Morning, everyone. Thank you for joining Keppel DC REIT's one half 2026 results call. I'm Renee from the investor relations team. Let me first introduce the management team on the call. We have CEO, Mr. Loh Hwee Long, and CFO, Mr. Adam Lee, as well as Head of Portfolio Management, Ms. Charmaine Cai . We will start with an overview of Keppel DC REIT's financial and operational performance, followed by the Q&A session. I will now hand the time to Hwee Long.
Thank you, Renee. Good morning, everyone, and thank you for joining us this morning. We delivered a strong set of results in one half 2026, reflecting organic growth, contributions from acquisitions, and continued discipline in capital management. For one half 2026, distributable income and DPU increased by about 19% and 11% respectively year-on-year. We are also disclosing contracted power capacity for the first time, which stands at about 95%. This means that about 95% of our power capacity is revenue generating. As revenue is more closely tied to contracted power capacity than lettable area, we believe this provides a more meaningful measure of utilization than traditional occupancy metrics. In one half 2026, we secured renewals and new contracts in both Singapore and Australia, extending our portfolio weighted average lease expiry to 6.7 years. Portfolio reversion for the period was around 10%.
This excludes renewals that have been secured but will only commence in second half of 2026. Overall, we remain well-positioned to pursue growth opportunities, supported by a competitive cost of debt of 2.6% and a healthy balance sheet of more than SGD 670 million, debt headroom to our internal 40% aggregate leverage limit. I will now let Adam take you through the financial updates.
Thanks, Hwee Long. Our strong financial performance in first half 2026 was driven mainly by contributions from organic growth and acquisition. This was partially offset by absence of income following the divestment of Kelsterbach Data Centre in 2025. Growth in distributable income and distribution per unit was also augmented by our increase in interest in Keppel DC Singapore 3 and 4, which increased the proportion of income attributable to unitholders. Finance costs were higher year-on-year due to acquisition-related borrowings drawn in fourth quarter 2025. We continue to deliver sustainable distributions to unitholders. The first half 2026 DPU of SGD 0.05714 will be paid on 18 September 2026. Moving on to capital management. We continue to maintain a strong and disciplined balance sheet. As at 30th June 2026, aggregate leverage improved to 34%, mainly due to loan repayments.
Cost of debt inched up quarter-on-quarter following hedging of Singapore dollar loans. Assuming no changes to our capital structure, we expect our cost of debt to hover between 2.6% and 2.8% this year. Interest coverage ratio is at 6.9 x. Before, we remained well within all financial covenants. Our capital structure remains stable and well managed. Debt maturity profile is well staggered with a healthy weighted average debt and hedge tenor of 3.1 years. We continue to have good visibility over our refinancing requirements. 87% of borrowings are on fixed rate, limiting interest rate volatility. We maintain a natural hedge of about two-thirds for our overseas portfolio, which helps mitigate foreign currency risk. In addition, we have substantially hedged our forecast foreign source distribution through first half 2027, with progressive hedging beyond that as well.
I will now let Charmaine take you through portfolio updates.
Thank you, Adam. Our portfolio remains of high quality and has multiple growth levers. As at 30th of June 2026, portfolio occupancy was 92.5%, reflecting the contract expiry at Cardiff Data Centre in the second quarter. Excluding this asset, portfolio occupancy would be 95.3%. About 95% of the portfolio's power capacity was contracted and is revenue generating. Contract renewals in Singapore and Australia extended portfolio WALE to 6.7 years. Portfolio reversion for the first half of 2026 was about 10%. We have also secured new contracts at Gore Hill Data Centre. We expect to see income from the asset more than double from 2026 onwards, with further upside as leasing progresses.
Our portfolio is well-insulated from power cost volatility, as most electricity costs are contractually recoverable. We continue to proactively optimize the portfolio through leveraging our significant co-location exposure to drive organic growth, evaluate asset repositioning, power intensification, redevelopment or capital recycling opportunities, as well as selective acquisitions of high-quality data centers that meet hyperscalers' requirements. Around three-quarters of our portfolio comprises co-location contracts, which give us the ability to capture organic upside as contracts are renewed over time. Our contract expiry profile is also well spread, with only a small portion of contracts due in the near term. These near-term expiries are largely smaller contracts that are close to current market rates. The majority of our income extends into the medium to long term, providing stable cash flows. We have built a diversified client base anchored by hyperscalers in line with our strategy.
