FY 2026 earnings call transcript

I know I'm going to pre-empt questions

how come this $9 billion is not going to go up?

How come you're not improving your VRD? We'll probably answer those questions later.

Ok?

And then the final slide is really a quick wrap up before we open our questions. Don't spend too much time. I think business performance you've seen the track record, we’ll continue to deliver that.

We will continue to actively manage our capital at $9 billion. Which has given us a very strong underlying NPAT growth. And then this allows us to declare a dividend of 18.5 cents, which is the highest ever in Singtel's history.

So with that, I'll hand it back to Lian.

Ms Lian Pek, VP of Investor Relations and Group Corporate Affairs (Host)

Thanks, Arthur.

We'll be taking questions now. I'm Lian from Corporate Affairs and IR.

For those present physically, do raise your hand to indicate your interest to ask a question.

Online participants can also use the raise hand function to indicate your interest.

Both cases appreciate, if you can wait for your name to be called out before making questions. And may we also ask that participants keep to a limit of two questions each, and we can double back if they're more later.

10

Piyush.

Mr Piyush Choudhary, HSBC

Thank you. Thanks for the presentation.

Firstly on the guidance, so just to clarify on the RE:AI. The revenue potential is $300 million annually, right? That itself will give $60 million of EBIT which is then 4% growth. So in that context like the guidance and you have DC Tuas getting active, Optus is improving, NCS is growing at the low single digit to mid single digit guidance, like what are the considerations if you can share that? That is first.

Second is on the asset monetisation. This morning we have said about Optus minority stake potentially. Airtel we know. There are other assets in the portfolio. You've already done $6.5 billion, so why not increase that $9 billion?

Mr Yuen Kuan Moon, Group CEO

Ok, maybe first on the guidance, I will explain a bit more. If you look at the DiCo first, because since you talked about RE:AI, you talk about the potential. We are first year commercialising RE:AI at scale in Singapore. So yes, the business cycle for RE:AI is shorter than data centres, but it's not immediate. You still have to build and assemble, and you will take 6-9 months to do that. So the first part of it is really about investment, and you may get some tail end, just like the Tuas data centre where you turn on the last quarter, you get some upside, or the EBITDA, the EBIT will flow through. But the first part is really investing and building. So that will take time. The good, but the good news is of course 80% of that is actually contracted. So there's also the potential upside of 20% once we turn it on.

So the second thing is also on DICo, if you think about the Tuas data centre. While we turn it on in January, we only turn it on floor by floor. Depreciation hit immediately for the whole building, the whole building. But our contracts is coming in floor by floor, because we are not turning on everything.

So you, so you get an immediate hit on depreciation for the entire building. But the floor by floor turn on is for the rest of the year. So, you will not see the full impact of the margin of the data centre immediately until the whole data centre is fully turned on. That means every floor is turned on, and we have also announced that we are almost full, contracted over 90% of the data centre. But we have not turned on over 90 % of the contract yet.

So again, you're not seeing the full impact of the margin flowing through in FY27. So, the next year you will see the full impact of the Tuas DC, right. So similarly, you can say the same for our regional data centre. They are all turning on stage right, depreciation hitting and, and the, the full margin is not being felt, and we continue to grow.

So you see that we are tracking that sort of progress, but it's not immediate. And we always say in Nxera, if you look at our EBITDA about $150-160 million previously. So you're looking at doubling it in the short term, on a stabilised basis, but it's not stabilised yet. Yes, OK. So that's really on the margin.

But obviously the second thing is we are also looking at some of the second order impact

11 on the Middle East crisis. Because first order, like I say or fuel prices or energy prices in

Singapore, Australia, we have got a bit of a long term power purchase agreement. So that's protected, mitigated. But eventually those contracts, you know it's not very long and you will have to renew and that's energy costs going up.

Secondly, more importantly is business and consumer demand, will it slow down? Because of that crisis, because of economic conditions is tougher, people may be tightening their belts. So it's still unclear. We'll see the next few months how it plays out.

And then finally, of course we were expecting market consolidation in Singapore, right. This year because it was almost done and you know, just last week we have got the new news.

So we will evaluate what's the situation. The market is still probably going to consolidate, but it may be another six more months, nine months before it consolidates. So it's still a lot unknown in quite a few of our businesses.

And the last one on the EBIT guidance is really about Optus. Optus is still, of course you see some positive momentum in the last quarter as well, but Stephen will be focusing on also building more resiliency, compliance of the Optus business while it continue to grow its revenue.

