okay.

[00:03:21]>> Yeah. Some of the best performing stocks at that point of time apart from the technology companies which were there were the hardware companies. Cisco >> viewers for context. Pank Muraka started his career as an IT analyst in 94. So just right there. Yeah. Please make a point.

[00:03:36]>> Cisco, IBM, >> HP they were some of the darlings of the market >> and they were all because big beneficiaries of proliferation of internet and PC revolution >> apart from whatever Microsoft and everything else.

[00:03:50]uh and something similar is playing out now. India doesn't India did not have a play in hardware when it uh the internet evolution of internet happened. We don't have a play on AI infrastructure be it the semiconductors be it the hyperscalers or the foundational model companies >> and therefore pankage the question is the point that you made about K-shap economy for different economies if you take the world as a single economy could the K-shap last wherein three or four or five countries and in fact you spoke about the performance we have graphics of that of the performance of various economies or markets South Korea Taiwan maybe US versus India Would the K shape last for a bit wherein India some of the other markets may not perform but the others may continue to perform?

[00:04:36]>> Answer is very simple to my mind. It's very difficult. It's not very difficult to time it. But once the technology is made made accessible this is like you know you building n on new technology available. It's like building roads and highways. Once you build roads and highways only then cars and vehicles can fly on that. Once the technology is made accessible uh the world has to access that technology. The diffusion of the technology has to happen by corporates and consumers. India definitely has a play there. India has a dominant role there.

[00:05:08]>> 70% of market share in IT services.

[00:05:12]These companies will play a very fundamental and very central role in diffusion of technology. So the K will flip the K will hopefully at some point >> there will be a mean reversion but I'm willing to yeah that's the question that I just want to ask.

[00:05:30]>> No I'm willing to I'm willing to what?

[00:05:33]Willing to bet willing >> I'm willing to bet there'll be a mean reversion >> when I think my best judgment in the next 6 to 12 months.

[00:05:44]>> Wow. Okay. Interesting. my best judgment next next 6 to 12 months.

[00:05:48]>> So I have a couple >> because also just a last point >> uh in the whole bill it's always a very normal thing it has happened cycle after cycle when such a foundational technology happens >> the world goes overboard and it ends up overbuilding or overinvesting into infrastructure.

[00:06:05]>> Yeah.

[00:06:06]>> Exactly same thing happened during the internet era in '94 to97. Yeah. And a lot of these hardware companies then uh crash. I was going to ask >> the same thing is happening now where companies are making investments, hyperscalers are making investments.

[00:06:20]Those investments cannot be rationalized basis cash flows and ROIs.

[00:06:25]>> Today those investments are made because no one wants to lose.

[00:06:30]So those investments do not have a rational ROI or cannot be justified with ROI. So you already know that you are at an you know things have gone overboard and one classic meaning uh I follow meaning one of the classic uh uh saying from amongst the 10 commandments of Bob Farrell one of who's very senior Mel guy uh one of the ex-colagues that uh markets have this tendency that the more they go uh go away from the mean they tend to rebound very sharp >> sharp So those last three or six months can be very ferocious but I think somewhere we are in the frag as far as as far as stock markets are concerned on the AI theme in terms of whatever you're referring to Korea Taiwan and all of that >> just on the point that you made right those three companies that you mentioned and specifically Cisco and IBM right for those people that don't know what happened after that and the years that followed just very quickly there was a significant amount of value destruction that took place as well for people that piled on at that time and held for the longest time particularly for Cisco the story that people have told over the years is >> it's very normal in a business cycle for companies to make excess investments >> and companies to go bust and some of those companies the very fact that some of these companies survived means they still had some underlying business a lot of companies went bust during that cycle >> so I'm not even discussing that obviously there are a lot of companies within us beyond this the top five or seven names which we know which have significant leverage on the balance sheet Yeah. And many of these companies will go bust. They'll not be around 18 24 months out. Meaning I I think there's no I I I think we should not even be discussing that. It's a foregone conclusion and it's a very normal thing to happen in a business.

