IFT - Infratil

[Snoopy]

That Livewire link had some details Snoopy. More about using their existing land bank and developer expertise/contacts than converting existing sites.

Thanks for the link Jaa. There is a graph in the link titled 'Exabytes of data created (Created in the cloud vs non-cloud)'. I think that is a global projection rather than Australian one, although the text doesn't make that entirely clear. But something doesn't look right about it to me. The graph does not tie in with the text

Why is the graph illustrating 'created in the cloud' instead of 'stored in the cloud'? Are the two meant to be synonymous? If we have something 'created in the cloud', say via a SAAS application, are we talking about the computing power needed to run such an application in a data centre? 'Created in the cloud' isn't mentioned in the text. The text talks about storage. So why illustrate the prospects of 'cloud storage' (concept 1) by graphing data 'created in the cloud' (concept 2)?

Then there is the bold header statement in the article that says: Cloud + AI = Storage (and lots of it). Now traditionally if you have a computer system in your office then that is 'non-cloud' whereas stuff stored over the internet on a server somewhere else is 'cloud'. But the header in bold would almost suggest that in this instance 'non-cloud' is referring to AI in this context? A later comment in the text when talking about cloud is that:

"Certainly, cloud storage is the mainstay, representing around 60% of global centre demand. AI is likely to be around 25% of the mix with the remaining portion largely taken up by data sovereignty. Data sovereignty is perhaps the least known of the three but is arguably the most important. It is tasked with protecting data from “external interference from foreign states and third parties."

So this would suggest that AI is indeed part of cloud being talked about in the article. This is saying: Storage (60%) + AI (25%) + Data Sovereignty (15%) = Cloud Capacity (100%)

But if that is what the article is talking about, it is in direct contradiction to the header: "Cloud + AI = Storage (and lots of it!)". What is going on here?

Next, the graph has split 'Storage' into 'Cloud' + 'Non Cloud'.

My reading of that graph is that 'data created' will increase from a total of 129,361 Exabytes, made up of Cloud Created Data 30,000 and Non-cloud Created Data 99,361 (2023) to a total of 291,122 exabytes Cloud Created Data 100,000 and Non-cloud Created Data 191,122 (2027). I calculate the respective growth rates as follows:

Cloud Created Data: 30,000(1+g)^4=100,000 => (1+g)^4=3.333 => (1+g)=1.35 => g=35%
Non-cloud Created Data: 99,361(1+g)^4=191,277 => (1+g)^4=1.925 => (1+g)=1.18 => g=18%
Total Created Data: 129,361(1+g)^4=291,122 => (1+g)^4=2.250 => (1+g)= 1.22 => g=22%

The writer of the article then says
"Certainly, cloud storage is the mainstay, representing around 60% of global centre demand."

But the graph shows the exact opposite of that. It looks to me that although demand for cloud created data is growing faster, even by 2027 it will only amount to 30% of all data stored, the remaining 70% being 'non-cloud'. So it looks to me as though despite the huge growth of data centres, still by far the largest opportunity is to sell the computing power for creating data 'off cloud'. Which then creates an opportunity for that data to get stored 'on cloud' (or does it)?

Jaa I have a feeling you are a lot more au fait with this datacentre stuff than me. The article seems a bit fast and loose with definitions. The net result being I can't understand the article. What on earth are they trying to say? I get the overall message that 'investing in data centres is the greatest opportunity'. But then the content in the graph in the article appears to not support that proposition.

Can you tech-translate for me?

SNOOPY

[Snow Leopard]

....
I also have no idea why they are throwing a lazy $US100m at "IoT, Big Data, and Security" venture fund, Clearvision Ventures. There goes half the money from the retail offer. Investors are paying fees on top of fees on that one. The only investment they disclose, Chargepoint Holdings has done terribly. Must be cool to visit Sand Hill Road and chit chat though.

From the Infratil Annual Report:

"The strategic objective of the investment is to help Infratil's businesses identify and engage with technology changes that will impact their activities."

The 50% gain on funds employed so far is just a bonus

Now if you thought that AI would be useful for something it would be digging up the details on private venture funds but it turns out AI is as useless at that as it is most other stuff.

But I do know that Clearvision have made at least 20 investments and I am pretty confident that they made a profit when they exited Chargepoint just over 4 years ago.