The good news is, we still see a 4% mobile revenue growth last year. So as price rises, they continue to price up and expect to see that. But it will be sustaining that momentum we've seen last year.

Because of all the unknowns, you'll see that our guidance is a bit more cautious. But if you look at our track record, we always will look at our guidance at the mid-year mark once we see a clearer picture of the Middle East crisis and the competitive landscape in Singapore.

On the second question on asset monetisation, I'll hand over to Arthur.

But first of all, I think you know, we have not even completed our $9 billion. So we still have

$5.6 billion. We still got a few billion to go.

So don't worry when we’re hitting closer to that, we will review. You know we have got more assets than the $9 billion. And when it comes, we will talk about it. There's no hurry because we still have headroom. We still have not fully crystallised the entire $9 billion yet, right?

So, as and when we are ready, we will update, you know, what's the monetisation or recycling of our capital. But what is important is the slide where you look at the $9 billion.

Can you go to the slide?

If you look at it, the $9 billion, about $5 billion of course we said is reserved for the VRD for the next 5 years up to FY30. And you see we have only used a portion of it because we didn't go to the maximum of 6 cents. This year, we declared 5.1 cents. So we still have got headroom to continue to support the VRD. And the VRSB, the share buyback, we only have committed, we only have used $200 million. We still have another $1.8 billion of which we say FY27. We will look at another $1 billion and FY28, another $800 million to $1 billion. So divide by two, right. So there's still capacity to do that.

12

And of course, the STT GDC we put in $740 million for that 25% stake. So the headroom is that, right? I mean unless we are running out of headroom, then you see after coming in,

OK, it's now not $9 billion, but it's another X, whatever X.

So, you know, wait for it, it's going to come. But when we are closer, we will tell you more.

And Arthur, maybe you talk about asset monetisation.

Mr Arthur Lang, Group CFO

Sure. So I want to add on to the EBIT guidance, I think it's also worthwhile noting that at this point in time, it's very early into the year, we only passed one month. I think we're seeing quite evidently that, in terms of second- order impact, telco services continue to be seen as an essential service.

You have heard our associates talk about it. I mean so far what they have shared is aligned to what we are seeing, what we have decided given that we are early in the year. It is better to be conservative because there is also an element in the markets, not just with Singtel, a bit of euphoria, right, in terms of the markets and what could happen in the Middle East.

So, we have decided this is the stance we are taking. But then again, as Moon said, middle of the year, there's always an ability to revise, if we need to.

Coming back to this, asset recycling. One more point, yes, VRD, we've paid 5.1 cents.

There's 0.9 cents headroom. Our core dividend, you've noticed we've paid 80% of between the 70 to 90% range. It's a big if, and this is not guidance. If all hell breaks loose, our underlying profit actually drops, right. We can still maintain the dividend because we've got a 10% headroom. So the headroom not only in the VRD but the headroom also in the core combined with potentially have a headroom of about 3 cents.

If you think that just fundamentally our dividend 18.5 cents dropped to 15 cents, we can still maintain an 18.5 cents next year. This is the worst nightmare scenario that we're talking about, OK? There's bigger problems to have if we really have to do that. But I just want to put things into perspective on this, right? OK, then now coming to this $9 billion.

Now we mentioned we announced the Optus transaction. To be very clear, we announced it because we are a Singapore listed company. We need to do it for SGX requirements.

It is not because next week we are ready to sign the deal, right. If we are ready to sign a deal then maybe we can consider upping the headroom, of $9 billion.

But we are just simply saying we are open, looking for Australian, we're looking for minority, and three months later we might have egg on our face and basically said no deal because maybe the terms are not right, valuation not right. But at the same time we're not kind of announcing this because we're not starting from zero discussion. But at this point there is no certainty there will be a transaction.

So, it is only prudent not to raise this $9 billion. In any case, there is the headroom.

Then you have all the assets. It's, again I want to reemphasise it is not just Bharti Airtel

13 that we are recycling, it can be other things as well. Many, many different kind of listed companies that we own, as well as some non-core assets which includes real estate and various things. So, still sticking to $9 billion for now.

Ranjan Sharma – JPM

Thank you for the presentation. I have two questions. Just want to dig deeper into what has been discussed already.

