Are telcos making a mistake by divesting their data centers?
The early years of last decade saw telcos rapidly build data centers, hoping to scale co-location services. Fast forward 10 years and almost every telco is now divesting this business. Telcos want to focus on core services while letting others manage the data centers for them. Is data center a difficult business for telcos to sustain given their steep CAPEX commitment for rolling network infrastructure? Or are they making a mistake now by divesting them?
Speakers:
- Sachin Mittal, Head of Internet & Telecom Research, DBS Bank
- Manoj Menon, Founder & Managing Director, twimbit
- Inderpreet Kaur, Research Manager, twimbit
Key Topics:
- Role of data centers in telco cloud and edge cloud?
- Is divesting the right way to treat data center assets?
- Should data centers to be left to specialists?
- Can telcos pull off data center transformations to create a valuation arbitrage?
Time: 4.00 PM SGT/ 1.30 PM IST
Date: 14/7/2022
Transcript
Manoj Menon
Okay, fabulous. Thank you everyone and good afternoon and good evening and welcome to this Twimbit immersion on our telcos making a mistake by divesting data centers. I’m Manoj Menon, MD and founder of Twimbit. And today I have two wonderful guests joining me to have this conversation. And I have with me interpret Carr who is my colleague, and she’s the research manager in telecommunications. So welcome into PRI.
Inderpreet Kaur
Hi, Vanessa, thank you so much. Excellent.
Manoj Menon
And joining us for this conversation today we have a special guest such an Mittal. Such an is with DBS. I think many of you have heard such and in some of the previous conversations that we have had, and Sachin is going to join us in this conversation because, you know, we have been debating this topic, there seems to be such a huge adoption of public cloud in this region in the last couple of years, the hyperscalers have built significant points of presence in many countries in Asia. And, and that is also fueling demand for data centers as well. The growth overall intensity of digitization should also drive demand for data centers. But at the same time, we see that many telecom service providers are divesting their data centers or spinning up them into independent entities. And to in order to understand that better we thought, hey, let’s bring together some experts who can drive that conversation. So delighted to have you here. Welcome, Sachin, thank you for joining us. Thanks, Manoj. Excellent. So maybe we start firstly, you know, first a key question, you know, is, is firstly, overall data centers, isn’t that a great growth opportunity, not only here, but globally as well. Such, maybe that could be a great starting point to drive this conversation.
Sachin Mittal
So every business when these cycles, okay, so what you’re telling me is definitely true. For the last five years, the data center has been a great opportunity. And now we are talking that it has become a selectively it’s a selective opportunity, it’s still a growth opportunity, but it’s for selective segments. So I think retail, as a segment has become a slower growth, we are talking about probably mid single digit growth for retail segment. And for the wholesale, it has become overly dependent on the hyper scalars. Now, and and the reason you mentioned that cloud. So if you look at the top three players, there’s a one 40,000,000,001 40 billion in annual well made of three players growing at more than 40% Each year, these guys are telling everyone, why do you need to own it equipments. You don’t need to order it equipment, not just from cost point of view, because refreshments and all the software updates they do they make it very, very competitive. So essentially, if you have lesser need to own it equipments, you’re listening for data centers. So I think this is my take that retail is slow growth segment. And Wholesale is more or less depend on hyper scalars.
Manoj Menon
So then what percentage of this total demand in the global data center is, let’s say driven by hyperscale? Is today? Is there a number? Is there a statistic to
Sachin Mittal
it? I saw a report by Twimbit. I think you guys have fun with those numbers. If I recall that that’s almost 7060 to 75% of the of the wholesale will be on will be governed by hyperscalers. Right? Correct. Okay. That’s our alliance. Right.
Manoj Menon
Got it. So clearly, you know, the hyper scalars are gonna drive a large part of the demand. So what does it essentially what does it take to win center part? Is it that you need specific capabilities to cater to that demand are hyperscalers you know, while some of their building some of the capacity themselves in many cases, they’re going out and partnering with other third party companies and and you know, leverage in their capacity as well? Yeah.