Hyperscalers remain a key focus for us, given their strong credit profiles and sustained demand for our data center capacity. At the same time, we remain mindful of concentration risk and manage this across our portfolio through diversification by assets, geographies, and lease maturities. The near-term macro environment remains uncertain, particularly with interest rates expected to remain elevated, which could continue to impact financing conditions and investment spreads. However, the underlying demand for data centers remains very strong. This is driven by the continued shift to cloud as well as growing AI workloads, which are expected to significantly increase demand for capacity over time. Keppel DC REIT's portfolio is well positioned to support such cloud and AI inference workloads. At the same time, power availability remains a key constraint and is increasingly shaping where and how new capacity can be developed.
Against this backdrop, we remain focused on enhancing portfolio quality and maintaining financial flexibility to capture accretive growth opportunities while delivering sustainable long-term value to unit holders. Thank you.
Thank you, Charmaine. We will now take questions. For analysts who are joining us on the Teams platform, please click on the raise hand button if you would like to ask a question, and wait for our cue before you unmute yourself to pose your questions. For the benefit of all participants on today's call, please limit yourself to two questions. If you have more questions, please re-enter the queue. For those joining us online via the webcast platform, please type your questions via the chat box provided. For the first question, can we have Terence from JPMorgan?
Hey, thanks. Thanks, Hwee Long and team. Very strong set of results. Could I just ask firstly on the management fees in units? From our calculations, it seems that the MFU has gone from about 27% in the second half of last year to 41%. Can I ask what drove the change, please? For the second set of questions, I wanted to ask on rent reversions, 5% in second quarter versus 50% in the first quarter. Does this include the renewal secured for Gore Hill? If not, could you give us a sense of what that would be if we include Gore Hill? Thanks.
Okay. Thank you, Terence. Maybe for your first question on the management fee in units, I'll get Adam to maybe share a little bit on that front, and then I'll reply on your second question.
Thanks, Terence. Proportion of fees in units increased over the last two years, as you have rightly pointed out, as manager has elected the management fees to be received in units for the newly acquired assets, aligning the manager's interest with that of the unit holders. The fees in units in first half 2026 increased from Tokyo DC 3, as well as the remaining stakes acquired in Singapore 3 and Singapore 4.
Okay. That's great. Thanks. The second question?
Okay. Yeah, that's fine. Okay. No, I think on that front also, maybe if you were to look at it also, even if you strip away the management fee units to hold it constant, I believe, I think our year-over-year, quarter-over-quarter is still our trend in that sense. Anyway, I think, Terence, to your second question on reversion. Yes, I think for the first half, I think obviously overall, the blended average is about 10%, right? You are right that I think for the second Q, the number was about 5%, right? Compared to 51% in Q1, right? On that front, I think, for reversions, because of the nature of it really depends on what contracts come up for renewal any one quarter that we talk about. For instance, I think if you draw on last year's pattern, actually, there's no clear pattern.
Certain quarters we have the strong double digits. Certain quarters, it's also a single-digit type of situation. I think what that suggests is that, I think, for example, in the past quarter, those reversions or the renewals that we were able to report, those leases they are probably quite close to market in terms of the in-place rent. That's why I think the jump is slightly smaller in that sense. It doesn't mean that those are bad outcome. In fact, it's still very good outcomes. I would say, overall, I think for this year, we should still be in a pretty good position based on what we can see at this moment from this point. To your question on whether it's Gore Hill, what we managed to renew or even sign. Is it captured in the existing so-called reversion number that you see?
Short answer, no. It's not captured in the current one that you see. You will see the effects from Gore Hill probably in the third Q numbers. Rather third Q reporting, in terms of that reversion. Yeah, I think from that standpoint, if you compare Q on Q, if second Q is about 5% for 3Q with the Gore Hill effect, that should lift us back up to, I think, a low teens type of range. Yeah.
Okay, that's great. Thanks for the answers.
Yeah.
Thanks, Terence. Next, DBS Dale.
Thanks, Renee, and thanks, Long and team for the presentation. Just two quick questions from me. Just wanted to follow up with regards to some of these vacancies at Cardiff, as well as SGP 1 and SGP 2. Previously, I think you did mention that potentially looking at some rejuvenation or enhancement work. Any updates on that? That's the first question. Then for the second question, just wanted to touch on the NetCo bonds. I saw in the announcement somewhere you mentioned that there was a partial repayment of these bonds. Could you explain more on that as well?
