Digital Core REIT

[swakoo]

From what has already transpired, it seems that the goose may have been chosen, exported and slaughtered for its meat. But that doesn't mean that the plate has no braised goose meat left though.

Admittedly, have not followed DCR in detail to be able to add to your interesting observations.

My post of a quick and dirty rough observation of huge gap in distribution yields (and it's potential implications in my mind) springs from cynicism (which may possibly not be warranted) of vendors choosing to list US assets in Sg instead of Wall St.

But glad to hear you think there may be some braised good meat left, for existing DCR unitholders.

[weijian]

From what has already transpired, it seems that the goose may have been chosen, exported and slaughtered for its meat. But that doesn't mean that the plate has no braised goose meat left though.

But glad to hear you think there may be some braised good meat left, for existing DCR unitholders.

Hi swakoo,

I will have to clarify - I only said there may be some braised meat left. I have not commented that they are "good". On the contrary, because these are data center assets and with higher demands of power, water and cooling required to operate increasingly complicated computing needs, I would think that technological obsolesce is a great risk.

But as investors, our job is to look at mis-pricings and not just whether the asset is good or not. Things get interesting when my "it doesn't make sense" radar detects something and I will decide to commit more time to look into it. That is mainly where I am coming from.

[swakoo]

I will have to clarify - I only said there may be some braised meat left. I have not commented that they are "good". On the contrary, because these are data center assets and with higher demands of power, water and cooling required to operate increasingly complicated computing needs, I would think that technological obsolesce is a great risk.

My typo error - meant "goose" not "good". Think  operative word here is "some".

But as investors, our job is to look at mis-pricings and not just whether the asset is good or not. Things get interesting when my "it doesn't make sense" radar detects something and I will decide to commit more time to look into it. That is mainly where I am coming from.

Yes, knew where you are coming from. Good play.

[swakoo]

News just out of CFA2 allocation:

Digital Realty, STT GDC among data centre operators awarded 50 MW of new capacity in Singapore

Potential pipeline for DCR of prized Sg DCs.

[weijian]

DCR recently had its Investor Day - giving forward guidance for DPU (which would imply almost a 10% dividend yield!) and also revealed its remuneration framework that it believes will align Mgt to OPMIs.

But I thought the most interesting portion is Mgt's view that stabilized data centers should trade at 2x replacement cost and DCR trading at ~replacement cost, is therefore "a bargain".

Valuation 101 teaches us that bargains on real assets happen when it is traded at below replacement cost. Therefore, this runs contrary to what DCR Mgt has implied.

So what are the conditions for a real asset to be traded at 2x replacement cost and yet deemed to be "fair valued"? First, this real asset must have really high ROICs - much higher than the cost of capital (which itself is in the range of high single digits to low teens) - probably something in the range of high teens to low 20s sort of range. Next, I would think that 2x replacement cost will also demand the high ROICs to be of "high quality", meaning that rents should either be increasing in the future or it has a long useful life time, so that the future cashflows discounted back will still have high present value. In a free world, high ROICs attract competition which eventually cause some sort of mean reversion for high ROICs. Finally, replacement cost is the theoretical cost to replace it. But what if you actually can't replace it? In this case, the replacement cost becomes just a mathematical exercise. Currently, there are well publicised regulatory push-backs and power constraints for new data center building. And even if the data center is built, there are also order backlog for the PPE to fill them up. Whether how long such practical constraints last, is another question again.

Investor Presentation

P. 12: FAVORABLE FUNDAMENTALS AT DISCOUNTED VALUATION

Cushman & Wakefield recently published an interesting report on data center construction costs, which we have attempted to summarize in the chart on the left. According to Cushman, the all-in construction cost per megawatt has increased by 21% over the past two years. As you can see from the chart, there is a fairly wide range between and even within markets, but the midpoint is right at US$16 million per megawatt. At the current unit price, Digital Core REIT trades at an implied value of almost exactly US$16 million per megawatt.

In either a U.S. or European market context, it is highly unusual for a stabilised data centre platform to trade at replacement cost. Data centre development has traditionally been highly profitable. In broadbrush terms, unlevered returns on data centre development have historically been in the 10%-12% range, where as cap rates on stabilised data centres have generally been in the 5%-6% range.

Said differently, stabilised data centres have historically traded at two times replacement cost, as visually represented by the chart on the right. The current valuation offers a unique opportunity to invest in a fully leased data centre portfolio concentrated in core global markets and managed by the leading global data center owner and operator – at replacement cost.

prepared remarks: https://links.sgx.com/FileOpen/Digital%2...eID=906669
ppt: https://links.sgx.com/FileOpen/Digital%2...eID=906630