Firstly, on your dividends versus cash flows, right. So if you look at your operating cash flows of about $4.9 billion, I guess that number is somewhat the same. This financial year as well. Then you have CapEx of $3 billion and lease payments and interest costs of around

$900 million. So we're left with like a billion dollars of free cash flow. versus the dividend of at least $3 billion, plus a backup on top, right. So that's a $2 billion plus shortfall, yes. So how will that be covered between asset monetisation or debt, if you can just try to get into more details.

Mr Yuen Kuan Moon, Group CEO

Okay. I think first of all the CapEx $3 billion, like I said the growth CapEx, if you look at the earlier slide, right, it's not entirely all from cash flow. Of that $1.2 billion, $700 million is already fully funded by equity for example, from KKR, customer prepayments, and it doesn't include some of the debt or project financing debt, right. So it is not the entire $1.2 billion that will come from our operating cash flow.

So similarly, if you look across the other parts of our business, we do have got debt capacity that we look at individual operating units. But Optus has its own debt to look at. Our finances is, is some of the CapEx. So it's not everything entirely coming from operating cash flow.

Mr Arthur Lang, Group CFO

So the CapEx is not this $2.5 or $3 billion is not all equity, 100% equity. I mean, for example, if we build a data centre. You probably lever it up 60% LTV.

So to your question directly, you have the operating cash flows which includes our regional associates, that one is effective. A good proxy is underlying profits. So we're paying 70 to

90% of that. Then there's the VRD, and VRSB as well. That is coming from our asset recycling, right. And again, we believe and we still are of the view, we can maintain 3to 6 cents up to 2030. Just looking at what we have said in the markets, right, that is a ready pool of capital that we can tap on, but we don't need to do everything today. So that's how we look at it.

That's how we make this dividend sustainable. The core dividend has to always be 70 to

90% of our underlying profits. And then the VRD or VRSB, think of it as a special dividend, but it's a programmatic special up to 2030. After 2030, we have also said that it is our intention to grow our underlying profits to a sustainable level such that we can wean investors off the VRD after 2030.

And that's why we are investing in things like RE:AI, we are leveraging our digital

14 infrastructure platform. We're thinking about possibly a REIT IPO that could generate management fees, right, if we manage the REIT, the asset and things like that.

Mr Yuen Kuan Moon, Group CEO

And of course the underlying business like NCS, Optus will continue, so growing the underlying profit of the OpCos is very important.

Ranjan Sharma – JPM

Thank you for that.

My second question is on, if you can get more into the details of RE:AI, we talked about unlevered IRR of I believe low teens. You're showing up to $600 million of CapEx. Also, $600 million of revenues has been contracted for three to five years, right. So how do we get to that low teens unlevered IRR?

Mr Arthur Lang, Group CFO

How do we…as in too low or too high? Or?

Ranjan Sharma – JPM

You have a CapEx of up to $600 million. You're also looking at about $600 million in contracted revenues over three to five years. I'm assuming that's the total for the time, not annual.

Mr Arthur Lang, Group CFO

Yes, total for that time.

Ranjan Sharma – JPM

It says 80% already contracted. So your IRR revenue is going to be much lower at 80% contract rate. So how do we get to the low teens?

Mr Arthur Lang, Group CFO

OK, the IRR is on a few things. First, is cash, right? So we assume up to $600 million of

CapEx upfront, but it also means, as you said, the customer cash flow, that means the advanced payments. So, the net CapEx may not be $600 million, if you know what I mean, right. So, we cannot share this, but hypothetically, if the customer says, you know, you spend $600 million, we'll pay you in advance 300 million. So, our net CapEx going forward is only $300 million. So that's how it says…

Mr Yuen Kuan Moon, Group CEO

Again is not cash CapEx, right? I mean, the CapEx is not all, not all equity.

Mr Arthur Lang, Group CFO

15

OK, we assume it is here. That's why we look at it unlevered. But then of course, on top of that, we may get some bank financing on top of that.

Mr Bill Chang, CEO Digital InfraCo

So, the $600 million includes fitting out the DCs, which would go beyond supplying more than the $600 million, right. Because you got to fit the DC for liquid cooling for more GPUs to come in that is basically going to be more than the $600 million later. So, you know that part plus upfront payments, you don't see that. That's why IRR in that sense.

Da Wei Lee – Morgan Stanley

I just have one question on RE:AI. Can you talk about the unit economics? Help us understand the volume and pricing, how many GPUs you bought, how should we think about the pricing dynamic? You know, with pricing coming down and how that's affecting your future growth rates if that happen, and margins looks to be like you said 20%. In fact, I think, in your disclosure its 29%. What kind of useful life in terms of depreciation are we using?