Sachin Mittal
So if you do ask me, I will say three factors. First is large space in a given location, because They at least need five to 10 Mega Watt minimum for hyperscale capacity. Second is, I think the second is a bit new, but it is a renewable source of power. If you if you look at and notice the announcements, public announcements by AWS they want to have renewable energy 100 per sent by 2025. For Google and Microsoft, they announced 2030 carbon free, they are under a lot of pressure, because these are in a power hungry data centers, renewable sources very important. Third is a regional presence. I don’t want to have a separate agreement each and every time the separate provider in this country, I want to have probably a master agreement. And that agreement allows me to expand the capacity as in when on demand because cloud demand itself is very volatile. So whenever they need it, they can ramp it fast. I think that digital presence becomes very important. FIFA
Manoj Menon
three critical three critical factors driving driving that. So within within that context, you know, given that this context that the market demand is driven by the hyper scalars and the telcos themselves. The telcos themselves are slowly beginning to adopt high, you know, public cloud, they are migrating many of their own workloads to the public cloud. So isn’t that you know, so is that a clear sign that telcos should be looking at divesting their data centers and start, you know, working in greater partnerships with these public cloud providers. And is that is that making this trend of telcos selling and hiving off some of their data center assets in need for capital to fuel their 5g investments?
Sachin Mittal
It’s definitely so you’re you’re totally right, that if you are a small taiko in a given country, and you might not succeed in attracting these hyperscalers that’s a valid point. And it is better for you to use no two divers. Today, you can sell your data centers for only 25 times EBIT, da, that’s not bad. telcos traded below 10 times, you know, five to 10 times EBIT, da. So this is a business for symbol EBITDA, you can get much higher multiples, okay. And you can use that money either for the 5g CapEx for your Reno making your shareholders happy with dividends, or other alternative opportunities, growth opportunities in the telco area. And, but if you are a large player, with the regional presence example, we be SingTel, they have decided I have partners in all the countries, I can work with them. And even that actually, even telco partnership is not important is not sufficient. In fact, they have gone in partnership with places like Gulf energy. So you need to have a very, I will call it a very dependable renewable source of power. I think that has become extremely critical in the last one to two years. Five years back, no one talked about it. Now, everyone, everyone focuses on renewable energy.
Manoj Menon
And so maybe the next question is, you know, we hear you know, with the growth of 5g, we see that maybe something like edge data centers could become important. And given that, you know, it is likely to be some time away before the big hyperscalers can build edge data centers in every, every city and let’s say in many parts of Asia. So isn’t that a logical place for the telcos to play and and start participating and, and building that play rather than exiting the market by selling of data sets.
Sachin Mittal
So it’s a valid point. But remember one thing, please, that 2012 or 13, we started with 4g, and even today, many developing countries don’t have 4g reach into each and every nook and corner. And that was a very good technology, a very capex efficient technology. And 5g is not even considered a CapEx efficient technology. It’s a good technology from latency point of view. But it’s an expensive technology. So you can assume that probably it will take more than 10 years for us to see a lot of 5g Rounders, it’s flexible. But in selected countries like Singapore, Hong Kong, I know they are clearly ahead, but rushed to the region will take time. And do you have that luxury of time wait for another five, six years, waiting for those edge data centers and how the technology evolves. So my sense is telcos need that money for that 5g itself, rather than waiting for the play on 5g.
Manoj Menon
Yeah. So you think it is a logical step for the telcos to to aggressively look at, you know, maybe raising some capital, using that capital to invest in build their enterprise capabilities. At the same time maybe for the better partnerships so that you can offload more workloads onto the cloud and in return be a reseller of some of the public cloud services to enterprise customers.
Sachin Mittal
Yeah, display to your first point that how do you work with these cloud players? Because cloud players are managing I mean, they they also need some resellers. They also need the people who can Will apps on their on their cloud services? So I think that opportunity is still there and will remain a long term opportunity for telcos and everyone else. And it’s better to focus here.
Manoj Menon
Correct. If you want to comment, Christian, from what you hear.
