Australia's Biggest Pension Fund To Cut Global Stocks Allocation on AI Concerns (ft.com) 29
Australia's largest pension fund is planning to reduce its allocation to global equities this year, amid signs that the AI boom in the US stock market could be running out of steam. Financial Times: John Normand, head of investment strategy at the A$400bn (US$264bn) AustralianSuper, told the Financial Times that not only did valuations of big US tech companies look high relative to history, but the leverage being used to fund AI investment was increasing "very rapidly," as was the pace of fundraising through mergers, venture capital and public listings.
"I can see some forces lining up that we are looking for less public equity allocation at some point next year. It's the basic intersection of the maturing AI cycle with a shift towards Fed[eral Reserve] tightening in 2027," Normand said in an interview.
"I can see some forces lining up that we are looking for less public equity allocation at some point next year. It's the basic intersection of the maturing AI cycle with a shift towards Fed[eral Reserve] tightening in 2027," Normand said in an interview.
Smart people (Score:2)
Well, somewhat. To not see the bubble and how there is no way in Hell most investors will even recover their investments requires you to be deeply in delusion at this time.
Re: (Score:3)
Oh? Really? So the dot-com bubble never happened, and didn't wipe people out? So 2008 didn't happen, and didn't wipe people out?
You're an idiot.
interesting question (Score:1)
S&P was $10 in 1940, and $6,845 end of 2025, for a ratio of 684.
Stocks are 30 times bigger than inflation. You are correct.
Re: Smart people (Score:1)
Do you think they'll switch to US Treasuries?
Re:Smart people (Score:4, Insightful)
I think the telling part of TFS is:
> but the leverage being used to fund AI investment was increasing "very rapidly,"
The leverage is indeed what's driving this (and all bubbles). Companies getting an investment then get to puff up their notional (or real) share price, which they then use to borrow against. They then invest that borrowed money in some other company, who does the same somewhere else... and so it goes on - perhaps circularly back to the first company in the chain.
The problem with leverage is it's a multiplier. It just takes a little wobble somewhere in the chain, and then whole thing collapses pretty spectacularly because the leverage multiplies the risk, multiplies the downsides etc etc. The companies who are leveraged really can't influence the outcome - if the leverage collapses, then so do they. What's left is arguably the "real" value of their businesses, and in a lot of cases will be almost nothing.
I'm not up to doing this, but I'll bet these analysts have already looked at (say) OpenAI and tried to figure out what the "real" value of that business is. That is, take out all of the leverage and look at assets, revenues and so on. All companies are probably over valued on that basis, so that alone isn't really the problem, but how much they're overvalued, and the dependency on other businesses to maintain that value definitely are a concern.
As for the Australian pension funds... if AI "pops", it'll take a lot down with it, but that doesn't mean there aren't companies worth investing in, or companies that are likely to be insulated from any fall out. Some of those companies may well even be American, European, Asian etc, so "reducing allocation of global equities" may be a bit more extensive than necessary (just because it's an Aussie business that doesn't mean it won't struggle).
Re: (Score:2)
Yes, that is basically how it works. The first steps still have some substance, but then hot air creates more hot air.
As to the analysts, they have clearly carefully looked and found a catastrophe in the making. And they did that some time ago and they convinced the ones deciding about the actual investments. Because what we are seeing here is public statements. And these public statements are typically made very late in the process and long after the decision making. They are made right about when the move
It usually starts out as a trickle.. (Score:5, Interesting)
but ends in a stampede. Some people are going to be left holding the bag... of manure.
OpenAI is projected to have a revenue of a little less than $13B in 2025. That is revenue. Not profit. They didn't have a profit for this year, but instead had a loss of $16B. Their spend commitment is $1.5T!!!!
This lunacy is not limited to OpenAI. People have lost their collective minds. This is not going to end well.
Re: (Score:3)
It really more or less is limited to OpenAI. If their investors lose faith it can all go away overnight.
Anthropic might be in a little trouble, but Gemini will be fine and so will Grok. Nvidia is obviously going to be sad that they're no longer printing money.
Re: (Score:2)
Why do you think will Nvidia stop printing money? People are only waiting for the resources allocated to OpenAI & Co. to become free, so they can buy GPUs. Not talking about the gamers, but about the data centers of smaller companies who also need GPUs. Should one of the giants stop buying, Nvidia has a lot of other interested companies who are waiting for GPUs.