Okay. Thanks again, Dale. Okay. I'll handle the first question, and then later I'll let Adam respond on the NetCo bonds. I think for the first one, yes, you are right. I think from a vacancy perspective, we took back possession of Cardiff towards the end of 2Q this year. At this juncture, it is sitting with us. I think what is ongoing at this juncture, of course, I think there's activity seen going on. Because the tenant has left us in that sense. That's why you also noted that based on the occupancy by lettable area, our occupancy actually took a slight dip from, I think, 95% to about 92.something% in the past quarter. At the same time, of course, I think a point to note is also, Cardiff is actually a relatively small asset.
I think it contribute less than 0.5% of our portfolio from AUM standpoint. As a result, incidentally, I think this quarter we also started to share a bit more on the power capacity type of contracting levers. I think that's probably a more relevant one. In this case, despite the so-called taking back of the Cardiff Data Centre, I think if you look at us from a contracted capacity perspective, still, I think very strong at about 95%. On that front, I think two parts. One, I think we continue to have conversations with potential customers to talk about leasing out that space. At the same time also, we are also quite open to other optionalities. I think, again, back to our overall strategy, today I think we are executing a strong pivot towards the hyperscale type strategy.
If there are interesting optionalities that are presented to us, I think we will also be quite open to those conversations as well. To your other part of the first question pertaining to SGP 1 and 2. I think SGP 1, I think the trend came down slightly again in the past quarter. Again, I think that is going back to our so-called midterm plan. For SGP 1, we are making preparations essentially for a potential, I think, redevelopment for the asset. I think from that standpoint, there will be a calibrated ramping down process for the remaining period of time until we take full possession of the asset.
In the case of SGP 2, that one is more of, I think, your regular type of. Because we run a co-location business, there's more of a rotation type of situation where some of the smaller customers left, and also a bit of it has got to do with non-DC space as well. Anyway, I think on that front, no problem. I think there's quite a few active discussions going on. We should see that being leased back up also quite quickly. I think on the second question, I think NetCo bonds, maybe I get Adam to give a bit of update on that.
The NetCo bonds, if you could recall, I think we did mention about cash flows of SGD 11 million per annum, and that comprises of principal and interest payments.
The partial repayment of the loan relates to the principal repayment
Sorry, Hwee Long, just wanted to follow up on the first question, the SGP 1 redevelopment. I see that
Yeah.
WALE is about 1.1 year. Should we be expecting the so-called redevelopment to only happen after this 1.1 years?
Yes. 1.1 year is obviously a WALE. It plus minus a little bit. Yes.
Okay. Got it. Okay, that's clear. That's all from me. Thank you.
Yes.
Next question, CLSA Yew Kiang.
Hello. Hi. Can you hear me? Hi, Hwee Long.
Hey, morning, Yew Kiang. Yes, we can hear you very clearly.
Yeah. Sorry, just two questions. One is on the contracted power capacity. What's the reasoning behind releasing this data point? Would you be showing a trend of this into the past and how has this been? What's the lowest level that we have seen for this level? When was that?
Okay.
Also, does it include the Chinese assets, the Guangdong assets? Does it fit into this number?
Okay. Is that the two questions or more?
That one is one question.
Oh, okay. I answer your first question first. Contracted power capacity, that's something that, as part of the evolving, as we find ways, because we have been hearing, of course, not only from yourself, from the investors as well. Historically the way that we announce or we share about occupancies really from an NLA perspective. We all know that for data center, because the way that the contracts are signed with the customer is really more, the unit base is really on a per kilowatt basis. That means on a power capacity standpoint. If we were to continue to use NLA as a basis, it's okay. It's a good proxy, but it may not give the full picture in that sense. My team has been, to be honest, took a while.
There's a lot of complexities around this switch. We thought that because of the way that the contracts work for data center, power-based type of occupancy metrics will probably be a much more closer proxy or rather a better representation of the business itself. That's why we have shared this. In terms of the trend, moving forward, obviously, we will continue to show this. To your point about the look back, okay, that one, I don't have a straight answer for you at this moment, unfortunately. Sorry. Let us go back and think through, and we'll come back to you. Your second half of your first question in terms of whether is the Chinese asset included in this.
Does it include? Yeah.
Yes, it is, it includes, yeah.
Oh, okay. My second question is on the M1 sale didn't go through. We should continue to see the income coming through, the interest income coming through, right?
Yes. You should.
Okay. That's it for me. I'll jump back. Thank you.
Okay. Thank you, Yew Kiang. Thanks a lot. Yeah.
Next one, Citi from Brandon.
Hey, morning, Hwee Long. Can you hear me?
Yes, very clear.