[00:08:16]>> You know, I mean that's a very straight candidate take when it comes to AI. But then okay let's let's take into consideration as a retail investor everyone is still thinking of what to do. Correct? it as a sector if you have to look at Q4 the sector around 12% is the kind of profitability growth that the companies cumulatively have put together and if you have to look at any of the other plays which are your AIdriven plays take into consideration your power sector take into consideration cap goods taking the data center play into consideration power has still somewhere outperformed cap goods not to that extent in terms of profitability would you be then saying that these are sectors which investors should then probably avoid right now because you are seeing some kind of a wave that will come in over the next 12 to 18 months because that's what people are betting big on right now. So I think power and utilities will continue to do well because we also need to understand the context that for a period of 15 years between 2008 and 2021 we did not have a single penny being invested into new power plants in India because in the 10 years prior to that we built so much of capacities >> uh 95% or 90% of the companies in India's power sector went bankrupt in 2010 and 2020 where we had this whole issue about big NPS in banks and all of that. uh so we're coming out of a phase where now actually demand is running somewhat ahead of supply. It's a sector where you cannot bring supply in a short span of time. It's a sector which takes 3 to 5 years to bring supply on place.

[00:09:49]So it's a sector where demand supply is in balance and as India's underlying growth you know sustains across manufacturing PLI and so many other things obviously rural electrification all of that. So supply needs to keep pace with that. It's a sector where we had a lost decade for the sector. So we need to understand that the context from where we coming from. So I think power and have done well I understand but I think they'll do well they'll continue to do well because apart from AIdriven infrastructure which is not a significant contributor to demand growth in India today >> unlike us but there are other drivers of growth and we have a sector with demand supply balance as far as it is concerned. Uh it's really depends on time horizon. If you ask me for any investor who has a time horizon of 3 to 5 years I think okay let me put it this way. I think in the last 20 years probably this is one of the most exciting opportunity or timing uh time at this point of time to be buying in this >> company.

[00:10:38]>> So you are not on the school of thought that Indian IT has missed the bus.

[00:10:42]They're not doing anything exciting.

[00:10:44]They're now trying to clamber on uh and has been left out of the whole race.

[00:10:48]We're not on that school of thought >> on their core business. They are doing what they're supposed to be doing and they're very much there. They've not missed the bus.

[00:10:54]>> So every new anthropic uh update is is not bothering you?

[00:10:59]Certainly not because eventually diffusion of technology for adoption of technology enterprises cannot do it on their own.

[00:11:07]>> Anthropic is saying that they are going to do what IT service companies are doing and their own NLMs.

[00:11:13]>> I wish I wish uh I wish okay I think they understand technology they don't understand enterprise ecosystem how enterprise operates >> enterprise are absolutely paranoid about their data. AI models are probabilistic model. The same question you ask the AI model three times in a day, it will give you different answer.

[00:11:33]>> Answers. Yeah.

[00:11:34]>> Yeah. Enterprise data systems, systems of processes are definite and deterministic models. You need finite outcomes. You know, just think about I'm just saying think about a large bank.

[00:11:46]Let's say a city bank operating in 140 countries. Each country different regulations in each country hundreds of products. Let's say home loan, retail loan, this loan, that loan and all of that, right? It's a very complex system.

[00:11:58]Do you think a city bank will open its system to Enthropic?

[00:12:02]>> Uh I'll give you a classical example which workday CEO gave. Okay. Enthropic launched an uh tool where you can uh run your own HR HRMS software where enterprise can and it's available for free. It's there on Enthropics website.

[00:12:19]>> Enthropic has 3,000 employees and they're using workday.

[00:12:23]First they need to use their tool for the owner employees before >> what is the reason you >> so the workday co is at pains to explain that all the three companies Google openai and entropic they run their HR entire employee management system HRMS on workday >> and each of them have launched a tool where you can build your own HRMS and you don't need workday >> but first this company need to use >> for their own self before you can expect others to adopt >> but they are don't work their systems.

[00:12:54]>> So I think uh you know there's lot actually the problem with markets is twofold. There's in the mark markets are an excellent discounting machines and they're very fast is processing data where markets make an error of judgment is processing information.

[00:13:11]>> Yeah. And the fear of unknown uh because there's so much of information coming the fear of unknown is a much bigger fear than the fear of known. When it's known, markets are very efficient at pricing that. So what markets create inefficiencies and which creates opportunities for long-term investors is when uh when the markets are processing information.