Mr Yuen Kuan Moon, Group CEO

So maybe the best way, Bill, you can about using the pilot as an example. That’s already nailed in, and then we can extrapolate and you can do your own analysis.

I think first of all, the big difference is the 1MW that we deployed $60 million is the different chip, right? You talk about the H100.

And secondly, I think if you look at it, we have taken a very conservative approach in terms of the depreciation of the business, which is really the contractual period of three to five years. We have not taken into consideration the residual value of the chips beyond that.

So, of course we do not know. I mean we want to make sure that this starts to run three to five years before we look at what's the residual value of the micro chip. That's, I'm quite certain that there is residual value, but it has not taken that into the modelling here.

Mr Bill Chang, CEO Digital InfraCo

We started with the H100. You know, 512 since 2024, those are fully subscribed today and will be fully subscribed to sort of what I see as the end of life in a three-year contract with these customers.

We then deployed GB200s, last May we announced that. And that's taking up now to the

1MW. We have secured contracts that would take a large part of the 11MW now to be deployed, and that includes latest GPUs including GB300s and we're preparing for Vera

Rubin in the pipeline. So we're innovating and transforming the DCs to go and carry that.

So, useful life, we basically price in three to five years contract tenure. And whether it's a three-year, the fully depreciate or five year depending on those customer contracts. And what we believe, although we've not built in our model now, it's after fully depreciated let's say a three-year, because we ran our first GPU on a 5G edge computing, A100s back in

16

2021-22 and that's already five years, and they're still running the A100s. So if I use the

A100 as a proxy that we have actually in our edge facilities, y the H100s can go beyond three years, we believe. Because we are deploying all liquid cooling, especially the lifespan will probably be better, because if you think air cooled they operate much high temperature and therefore the lifespan they say and higher failure rates. We are operating all liquid cooled. So the liquid cool, we believe we could actually take it and so we have, in this model when we have all those fully depreciated, those GPUs are now up for recycling.

And the recycling has got value beyond that. So obviously we're keeping that aside first until we get to that.

So it's a three to five year basis. And from a pricing standpoint, depending on the GPUs, different prices for different GPUs, right? And whatever you see up there is a net result of pricing, largely the H100s and it's coming with the GB200s blended in already.

Da Wei Lee – Morgan Stanley

So if I can follow up, so the pricing, is it fixed for the full three years? Are there any escalations, or?

Mr Bill Chang, CEO Digital InfraCo

Basically, it's priced fixed with that tenure of three to five years depending on the customers’ energy pass-through. We've discussed this and energy pass-through is because we want full transparency with those customers. And also because it's a Scope 3 issue that we don't have.

Mr Yuen Kuan Moon, Group CEO

I think energy pass-through is very important, especially in the current landscape. So we don’t want to take the risk of energy price up or down, the client actually takes care of that.

Mr Arthur Lang, Group CFO

And to be very clear, this point is a very important point. The contract is priced as if after year three, the value of the chips is zero. But in reality, we can always recycle the chips for something else.

Mr Yuen Kuan Moon, Group CEO

As I said, the business model today that you've seen here is more conservative. We are assuming zero residual value. But as what Bill have said, even in the early deployment there, there are some residual value, but we haven't got to that stage yet. So we do not know how much is that worth. And the chips are actually evolving very quickly.

So we've decided to take a more conservative approach in the modelling and said we will depreciate it fully across the lifespan of the contract.

James Druce – CLSA

Just following up on a couple more RE:AI questions. Firstly, just on the 11 megawatts that

17 you have contracted, are you able to share the nature of the customers that you've engaged? Are they governments? Are they hyperscalers? Are they enterprise? Size of the deals? Are they half a megawatt? Are they 5 megawatts?

Mr Bill Chang, CEO Digital InfraCo

I would say it's a combination of public sector agencies and enterprises. Not hyperscalers because of the limited energy, hyperscalers might buy a much bigger megawatts. So we don't want to be reselling this. We are really catering to end customers directly, and they include public sector agencies and enterprises. Enterprises in a few sectors and research.

You know, we just announced with what we're doing with the IHL including advanced research, sensitive research and all that. I think it's those 3 segments, how many megawatts and all that, they range from anywhere from more than 5 megawatts to those that are like 250, 150 kilowatt, the full range of that. And then GPUs range from H100s to

GB200s, soon GB300s and Vera Rubins are all in discussion, and we're preparing our rollout.