Hey, great. Hey, just wanted to ask a bit on the contracted occupancy. Are you comfortable to share where by markets this number is? Specifically, the co-location markets of Ireland and Singapore? Yeah, that's my first question.
Okay. That one, let us take it back also. Not at this juncture in this call. Apologize for that. Let us think through. I think on that front, you should also be, logically, the co-location will have so-called relatively lower so-called occupancy in that sense compared to the master leases, because the master leases are by default 100%. In terms of the split by country, let us get back to you. The co-location, I think the business, that's the nature of it. It's relatively short WALE, and there's always people coming in, people going out. That's also part of the strategy, how we U up the assets in that sense. That's, I think, a big plus in terms of the business model that we run.
Through the co-location, I think that allows us, I think, a lot more levers to drive very meaningful organic growth compared to if we were to be saddled with only pure master leases.
Okay. Can. Also on the MFU. Is it correct to say that going forward, your future acquisitions, you will try to pay out, what, 40%- 50% in units rather than the traditional 20% that you have been giving? Also with relation to that, what's the reason for choosing higher MFU? Yep. Thank you.
Okay. Maybe I get Adam to answer this part of the question.
Brandon, this MFU, management fees in units. I suppose that's what you are trying to allude to. It is going by each deal that we assess on a merit basis to see whether we want to elect or not elect the management fees to be received in units. For example, like I mentioned earlier, the Tokyo DC3
We collected management fees in units, subsequent acquisition of the remaining interest for Singapore 3 and 4, that we also collected the fees in units.
Yeah.
On a pro basis, they know that we will review it on a deal-by-deal basis. I think that's what Adam is trying to convey. Yeah.
I just want to understand more the rationale. How do you decide whether you want to take 20%, 30%, 31%, 42%? I just want to understand what's driving that number.
Yeah. Certain assets that we acquire, we do collect 50% base fee. Certain asset we do collect 100% base and performance fee, like the Japan one. This 40% is a resultant of a portfolio of assets that receive management fees in units. It's not a number that we apply on a portfolio level as to how much we receive in units. Yeah. It's not a percentage. Next quarter, if everything remains the same, technically this percentage shouldn't change too much. It should change based on the performance of the REIT.
Okay. Can. No worries.
Yeah.
It's okay. Okay, thanks. I'll move on. Thank you.
Next question, Macquarie, Rachel. Rachel, can you hear us?
Hello. Hi, can you hear me?
Yes, great.
Okay. Great. Hi. Morning. Thanks for the presentation. Maybe first question on interest cost. I did see a slight tick up in second quarter. Just wondering, with SORA slowly creeping up, should we be expecting a turn in your interest rates from now onwards? My second question is on potential acquisitions. I think you have been a bit quiet in that front. You have a lot of debt headroom as well. Should we expect an acquisition soon? Which country are you spending most time in evaluating proposal?
Okay. Thank you, Rachel. I think first question on interest, I get Adam to also answer that.
Thanks, Rachel. In second quarter, the full effects of what we have done in first quarter on hedging has contributed to the so-called slight increase from 2.6%- 2.7% quarter-on-quarter. If you look at our fixed rates and floating proportion, only 0.2% of our entire debt profile are in Singapore dollars. The SORA increase will have a minimal impact to our so-called interest cost.
Oh, okay.
I think on your second question, Rachel, I think pertaining to potential acquisitions. I think for us, of course, we are constantly evaluating deals. I think no particular hurry in that sense, because I think the key thing is really to get the right deal that fits our Keppel DC REIT's objective. Also meet the investment requirements and hurdles that we can deliver to our unitholders. Notwithstanding that, of course, I think from a strategy, from a market selection standpoint, I think we are still very focused in terms of the type of markets that we choose to focus on. Generally, what this means will be markets that possess, one, high barriers of entry from a supply standpoint. It can be coming in the form of power constraints markets or difficulty in terms of construction resources to be able to deliver new projects.
At the same time, demonstrate strong demand profile. On that basis, we continue to be very focused in terms of scouring for opportunities, not only in Singapore, but also in the regional markets of Japan, Korea. Also, I think we are starting to see there might be interesting opportunities out there in the core FLAP markets in Europe, where we have presence as well. Yeah.
Oh, thank you. Just following on that, it seems that you are a bit more delaying your acquisitions. I was just wondering, are you keeping some of this cash for like, say for example, your Cardiff, whatever that you are trying to do on Cardiff? Should we expect something soon from Cardiff? Redevelopment or AEI?