Inderpreet Kaur
Yeah, I mean, if I removed the data centers, or particularly in other day edge data centers given, there are so much infrastructures that telcos already have their central offices, all the pop infrastructure that they have, if we removed that, what else is remaining for a telco in an edge play? Because if you look at the upper stack, you know, a cloud player already has the applications that have to run on top. It already has the platform that has to work in an edge data center. So if telcos today, say I’m going to divest all of this infrastructure, are they hampering their long term view in in, you know, the edge compute opportunity? So rather than looking at divesting all of this, should it be, you know, even a thought for them to consider maybe, you know, carving it into a separate company or in separate entity, but still retaining control of those assets?
Sachin Mittal
It’s a good point. And I agree that some telcos will definitely who, who don’t need so much of capital, they can split the data center business listed separately, if it’s sizable, but you know, the condition is it has to be sizable business of decent size, and it can be listed. And many of the telcos don’t have such a big data center business, and it will be too small to be listed and split as a separate business. So while while you’re right, in terms of intent, and the process, the question is, not everyone has that size of the business of data center. And that will be a barrier here. But we should not forget that enterprise business is opportunity for telcos with or without cloud, because cloud players cannot talk to local enterprises, and medium and small enterprises. And even the bigger ones in each and every country, these are global players. Okay, so they need someone to talk and ultimately give the end application to the enterprises. That’s point number one. Point number two, one of the problem with the with the this whole hyper scalars is that there are only four or five hyper scalars in the whole world. And if these guys are renting the space from you will only dictate you the price will have to dictate you how much I’m going to pay you for this megawatt. So this is a concentration of customers, right. And that is a serious issue here. So I will not be very clear. I mean, I think retail is the space, at least there’s a diversified customer base, that it may not be very high growth opportunity, where at least the customer base is diversified. It’s a very concentrated customer base.
Manoj Menon
And, you know, in many parts of Asia, to be fair, we still do not have hyperscalers having their own data centers and points of presence in many of these markets. So as a result, you know, it will be some time before you know, so there is a role for telcos to continue to have. So for example, Philippines today is not served by you know, any of the hyper scalars directly, you have to either opt for a Singapore or no one of the other locations. So, so maybe the strategy will vary based on which country you are in, in Asia, right? Clearly markets like India, and Australia and Singapore and Indonesia, are big markets. But then as you look at, you know, distributing your risk across multiple providers and adopting a multiple cloud, so you would, you would still have some relevance, but maybe you will not be as competitive as the hyper scalars.
Sachin Mittal
Yeah, that’s a good point that you brought up just now supply, I think, in many countries in Southeast Asia, whether Singapore, Hong Kong, due to due to partly due to COVID. And partly due to regulatory, I think regulators focus on power efficiency, you know, that supply has a little bit lagging behind. So why, you know, demand demand is, is probably maybe a bit slower than expected, because of the cloud. So it’s a supply and that actually means that in the next two to three years, we are not seeing immediate correction of the data center pricing or the other living rates. So, there is no there is no such disbalance as of now, because of this COVID issues which happened oversupply Yeah.
Manoj Menon
Okay, fabulous. The other question is as telco as you know, on one end we’re seeing telcos should partner maybe become a reseller of some of these hyper scalars. At the same time, you know, you also cover the telcos extensively. You you put your buy or sell recommendations on these telcos, the telcos are beginning to, you know, from a high beta margin business, as they get to start reselling public, hyper scalars business. This clearly has implications on their overall margins. So how do you rationalize it? And what is your you know, what do you expect these telcos to do? You know, as they start addressing this new opportunities, which have a relatively lower EBITA margin as compared to the traditional businesses.
Sachin Mittal
So actually, you’re right, this was an issue five years back when when telco margins were even higher. So they were so high that anything in any every business look lower margins to them. And but now, I think telco margins have also in the last four, five years, most countries have seen a 4g new entrant whether big or small, and telco margins have corrected a lot. And now they have come to a level if you look at the investment, because the capex involved right even for to get that margins of 20 25% There’s a CapEx to be put into the network’s while all these reseller businesses, they will not require that much of network capex, they will have a low capex, they will it will be more OPEX which is factored in the margins, right? The operational manpower expenses. So, if you calculate the ROI, see return on invested capital, actually could be quite similar, or even better for these new businesses because they have higher growth profile. So, looking at the investments, it doesn’t look bad, but that case was not there five years back that I agree,
Manoj Menon
interpret anything else.