Mr Arthur Lang, Group CFO

And Bill, it’s fair to say 11 MW is constrained by power, not by demand.

Mr Bill Chang, CEO Digital InfraCo

Yes. The 11 MW is constrained by power. And obviously whatever we're thinking of in

Nxera, in our data centres, whatever power we will allocate to RE:AI first.

Mr Yuen Kuan Moon, Group CEO

Of course, the 11 MW, we have already secured the power.

Mr Bill Chang, CEO Digital InfraCo

The other piece is obviously we're thinking of, like what Arthur said earlier, regional expansion now, because the demand is high and customers love our product. If you see what is Frost and Sullivan's recognition as the best GPU-as-a-Service provider in Southeast

Asia, it's RE:AI. That's an independent report from Frost & Sullivan. And Morgan Stanley also said, you know of its capabilities.

Mr Yuen Kuan Moon, Group CEO

So we will… of course this $600 million of CapEx are all purely in Singapore. We are not giving any guidance on beyond Singapore. We got to get this right. You know, just like what we have done in the past in piloting, even if we go overseas, we will also do the pilot first because the operating model will be very different for different countries. So we are not jumping straight into it. So I just want to make it clear that this projection here is really for

Singapore, sovereign AI.

Mr Arthur Lang, Group CFO

18

Sorry, on that point about reselling direct to our customers, not through hyperscalers.

Hyperscalers always will cream off some margin as you all know. I think the second point is, by selling to enterprises, we can give, or RE:AI can provide a full suite of services. So it's not just the RE:AI service. You got the DC house, the RE:AI. If they want connectivity, then we've got the dark fibre and all that. So it's a whole solution that we're delivering to our enterprises.

Mr Yuen Kuan Moon, Group CEO

And this is standalone business from data centre. So basically RE:AI will have to rent data centre if they want to build more beyond this. So then data centre will have its own business case of renting the, the data centre capacity.

James Druce – CLSA

And then one follow up, if you, if you are entering into a three to five year contract say, how much of a discount do customers get off the spot rate that you can see in retail.

Mr Bill Chang, CEO Digital InfraCo

At the moment when we say 80% contracted, more than 80%, actually it's 100% deployed today, all based on contracts. The reason why we are just using that more than 80% contracted is to give us what's called float capacity, flex capacity, to acquire new customers.

That flex capacity when we have available data centre, when we put this up or we go to a region we have debt, we now have the ability to now go, spot pricing for short term?

Mr Yuen Kuan Moon, Group CEO

So, so actually when you talk about spot rate, right, based on retail prices. First of all, you have to understand that a lot of this capacity is deployed for sovereign AI. Even you've got a spot rate somewhere else, y, you are not serving my purpose. That's why it's very country-specific. You need to look at the capacity and the demand from a country perspective. And in Singapore, this is constrained by energy. Like I said, you know, we got more demand than the 11 MW that we've got. But whatever we can harness in terms of energy, it is where it is.

So from that standpoint, you're not competing with the spot rate somewhere else in another foreign country. And I don't think there's a lot of spot rate for Singapore. So this is the rate that we've contracted. The customer is willing to pay for this rate. We're willing to sell at this rate. So if you go into any region, we should have to look at that solution and say

I'm providing this RE:AI not just to the hyperscaler, but really for a sovereign AI solution.

Hussaini Saifee, Maybank

Yes, thanks for the opportunity to ask a question. If you look at the you know globally, especially in US, this neocloud, scale is the name of the game. So just trying to understand it, you know, the Singtel is starting now, are we looking at a multi-year investment cycle in this space because we are just now at 11 MW, whereas in US we are talking about the gigawatt kind of capacity.

19

Mr Yuen Kuan Moon, Group CEO

I think you're talking about a very different model. I think you talk about how you probably have to go back and look at the neoclouds, who are their customers? Are they end user enterprise customer? Are they government agencies or are they selling to hyper scalers of the East and West doing training workload. I think it's all very different. I'm not going to give a guidance beyond what we have said, because this is a very new business, we have to really appreciate and understand the business model. The unique economics, does it scale, is it sustainable, right, the returns, is it justify the investment. So we have done the pilot in

Singapore of 1 MW which I, we, we tried, tried about 6-18 months ago. Now we are confident and say, hey, I'm going to put in another $600 million into CapEx to really scale this in Singapore.