No, we have not announced any redevelopment or AEI plans for Cardiff yet. I think at this moment, as I mentioned earlier, I think a key drive is really, one, focus on re-leasing. At the same time, also keep our options open to other optionalities. I think you'll note that, I think that site is relatively small in the bigger context of things. I think from a submarket perspective. Not a bad market, but it's probably a Tier 1.5 , Tier 2 type of market in the European context. That is, I think at this moment, what we are looking at. On the comment around are we delaying our acquisition drive? No, I don't think there is an active so-called delay or accelerating type of push on that front, on the acquisition front, right?
That one is really, I think because as with all acquisitions, it's really more opportunistic, right? I think we want to make sure that we only press the button if the right deal were to come about. I think that is more what drives the so-called, for example, in the context of 2026, right? First half, we have not done so-called any major acquisitions to date with the exception of the small interest in the SGP 3 and SGP 4. Yeah.
Okay, great. All right. Thanks, Hwee Long. Thank you.
Yeah. Thank you so much. Thank you so much, Rachel. Yeah.
Next question, Goldman Shen.
Hi, morning. My first question is on rent reversion computation. Can I clarify if that has changed? It seems that previously it was based on when the lease was signed, and now it's based on the actual expiry. Yeah, if I look at your footnote. Just want to clarify that. Second question is on DPU growth 2026 versus 2027. If I look at first half number, higher MFU and lower CapEx reserve actually boost your DPU number this year when the organic growth is really strong. But if I think about 2027, and if you ramp down SGP 1, that's where DPU will be negatively impacted. I guess to Brandon's point earlier, why raise MFU now when your organic and inorganic growth are so strong and not use it for 2027? Thanks.
Okay. Thank you. Thanks a lot, Shen. I think on the first question pertaining to the computation methodology for reversion. No, I don't think there's a change in the methodology. I think it's quite consistent approach that we have. That's the part. No change. Second one, I think in terms of from a DPU growth perspective. Between what 2026- 2027. I think like what Adam mentioned, I think in terms of the election on the MFU for different quarters is really, I think dependent on, for example, individual deals and assets. I think in this case for the Japanese acquisition, we have opted to receive, I think, some of the fees in units. That caused it to go up, spike up for this current so-called first half. On a forward basis, again, I think that is something that not cast in stone.
In terms of that approach. Secondly, I think in terms of the commentary around the profile, how that matches with SGP 1, ramp up, ramp down type of arrangement. Yeah. I think on that front also, I think in terms of a bit of a crystal balling. While there is a so-called ramping down trajectory for SGP 1. I think on that front, rest assured that will be something that we'll manage very carefully. Such that overall, the objective is really to balance to ensure that we preserve the DPU growth trajectory. Yeah, that's what we will aim to do. Yeah.
Quick clarification on slide 12, footnote five. I think this footnote is new, right?
Maybe I get Charmaine to give a bit of commentary on that.
I think we added this footnote just to clarify the actual previous footnotes. Let's say footnote four, to make it clearer that it excludes the impact of the renewals.
Okay. I think if you look at that, I think the footnote five you're saying, I think what she's trying to say is that, or rather, sorry, Charmaine, in that comment was more of it excludes the impact of renewals in one half 2026 that will commence in two half 2026. I think the methodology no change. I think what that is trying to tell you basically, I think to the first question this morning on Gore Hill. Because Gore Hill was executed in first half, but the actual impact will only come about in 3Q onwards. That's why when you talk about reversion this quarter, when you talk about the overall first half, 10%, the Gore Hill effect is not there yet. You only see it in 3Q reporting onwards.
Okay. Got it. Thank you.
Yeah. This methodology, we have been applying it consistently, historically as well. Just to clarify. Yeah.
Next question, DBS Derek.
Hi. Morning. Just a question, I guess on contracted power occupancy. Hwee Long, you alluded to it being more helpful to divulge such information. It would also be helpful if you can divulge the power capacity of the assets. Is this something that you can share with us?
Okay. Thanks for the feedback. Let us take back to delve a little bit more in terms of asset by asset. Again, I think the challenge we always face is really at the asset level, if there are certain single tenancies or larger tenancies, we have to balance with the confidentiality arrangement that we have with our ultimate underlying customers. That part let us delve a bit more.
What about on a portfolio basis? It's more anonymized at a portfolio level, the aggregate power capacity.
Correct. Today, I think that's why we are sharing the 95% on the whole portfolio basis, right?
No, I mean the total capacity in megawatts.
Okay. Sorry.
Yeah.
I misinterpreted. Okay.
Since you already shared the occupancy by power capacity, why not share the Yeah, that's what I meant.
