IFT - Infratil

[Jaa]

IFT sold it's stake in POT.

What return did they miss out on?

IFT sold POT down over time, so would take a lot of work to calculate, they timed a sale well in 2021 but it didn't all go smoothly. From 2002:

Infratil is selling down its stake in Port of Tauranga, saying it needs to free more capital for new investments such as overseas airports.

Infratil trades sea for air

Probably a lesson there.

[kiora]

You are welcome Kiaora, Sharesight is great, tracking against the benchmarks keeps you honest as an investor. A Kiwi success story too.

Talking about success, I found another 20 year winner, Southport at 11.24%. Why bother with IFT's high finance, capital raises and wheeling and dealing when you could just own a boring NZ port huh?

Thanks Jaa
I'll definately look at Sharesight a lot more.Very Useful

I did own POT for a while but earnings growth wasn't great.Looks like been pretty OK returns since
Same NPT,sold at start of Pandemic when they ere building a new wharf

Glad I did,1 % compounding !

[kiora]

IFT sold it's stake in POT.

What return did they miss out on?

13 % compounding?

[kiora]

Oz election commision moves from NextDC to CDC

The market gives with one hand and takes away with the other

Makes sense?
National Data sovereignty & security?

[Snoopy]

I didn't even look at the chart and now I have, neither should you. No one really knows what the growth in data centre demand will be so there isn't much point in analysing it to death.

I think there is a middle ground between 'analysing the demand for data centres to death' and 'zero analysis' though.

I get the impression that Carl Capulingua who wrote the livewire article "Wondering how to Invest in AI? Take a look at ASX data centres"
https://www.livewiremarkets.com/wire...x-data-centres

had already decided that demand for datacentres would grow in the future exponentially. So he just cast around for a graph showing 'exponential growth of data' and put that in his article. However, he somehow failed to notice that the graph instead showed where data was being created, not where it was stored. And it was clear that despite data creation at datacentres growing exponentially, it was still well behind data creation not in data centres that was also growing exponentially. So the graph did not necessarily show a huge increase of demand for physical datacentre footprints. And this alternative view is supported by another reason you mentioned Jaa:

Assume that data growth will continue and humanity will find more things to do with computing power. Moore's Law means this doesn't necessarily require more data centres as processors and storage are constantly becoming more powerful, efficient and cheaper. To give you some idea, someone tried to prove you could run Twitter on a single modern server. Likewise software, GTP 4o is much more efficient than GTP 4 for example.

Improvements in processing and storage can keep occurring, without constantly increasing the space to house such improving equipment. So it looks to me as though Carl Capulingua put zero effort into researching the industry background in his article, and he didn't even realise that he put up a graph that if anything was evidence against his biased self created opinion of 'demand for the footprint of physical data centres always growing.'

There is a nascent but growing movement away from Cloud due to its increasing complexity and opaque and ever increasing bills to owning hardware again via colocation which could be beneficial for the likes of CDC.

Now it looks like you have confused yourself Jaa. 'Private cloud' where a company puts their own equipment into a dedicated remote data centre (this appears to be the model operated by CDC) is still 'cloud', is it not?

Basecamp, one of the first SaaS companies and considered thought leaders, reduced their cloud bill from $US3.2m/yr to a one off buy of $US600k in modern hardware and $US7m in expected savings over 5 years and have declared hardware is fun again.

Basecamp is an example of a company uprating the hardware on their private cloud in a 'single rack server'.
https://world.hey.com/dhh/hardware-i...again-b819d0b4

At least that is how I read the article at first (good for the likes of CDC). But I suppose it could equally read that Basecamp could operate their own small server on their own premises away from any data centre, bought for $600k?

This interplay between hardware and software is very hard to predict and thus so is data centre demand. Very different from IFT's traditional infrastructure investments.

Is the interplay between hardware and software another way of looking at the shifting balance between 'data storage' and 'data creation'? As you can see, I am still trying to get my head around what 'data centres' are really about.

SNOOPY

[Snoopy]

Oz election commision moves from NextDC to CDC

The market gives with one hand and takes away with the other

You have shares in both Infratil and NextDC?

From the referenced article:
"The AEC spokesperson said the timing of the data centre migration is "all about ensuring our IT election ready systems remain safe, stable and secure.” "

That sounds like NextDC have not updated their security protocols, and they have old unstable equipment? What were they thinking at NextDC?

"CDC Data Centres won the contract through a competitive tender to participants of a Digital Transformation Agency-assembled panel."

"The AEC spokesperson told iTnews that it is “taking the opportunity” presented by the move to a new facility to also “downsize” its on-premises infrastructure. Its data centre hosts core WAN infrastructure, unspecified “internal business systems” and election delivery services. The intent is to make “further investments into cloud capabilities” as the AEC shrinks its data centre space requirements."

This sounds like the AEC are moving more of their 'on premesis' computing equipment into the (private) cloud. But furthermore as computing power continues to increase and computer processing equipment shrinks in size, AEC are going to put more computer grunt into the same rack space. Thus they are planning to increase their computing power without the rack space rent going up - very clever. This is a marketing advantage of 'private cloud' (seems to be the CDC model) that I had not considered before.

SNOOPY