Upon successful scaling it in Singapore, obviously concurrently we are looking at the region. Is there a similar play of Sovereign AI? That's our approach in our own neocloud.

Other neocloud players will have their own model, right? I don't know what is their target customer. I also don't know which region are they playing. Are they serving only a specific

Chinese hyper scaler, which many of them are, but that's not our model But whether forward beyond that, if we can find a niche that we are unique, that we can differentiate ourselves, we will do so. But it's too early to tell. So our focus is to make sure we deliver this first 11 megawatt first so that we can monetise it faster.

Rachel Tan, UBS

Hi, this is Rachel from UBS. I have three questions. Earlier I think you mentioned monetising your 5G SA network. Could you elaborate on that?

My second question would be for the sourcing of customers for RE:AI, how active is NVIDIA in introducing customers to you, and will you rely on them when it comes to your

Southeast Asia expansion in terms of partnering with them for the customer network?

And my third question is how much do you expect AI to be a part of your operating cost?

For your partnership with Open AI, what is the nature of the contract? Are we looking at monthly subscription per head or is it like token based billing? Thanks.

Mr Yuen Kuan Moon, Group CEO

So the third question I'll answer it, the second question is easy. I'll leave the first question to Tian Chong to talk about monetisation of 5G SA network. I think just a quick answer to that is, you know, we are really monetising it both from enterprise and consumer. I'll leave it to Tian Chong to explain later.

For the RE:AI, if you talk about sovereign AI, the customers, we have to look for it ourselves in Singapore. Of course, NVIDIA is still a very important partner because they do have the visibility of where are the demand, where the needs, and I'm sure Bill can add on a bit more about our partnership with NVIDIA later on.

Finally on the AI operations of OpenAI partnership, we are one of the early partners in terms of adopting ChatGPT on the enterprise grade. This is something that we go in with our eyes open to enable our staff to know how to use what we call the frontier model of AI

20 to help on the daily operations. But this is not just about personal productivity, which everyone of us uses. It's also about transforming our core business, our core operations.

Tian Chong, of course, is working full stack across all its operations from call centre to network operations to product development and also to sales and marketing.

So we are using or deploying AI across both cost productivity and revenue drivers to improve our margins and performance. So, it is not just a paying subscription or tokens.

It's probably a combination of both subscriptions and token to deploy AI well and to capture value, you definitely have to look at token management, and that's why we are not relying on only one AI partner. That's why we also talk about Sierra AI. We also build in our own models especially on the network side so that we can optimise our token. I think later on I can also ask Sam to talk about how he's helping customers to deploy AI coding using different versions of it. So it's the full works of the entire AI stack that we are working on.

So maybe now I hand over to Tian Chong to talk about the 5G monetisation of both consumer and enterprise.

Mr Ng Tian Chong, CEO Singtel Singapore

Sure. I'll quickly talk about that. On the 5G standalone network. I know I've spoken to many of you on this topic before, but we completed our SA network set up in 2022. Since then, our SA network is very unique in Singapore because we have progressed very far in working with public sector first on how to do slicing, and we've done many types of slicing to support use cases for homeland protection, homeland defence. And then later on, we have progressed to now supporting the automation of the most complex and biggest automated port in the world, PSA in the west, and then of course the automation around the airport ecosystem of the east.

So for us we look at Singapore is like a dome, you know, with the 5G SA network wraps around and what Arthur presented earlier on the enterprise case, if I can just simply explain, is because of the unique capabilities, We've gone into multi-year, co-innovation partnerships with organisations like SATS in the east, PSA in the west. And then the government agencies like LTA to build a smart city, smart junction, where Singapore uses a lot of data. And of course, today we are at 5G advanced stage going towards 6G by 2030. So the uplink capabilities are really solid and support many of those with video capabilities and then now the way we describe it is the network is supporting network for AI, network for robotics going forward and autonomous driving and all of that. So we monetise a lot in the enterprise case.

In the consumer space, the reason why we have developed the tri-brand approach in

Singapore, is because you know we are, we really want clarity. We want to offer the right value at the right price for every segment. But more importantly it allows us to really protect and invest in our premium brand. And it is because of our uniqueness of our 5G standalone network.