Understand. Sorry. Okay. I think on that front, happy to. I think today, if you look at our entire portfolio, right? Sorry, again, I have to give you a range.
Sure.
To try to address our part on the balancing. Today, I think look at our portfolio, I think our capacity perspective, we are in the range of around 350 MW-450 MW.
350 MW- 450 MW for the entire portfolio.
Yes.
What about Singapore portfolio?
That one, I think we are not ready to share to that granularity yet. Yeah.
Okay. Got it. Okay. No worries, Hwee Lon g. Thanks for sharing that part anyway. For the reversions outlook for the full year, given what you've guided so far, safe to assume we're looking at 10% or somewhere around that region for the entire year, right? Because third quarter is really low teens.
Yeah. We'll definitely work very hard to get ourselves to a good position now. I think some are still under negotiation, still no firm numbers yet. Yeah, mathematically you are not wrong.
Okay. Can leave it as that. Just one question for Adam on tax expenses. I think there's some bit of movement in this particular first half, especially with the deferred tax expense at the distributable income side. Could you give some color? It seems like you're kind of getting a tax credit.
Yeah. Okay. Tax expense in first half 2026 were lower due to the tax transparency benefits that we have, as well as absence of some one-off tax items that were recognized in first half 2025 for Keppel DC Singapore 7 and Keppel DC Singapore 8. These are items like income support, which were not eligible for capital allowance deduction. All these were partially offset by the absence of the deferred tax asset recognized on the divestment of Cardiff Data Centre back in first half 2025. There were some movements in terms of the deferred tax balances relating to Guangdong Data Centre itself.
Okay. The deferred tax expense that we saw in first half that was plugged back into the DI, that should not repeat itself in second half onwards, should it? The six over a million dollars.
Six over a million dollars. There will be some that will be repeated because I think our tax transparency was obtained with retrospective effect in February. There should be still some number there. I think in second half, what could potentially cause the deferred tax adjustment would be also valuation movements, and also capital allowance claims in overseas jurisdictions.
Okay. Yeah. The reason I ask is because once you put in deferred tax, it's almost like you have a slight tax credit to your DI, and that's kind of moving up the DI as opposed to cutting into it.
Okay. All right, cool. Thanks, Adam.
Hey, we have some questions online. People are asking about an update on Guangdong Data Centres in terms of the backfilling as well as the outlook.
Okay. Thanks for the question. I think for Guangdong or for China, obviously, I think China, from a DC standpoint, in the past 12-18 months, there's been definitely marked improvement overall market. I would say a lot of the excitement around last year was DeepSeek. Then more recently, I think in the past couple of days, you hear a lot more about things like Kimi K3. I think there's a very strong AI trend going on in China, which will bode well, I think from a future demand perspective. Today, of course, I think the demand growth, demand update today, I think is still quite patchy. It's still quite concentrated around the greater Beijing area. In our part of the world, our part of China, at this juncture, candidly, no material uptake at this juncture.
I think the positive that we draw from it is that, one, in terms of the conversations that we have with potential customers, there is definitely, I think, a lot more constructive conversations going on around there. The other part is also, I think, the consistent feedback that we have gotten from some of these potential customers is also, I think they're still generally a bit lacking in terms of availability of already getting access to GPU chips, which seems to be hindering their ability to scale up faster. Those are things that we continue to monitor quite closely. What we are doing at this moment as we prepare for the asset recovery, we will increase, we intensify, we constantly work on the customer conversations.
Such that hopefully, I think as the demand profile further strengthen in a more even manner across the entire Chinese market, hopefully we'll be in a good position to benefit, I think from that trend. Yeah.
Next question, HSBC, Joy.
Hey, morning, Hwee Long and team. Just a few clarification. First of all, in terms of the Guangdong Data Centre, this is a new tenant that is renewing the lease. Are we going to see 100% occupancy or they're still staying at 80%?
Okay. Hey, thanks, Joy. I think there are two parts to Guangdong Data Centre. Guangdong Data Centre, there's a renewal for the existing customer. We have also managed to bring in new customers. If you recall, I think last time historically, we used to be operating at 100% occupancy, and now we are at about 80%. You do expect that 80% to go up with the new leases that we managed to secure in the past quarter. I think those new leases are pretty interesting. I think we have managed to secure quite good rates. As a result, of course, I think the new lease will, I believe, the first one will only start to take effect towards the end of this year in terms of the customer coming in.
