I agree that, around Auckland at least, industrial land is a scarce commodity. So it does make sense to value such land on a 'best use' basis. Data Centres are near the top of that 'best use' list, right now, as far as I can figure out. And you can measure that by looking at the cash-flows being generated from them.
But what about that vacant paddock of land that you have purchased out the back of your data centre? How do you value that? You would have to make guesses as to:
a/ When the ramp in demand was sufficient to require a data centre to be built on that land. (PRFY2014 'slide 11' suggests that Spark are looking at building 118MW of new data centre capacity within the next 5-7 years. C.f. current operating capacity of 22MW).
b/ What the profit margins would be like when the utility in those data centres was sold off, and at what rate that customer take up would occur.
This is where I take issue with CFO Stefan Knight saying that he thinks the IRR on his Spark data centre projects will be 10% (that part of his modelling IMV is believable) but for the uncommitted riskier part of the build out, the IRR would be 15% (IMV not believable).
I think I get where Stefan is coming from (see FY2024 Presentation, slide 11). He is saying that in order for the future 'pink and grey' part of the potential data centre site capacity to be built, he wouldn't do it unless he could get customers signed up to contracts generating an IRR of 15%,
But I come back to the Madison Reidy interview with Jason Boyes on the subject of data centres, and the comment below by Jason
------------------------
https://www.sharetrader.co.nz/showth...=1#post1065389
9:30 Jason Boyes CEO Infratil: "I think it is prudent to assume that (the data market) does become more competitive in the future. It is how we think about any investment.
You should not assume your margins will stay the same forever. You should assume they should contract to a pretty stable norm over time. That is a sort of planning for the worst. Then there is a sort of
hope for the best, which is down to how you manage your business keeping an eye on what demand and supply really is and whether or not we are reaching that point where actually there is going to be a lot of competition, and we should look to move on. And that is the power of Infratil. If over time you don't like the sector you are in, you can move on. And this is definitely not one of those ones in moving on territory at all, but that is the flexibility we have at any given time."
-------------------------
I get the impression that Spark CFO Stefan Knight is definitely in the 'hope for the best' camp. Don't get me wrong, there is nothing wrong with Spark being optimistic about their datacentre prospects. But I find Stefan's argument partly unravels when he talks about known datacentre customers generating an IRR of 10%, whereas with new customers he expects an IRR of 15%. So as new customers sign up they become 'known' and so their expected IRR to make the sign up viable drops from 15% to 10%. Huh? Maybe I am just 'data-centre dumb' and don't understand the business model. But it looks to me as though Stefan is banking on more data centre capacity demand and higher margins, just as the competition from the international mega datacentre providers is ramping up in parallel. I find Jason Boyes more measured outlook of data centre margins contracting to a more stable norm over time more believable. Am I being overly negative with my thoughts on the future profitability of Spark datacentres?
SNOOPY
Bookmarks