We're the only operator in fact globally that have taken this slicing capability and opened it in the consumer space where today on our premium plans 5G Enhanced that means a $30 price plan upwards. We're able to now bring our mass market into the slice and offer them faster speed and more secure protection. And we have done that. And in fact with this tri- brand approach, it allows us to confidently get to do this without a price war at the low end,

21 because we've got other price. So Scoot do what they want, Singapore Airlines can focus what they need to do, using that metaphor. And Singapore Airlines doesn't have to deal with AirAsia, so to speak.

So we have done that and our focus is on revenue share. Our focus is on premium, and in fact on the 5G+ with monetisation effort, most recently we've seen a 37% year-on-year uptick, you know in our premium plans. So we continue this strategy and this is independent of any market outcome of merger or no merger. You know we have executed that plan and will continue to focus on that.

Mr Bill Chang, CEO Digital InfraCo

In the market of Singapore, we access those customers directly because public sector agencies, enterprises. However, when we expand across the region, the customer segments that we approach here is because of the sovereign AI local champion, will be very different in those markets. It is not going to be ‘oh, we want to choose the same customer so we can go over there’ because you are not a local champion there. So, it's a different dynamics. You know, dealing with that could be enterprises, hyperscalers who will buy from us, right? So this is where ecosystem partners want NVIDIA is very key. And why are they key?

One, as the leading sort of GPU-as-a-Service provider in this region, we get access to the latest chips. And that's key in a world where you think that heavy demand like supply getting access to those chips.

Secondly, they've got a network of customers, referrals and all that. We do need them in the region. And they do value-add to that process. Beyond NVIDIA, we also have Mistral, second partner. We have Cohesity, the third partner. All these partners do bring contracts with us and that's why we set up this whole ecosystem to pursue across the region and in

Singapore.

And finally, our partnership with NVIDIA brings a third element other than just referrals and chips. It's this whole thing about designing and co-designing our data centres to fit their GPUs in 28/29 so that Nxera is getting those capabilities way ahead of other data centre companies. And by doing that, that's why Nxera is at the leading edge of being sold out when it comes to AI DCs. Because we deploy RE:AI and we're deploying those GPUs and Nxera exists to support RE:AI to achieve that. And so that symbiotic relationship drives back the value into Nxera, and Nxera it's getting sold out on that.

So those are the three reasons why we built NVIDIA as a very key partner, and as well as other key ecosystem AI partners that we are now assembling.

Arthur Pineda, Citi

Thanks for the opportunity. Two questions please. Firstly, with regard to Bharti, I'm just wondering, given that there's been a recent change in the gap between yourselves and the promoters, how does that actually change your outlook in terms of capital recycling?

Because presumably Bharti was a big part of your funding source for the $9 billion? That's the first question.

22

Second question I had is with regard to Australia, just wondering where we are with regard to all of the government investigations. Is that all behind us now or is that still ongoing?

Mr Yuen Kuan Moon, Group CEO

I'll ask Stephen later to talk about where we are at all the regulatory conditions that are set on Optus. Maybe Arthur, you talk about Bharti. First of all, I think there's still a gap even with the recent transaction.

Mr Arthur Lang, Group CFO

Your assumption is that it's a smaller gap now is itt? To be clear, the $9 billion that we set out. First of all, Bharti is not the only one, but it is a main component of the $9 billion. If we say we want to monetise our direct stake, the whole 7.5%, you know that number is way bigger than that. So I think that's the first part.

Second point there's, there's still intention to equalise. I mean, that's assuming that SBM continues to hold whatever stake that he will hold after the share swap. So, in the past it's been changing. But there's also a possibility we decide to both sides sell everything to BTL.

There's a lot of possibilities. I think the takeaway, you should not second guess that $9 billion number. It is still very much intact. So you should think about it as something that we have thought of. I mean give us some credibility. We have thought about all types of scenarios and it's a number that we still stand behind.

Ms Lian Pek, VP of Investor Relations and Group Corporate Affair (Host)

Optus, on the government investigation.

Mr Stephen Rue, Optus CEO

The network outage in 2023 is done. The unconscionable conduct and mis-selling of the

ACCC that's also long term undertaking which were part of, but in terms of the regulatory investigation.

There are still some outstanding matters relating to cybersecurity, so that is the OAIC or our data breach in Australia. So that's still, involved, but that's 2022.

And in terms of Triple Zero, the Senate is due to hand down its findings at the end of June.

And the ACMA still has not completed its investigation into the Triple Zero outage so that we're still awaiting. And there will be some coronial investigation.