Post that, what you can expect is that I think in terms of the income profile for Guohua itself with the new lease coming in, you should see, I think more than doubling of the existing revenue contribution from what you see in place. On top of that, I think we still have additional capacity to sell. On that front, again, I think we do have quite a few interesting leads that we are talking, engaging quite closely on. That one, have a bit of patience. Give us some time. We will do our best to negotiate the best outcome, and hopefully, I think if we are able to get some of those across the line, we should be quite close to full sell-out for Guohua in 2027.
I see. That's very helpful. Second question, in terms of the lease expiry in the next 18 months all the way till 2027, what percentage is actually coming from SGP 1?
One second.
Sure.
Okay, never mind. Yeah, I can get back to you on that question later on, if you don't mind.
Okay, can. Then one last question. Within your current portfolio, do you still have untapped power capacity?
Untapped power in the sense that we have power available but unsold, is it?
Not unsold, yeah.
Oh, unsold, yes.
No, sorry, not unsold. I think you haven't maximized your power when you first built the DC.
Oh, okay. From that perspective today, at this moment, no. Actually, I would say, that's why you look at it, the new metrics that we share, the 95%.
Yeah.
Contracted power capacity, that will show you that today we have about 5% of the portfolio power capacity that is so-called unsold.
Yeah.
In terms of any incremental, no, nothing that we have firmed up yet. I think that will be a little bit more, ad hoc in a sense. For example, if we are through some of these power intensification conversations, if we manage to get a positive outcome out of that, then only in those instances there will be a so-called jump. I think at this juncture, don't have. I think we are operating within the power that we have on our hands today.
Okay. Got it. Thank you.
Thank you.
Next question, UBS, Terence.
Hi. Good morning. Just on the reversions outlook for Singapore assets, is it fair to say that we have mostly repriced those leases up from their low bases such that, I guess when we think about reversions from here on, it looks like we are probably moderating away from the high 50% level seen prior now towards the teen levels, which I presume more so reflects the rate of market rent growth. Would it be a fair statement?
Okay. Hi, Terence. Thank you for the question. In terms of reversion, again, that one you have to look at the nature of contract. The big ones tend to be the most significant ones. If you look at our overall WALE, of course, the good news this quarter is that we managed to improve the overall WALE profile up to 6.7 years. Of course, if you look one layer down between colocation and the master leases, the colocation are the one that we keep it short so that we are able to better flex our muscles when it comes to driving organic growth. On the colocation part is about three years or thereabout. And essentially what that means is that for colocation part of the business, we are operating on a 3-4 year of renter cycle.
The current cycle that we have with the volatility globally today, we are quite fortunate that the current cycle, the bulk of the big renewals, we managed to do it in the past two years. That also suggests to you that the next so-called cycle probably, it should come maybe a couple of years later. On that front, in terms of how much we can push, again, it's a bit of crystal balling. It depends really on how the market continue to trend. Given that we have a very heavy presence in Singapore on the colocation side, and we do think that from market dynamics, the constrained supply type situation should still persist. We should still see pretty good outcome coming out from the portfolio in the years ahead.
The other part probably to also consider is, probably to your question, is really have we so-called mark to full market today? A part of us is that, you look at the portfolio that we have today, I believe there's still some level of under rentness in a sense which will put us in a good state when the next big rental cycle comes about. Yeah.
Okay. Maybe this is now just a broad question. If you could just remind us again, assuming inference workloads scale from here, is there any direct mechanism that the REIT will benefit from, be it increased power utilization, or is it going to be more indirect, like counting back or based on the, I guess, periodic lease negotiations and reversions?
Yeah. No, I think the utilization, we are limited by the amount of power that we are being allocated. If we have, for example, 10 MW, 10 MW is all we have to sell. The utilization will not necessarily increase. What that means with AI inferencing shooting up from a potential demand standpoint is that it adds on incremental demand layer which means that for every 1 kW of capacity that we have out there for sale, previously maybe you have X kilowatt of demand. Moving forward with AI, it will mean X plus Y kilowatt of potential demand, which I think from that standpoint, hopefully what that means is that it will give added pressure in terms of from a pricing standpoint. Yeah, which I think will be quite beneficial for existing owners of data centers.
Okay. Thank you.
Yeah. Thank you.
Our next question will be [Ke Gan] John.
Good morning. Thanks for taking my question. First question relates to Cardiff and your overall power capacity utilization. Your answer seems to be alluding that Cardiff has a lot of physical space, but maybe small in power capacity. I hope I read that correctly. My question is on your contracted power capacity, 95%, is it for first half? Could we have the number for 1Q vis-a-vis 2Q so that we can see how that has changed? My second question relates to how power is allocated. It's just my curiosity. Say, if you have SGP 1, say, down because of redevelopment, if you do have a data center nearby, could that power be temporarily, 2-3 years timeframe, allocated to another data center nearby?