Inderpreet Kaur
I mean, I completely agree with Sachin on the, you know, themes of having scale, you really need to have scale to have these business or the data center business run. And as well as you need to have, you know, regional presence, if not a global presence and that we see like NTT and someone like China Telecom is still successful, they do have their data centers across you know, most operation countries, but yet, you know, if we see another trend that is emerging, the primary market markets are established, but then there are these new emerging secondary markets, for instance, that we take example of India. Recently, there was a partnership between Adani group and edgeconnex. And likewise, you know, we see so many developments happening in Indonesia or even in you know, Malaysia. So, is, is that still an indication that you know, there is still some merit even if you are the local player, but you have large enough scale, you know, to support that data center business, particularly looking at the Indian market or the secondary markets, such and how do you see that?
Sachin Mittal
So, I think, if you are a greenfield operator, and you are beginning to focus on edge opportunities today. I think that’s fine, because, you know, you don’t have any other businesses to take care of. That is not the case for telcos. You know, can telcos have existing commitments to the shareholders? And, you know, so they’re in different ballgame. So I agree with you that, while he will be small opportunities, it will take time, if you can wait for that time. And, you know, and that’s the business, which has a very small base anyway. So you know, it will look like a good growth opportunity. So I think as a Greenfield, it’s fine. But as a, as a business as a site, additional business, it doesn’t look that attractive. I think that’s a key difference between telcos and someone like Danny, because if I if I hear correctly, Danny is going for Danny group is going for enterprise 5g in India. They actually are the news, public news, was it two days back? They are focusing on a race for 5g for enterprise business. So that is, so they’re focusing on this whole niche of Edge Plus enterprise 5g. That makes sense, okay. And the question is, how much how much capacity you have, how much can you invest? So every player will have different capacity and according to a balance sheet,
Inderpreet Kaur
so So if we hear it right, if you have the capacity, you know, to keep these investments going, telcos should still own the data set. First, they should
Manoj Menon
have capacity and scale. So the summary is your capacity and scale, and you continue to play in it. If you do not have capacity and scale, it’s best to partner and get deeper into the relationships and becoming a solution provider.
Sachin Mittal
That’s true. Yes, that seems to be the key message. But remember that many telcos are merging in many countries, because they don’t have the capacity for standalone 5g. Correct? Correct. Yeah. Yeah.
Manoj Menon
Excellent. Excellent. I think, pretty much, you know, that summarizes of a conversation for today. I didn’t know is there anything else? Such in you wish to add? Are there other players who, you know, what is your proposition to other players, let’s say, your player data center players like Kleenex and others, you know, in this region, what is the outlook for companies like them, you know, they can continue to obviously serve the big hyperscalers. And they are becoming a big part of the customer demand for them, do you see them continue to thrive in this part of the world?
Sachin Mittal
So, exactly, that is what has happened that in the retail space, are pleased like Equinix had depreciated themselves. One because of global size one because of your interconnect offerings. So I think selectively plates these are these are very good players, they can expand. No, these are, I would say leagues apart from many of the tapes, they will continue do well, I’m not taking a view on the share price. I don’t cover them, you know, how much is priced in how much is not? That is for the market to understand.
Manoj Menon
Okay, fabulous. Excellent. I think if there are no other questions, then thank you very much searching, I think, you know, your your response and outlook is very, very clear. You need scale. To compete, you need capacity, while clearly there is a huge opportunity in the business is about serving the large, hyper scalars efficiently because that is where the bulk of the demand in the next couple of years or next four or five years is going to be over a period of time. Yes, there will be demand for edge data centers, and you have the capability to ride through the next five years without a significant need for capital, then you can continue to play in that space. That’s a that’s a fairly good summary, as would you say.
Sachin Mittal
Excellent, excellent.
Manoj Menon
Okay, thank you very much searching. And thank you very much in the feed for that conversation. Appreciate and if there are other questions, feel free to reach out to us, and we will be able to respond and support you guys. Have a great rest of the evening.