[fastbike]

Refunds have been paid today.

I got 1.077 of the allocation 136/1000 in the prospectus, however the small amount of over subscription has been credited back today, so no complaints.
Apparently my average purchase price is NZD 5.646, so I appear to have been fleeced

[kiora]

I applied for less than the 136/1000 otherwise the RC would have been savaged & got 100% of applied + $3.25 refund.
I see the Morison shareholding has been diluted down as well.

[huxley]

I applied for less than the 136/1000 otherwise the RC would have been savaged & got 100% of applied + $3.25 refund.
I see the Morison shareholding has been diluted down as well.

- $3.25 right? They don’t refund amounts under $5.00

[kiora]

- $3.25 right? They don’t refund amounts under $5.00

You are right Huxley.I didn't read the fine print & the $3.25 isn't in the bank account.

"2 In accordance with the Offer Document, your refund will be issued within five business days following the Allotment Date. Any amount lesser than $5 due to rounding or scaling will be retained by Infratil."

[Jaa]

"The strategic objective of the investment is to help Infratil's businesses identify and engage with technology changes that will impact their activities."

I was trying to be glib Mr Leopard... but gee identify and engage sure does sound like expensive chit chat.

The reality distortion field of venture firms mean they only disclose profitable exits. IFT did post bubble bingo winning news on Linkedin though, I bet most shareholders have no idea that IFT is back investing in buses. Guess with your views on AI you won't be letting it drive the kids to school?

One of the ways Infratil seeks to identify and engage with technological change and disruption is through its investment in Clearvision Ventures. Clearvision invests out of Silicon Valley in early-stage companies, focusing on innovations in IoT, Big Data, Security Technology, and tech-enabled sustainable infrastructure.

One of these investments is Zūm. Zūm provides modern transportation solutions aimed at transforming school transportation.

..

Zūm is providing a fleet of 74 electric school buses and bidirectional chargers in Oakland, managed through its AI-enabled technology platform. The all-EV fleet will transport students sustainably and play a critical dual role as a Virtual Power Plant, giving 2.1 gigawatt hours of energy back to the grid annually.

While Infratil’s look-through investment in Zūm is modest, it is an investment that provides valuable access to the thinking and technology driving an emerging class of tech-enabled sustainable infrastructure.

https://www.linkedin.com/feed/update...5729899130880/

I also note that the founder of Clearview Ventures, Dan Ahn runs an identical looking VC fund called Envision Ventures out of the same expensive office on Sand Hill Road. Wonder which fund gets the best deals? See if you can spot the difference:

https://www.clearvisionventures.com/
https://envisionvc.com/

[Jaa]

Thanks for the link Jaa. There is a graph in the link titled 'Exabytes of data created (Created in the cloud vs non-cloud)'. I think that is a global projection rather than Australian one, although the text doesn't make that entirely clear. But something doesn't look right about it to me. The graph does not tie in with the text

But the graph shows the exact opposite of that. It looks to me that although demand for cloud created data is growing faster, even by 2027 it will only amount to 30% of all data stored, the remaining 70% being 'non-cloud'. So it looks to me as though despite the huge growth of data centres, still by far the largest opportunity is to sell the computing power for creating data 'off cloud'. Which then creates an opportunity for that data to get stored 'on cloud' (or does it)?

Jaa I have a feeling you are a lot more au fait with this datacentre stuff than me. The article seems a bit fast and loose with definitions. The net result being I can't understand the article. What on earth are they trying to say? I get the overall message that 'investing in data centres is the greatest opportunity'. But then the content in the graph in the article appears to not support that proposition.

Can you tech-translate for me?

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.

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.

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. 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.

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.

[Jaa]

I'm always interested in investing in companies in NZ that reward their shareholders with higher compounding returns than IFT has over 30 years.

Do you know of any?

IFT have done well Kiora but they are not the only option.

Using the share checker in Sharesight, which includes dividends and capital gain, for the max 20 year period, a few shares I checked beat IFT's 11.85% p.a total return over the last 20 years. They are FPH, MFT, RYM and EBO. POT and AIA were both close too, with 11.19% and 9.8% returns respectively.

Obviously past returns don't guarantee future performance which is why you should keep an eye on your investments.

[huxley]

I for one look forward to reading the the comments about what constitutes infrastructure when the Console Connect transaction goes through