Because it is maybe same geographical location, would the utility company allow you to maybe use some of that in another data center nearby? Thank you.
Okay. Hey, John. Thanks for your question. On the first question, you are spot on. Cardiff is one of those whereby from a space perspective, quite significant, but power capacity, very small. That's why I think from a so-called contracted power capacity standpoint, first half is 95%, right? If you were to have a rough gauge between 1Q, 2Q, I would say the number will still be quite stable. Because of the small power capacity for Cardiff, so it doesn't really shift that 95% number significantly. In terms of your second question on the so-called power allocation, that one again, I think is very specific to individual scenarios in that sense. Because there's a lot of moving parts to shift power temporarily to another site.
I think the conversation will have to be had with the grid company in terms of, one, whether the new site that we are proposing to shift it across to in that sub-location, is there sufficient transmission capacity, power generation capacity, et cetera. On the assumption that there is, and us getting so-called green light from the appropriate authorities. I think potentially that can be done. Of course, I think there's a lot of hoops in terms of actual execution that we need to get clearance from before that can be effected.
Okay. If I may just have a short follow-up regarding outlook for 2027. You have maybe like 3.2% from co-location facility. Could you share if there are any large leases in your expiry for 2027?
2027, I think it will be quite similar to 2026 in terms of the profile. I don't know whether the leases are big or small, so it's probably more of the smaller leases. Again, back to the way that our WALE profile for the co-location works. I think you'll start to see more significant leases probably, in subsequent years, like 2028 onwards. 2027, there will be some probably more gearing towards the small to mid-size type of renewals.
Yep. Thanks for the color. Yeah, very helpful. Thank you very much. Thank you, Hwee Long.
Thank you.
Maybe our last question, we have two minutes, so we can take it, RHB Vijay.
Yeah. Hi. Morning, Hwee Long and team. Thank you for the opportunity. Just two questions from me. Firstly, in terms of Keppel DC8 and 9, I recall that there was some redevelopment potential of adding a floor into this asset. Maybe can you elaborate where these plans are at this point of time? From a redevelopment perspective, is Keppel DC1 the only asset, or you have other assets that you can redevelop? That's my first question. My second question, in terms of rent reversions, for the second quarter, 5% rent reversion, can we know where this came from? Since this rent reversion for Singapore is jumping 10%- 50%, would you be able to give some color in terms of what is the gap between spot rents versus your in-place rents for your assets in Singapore at this point of time? Thank you.
Okay. Thanks a lot. Okay. I think few parts. I think first thing in terms of the seven and eight, you are right. I think when we acquired, there was three empty data halls that we have. Of course, I think from a power capacity at that point in time, we have so-called fully utilized in that sense. Out of the three, I think what we have announced, I think towards the end of last year was that I think one out of the three, we are able to unlock it. Not because we are able to get new power, but more because we were able to post stabilization. We were able to further optimize, I think, how we run the asset. As a result, able to improve on basically your utilization efficiency such that we can squeeze out additional saleable power capacity.
That allow us to unlock one data hall last year, which we are now in the midst of fitting out. That will come online, I think in the second half of 2027. We are very well on track from that perspective. Of course, I think what happened to the remaining two data halls. I think today, as we are now, I think we do not have ability to bring in additional power yet. I think in the future, if we can bring in new power, obviously, I think those will be the very low-hanging fruits that we can fit out and turn those into saleable data center space. That is the status. I think in terms of your second question pertaining to, I think for SGP1, that's a strong, I think, redevelopment type of candidate.
Apart from that, I think if you look across the rest of our portfolio, I would say nothing so immediate yet. The rest are still, I think, strong, still churning, I think good income for us. In terms of the last question pertaining to so-called spot versus in-place, I think on that front, I'll probably link back to my earlier commentary. You asked me, I think our portfolio today our assessment is probably still under-rented compared to spot rent. In terms of the so-called magnitude of that, I think we have to look at individual leases because some of the smaller ones will be obviously quite close to spot rent. Some of the bigger ones, I think those will have a lot more avenues for us to push harder in future rental cycles.
Got it. Thank you. That's all I have.
Yeah.
Thank you. With that, we have come to the end of Keppel DC REIT Webcast. Thank you for joining us today. Have a pleasant day ahead.