Prem Jearajasingam , CGSI

Thank you for the opportunity. Just given all that's been happening and you know what you're saying about RE:AI and also the Middle East situation, the power situation globally.

How is that changed or modified your views with regards to the DC rollouts over the next three to four years? Is there a chance that we double what we thought we would do by

2028, 2029? Where are you going on all of this?

Mr Yuen Kuan Moon, Group CEO

23

Well, I think first of all, if you look at our plan, we have communicated very clearly what's our medium term aspiration of rolling out data centres both in Singapore and in the region under Nxera.

The first priority obviously is we want to win the CFA2. That is probably the crown jewel, the prize for Nxera and I'm sure Bill have put in a very compelling proposition to the people who are evaluating and accessing that CFA2 process, and that is a priority. And of course, a lot is dependent on how much capacity if we win, we've been awarded, right. Because there's a lot up for grabs, but equally a lot of high interest as you can imagine in terms of the number of interested parties to also put in a bid for it.

Beyond that, I think we have also articulated in Thailand where our growth potential, in

Malaysia, where our growth potential and also in Indonesia and Batam. And I think we've given a very clear update of some of the data centres is coming up very soon ready for service. And you can see the various stages of pre-selling, from 60%, 80% before we roll out.

So again, it is constrained by energy. If you look at it, sometimes it's not just about customer demand. In fact, there's a lot of demand. I think Bill can talk about the demand that is coming in from hyper scalers in the region, but it's also availability of energy. So that's where finding the right partners is very important, can help us secure energy in the region. So Bill maybe just touch on the region as well as the CFA2.

Mr Bill Chang, CEO Digital InfraCo

CFA2, I think what we're doing to ensure that we put one of the, without going too much into details, looking at renewable energy that basically is a factor additionality pathway. So that's not taxing on the current ground nor the green energy that they've got countries secured, but this is additional pathways, right? Additionality they call it. So which means you got to pay a higher cost, but it's something good so they don't overtax the grid. And you know, obviously we're looking at the most sustainable platform from a PUE and WUE, right? So I'll just leave it as that. And the other one is obvious, designing it to carry GPUs of the future. Because we see this need of sovereign AI, so I'll just leave it as that.

Across the region we are doing the same thing looking at renewable energy to access that early. Malaysia, and think about, Thailand and Indonesia, and our partners there are some of them are energy companies. The question it's not just energy, it's also water. In Malaysia for example, parts of Malaysia have water, so we have to think about that. So other than driving efficiency of water can be used, water from not in the grid of the water grid. You know wastewater and all the collection and all this stuff. How do you do DCs with closed loop systems and working with clients who operate at much higher temperatures, even though with liquid cooling, and how do you then create digital technologies to sense and be able to react quickly. So these are all the innovations that Nxera is doing to really differentiate and to be able to make sure that we are the first we're the first to roll out 200 kilowatt per rack, and we're pushing the envelope to be the first to go 600 kilowatts per rack for the Vera Rubin Rubin Ultra. And then the 1 MW, we're pushing that very hard to achieve payment. So we're designing all that so that RE:AI can move, at the same time

Nxera becomes sold out every time where we go. So now accessing land and energy and water becomes a key part.

24

We are obviously looking beyond Southeast Asia now. You've seen press releases and

LinkedIn messages looking at partnerships in Japan, tier one markets, and beyond

Southeast Asia. We're looking at all those markets. At the moment, we don't want to talk about anything that's 2-3 times whatever. We just want to contend with the current and getting solved. You know this thing around sustainability and getting the energy, otherwise more DCs out there and you are just rolling out DCs, you're going to create an energy crisis. So we're doing this responsibly and sustainably. And engaging a community closely.

Mr Arthur Lang, Group CFO

Just to clarify, you talked about the reference of whether we are still confident of doubling the EBITDA. To be clear, that EBITDA doubling is based on just simply DC Tuas fully stabilizing. So that specific question, yes, we are confident. Then on top of that is what Bill has been saying all the expansion, the longer term.

Ms Lian Pek, VP of Investor Relations and Group Corporate Affairs(Host)

OK. So I think on that note, we probably have to wrap up. Are there any further questions?

We're 80 minutes in actually. So it's flown by, lots to talk about, and we will keep you posted as to you know what transpires next.

On behalf of the management and the IR team, thank you all for joining us today, both online and here in person. Thank you.

End of Transcript

25