Re: (Score:2)
Even expensive GPUs are cheap compared to AI accelerator cards, and the profit margins are lower (in percent) as well. An H200 card reportedly goes for $30,000 and up, about 20 times as much as a RTX 5080. Nvidia's margin (in percent of sales price) is probably two or three times as much for the H200 -- so they need to sell a hugely larger number of GPUs to make the same profit. If AI demand went to near zero, Nvidia would still have a very healthy business but their profits would drop enormously.
Re: (Score:2)
Those AI Accelerators might as well be called ML accelerators, once you remove the hype. The market may lose interest in LLMs but the applications for machine learning are real.
There are lots of enterprises with plenty of data to crunch. One thing that has surprised me so far is I have not seen a lot of support for these thing is MOLAP products or in traditional DBMS systems for ROLAP work yet. With the right storage (read giant memory caches) you could do dimensional data on top of normalized data withou
Re: (Score:2)
Name a specific problem and couch it in terms of memory bandwidth, memory capacity, and FLOPS -- then we can figure out whether GPUs are a good fit. Arm.waving about "data warehouse capabilities" doesn't cut it. Most of the time, the difficulty will be in getting the data you want into the database you want to process it, rather than how far the database can give you an answer.
For example, people don't tend to put databases on GPUs unless they do full scans of the data multiple times each second, because
Re: (Score:2)
What about weather simulations? You never have too many GPUs for those.
Re: (Score:2)
AI demand won't be near zero just because one of the big companies is out of business. The others would share the market then. But even if it were zero then, then you could still do a lot of other computations on a H200.
It's a bubble (Score:4, Informative)
Get out as soon as you can. The ultra rich are the one ones that are going to walk away unscathed.
Re:It's a bubble (Score:4, Informative)
It seems he has his dirty little hands in quite a few places:
https://www.nytimes.com/intera... [nytimes.com]
It is sickening to read. One of his latest gambits, admittedly small potatoes, is his defrocking the committee running Wash. D.C.'s golf courses. He intends his golf company to run them instead. They will squeeze out the common folk and the only ones allowed to play are the ones who can pay the exorbitant fees membership in one of his stupid clubs requires. There's no enterprise small enough that he cannot see fit to destroy it.
Re:It's a bubble (Score:4, Interesting)
Get out as soon as you can. The ultra rich are the one ones that are going to walk away unscathed.
I'm less concerned about getting out unscathed as I don't have a dog in that fight... However as I'm not a member of the ultra rich (just a peasant who's managed to do OK for their humble beginnings) I'm worried about the aftermath.
The AI bubble is the only thing holding the US economy above water. With Trump's tarriffs screwing everyone who's not a billionaire six ways from Sunday, his financial illiteracy destroying what little that isn't screwing over and his overlords more concerned with keeping the ultra rich onside as they proceed with their Great Re-Whitening... Once the AI bubble bursts it's going to be carnage, metaphorically with huge job losses (which have already started) and possibly literally with the huge amounts of guns in the hands of not so stable people.
As someone living in the UK, we'll be somewhat insulated from this as business flees the US, it'll come to the UK and EU first, with Canada, Japan, Australia and a few others getting some as well. This issue for me will come as aftershocks, a lot of British businesses have swallowed the Cloud and AI fads hook, line and sinker so they'll struggle to get away from large American tech corporations who will be desperate to fleece whoever they can (and it's not like I didn't warn them when they sold off their local infrastructure for "cheap" cloud). Also the fact that we'll see a lot of Americans immigrating over to Europe, meaning more competition for the good jobs.
Re: (Score:3)
You act as if your economy won't follow ours. If the US economy goes into a major recession or a depression, the world will certainly follow. Even with all this tariff business going on, we are still too connected as a global economy for this to not be true.
So hope we can avoid a bad recession because it's bad for everyone.
Re: (Score:3)
I've been reshuffling my retirement savings to mimimize exposure to tech stocks. All mining and Agriculture, because by my reckoning, no matter what happens people will still need to dig holes and eat food. Fairly safe imho.
caveat: Not american agriculture however. Those Tarifs are doing *stupid shit* to the sector.
Pop! (Score:2)
...goes the bubble.
Let this news be digested by Wall Street later today. Others will follow suit.
Re: Pop! (Score:1)
Are you saying the sky is falling?!?
Re: (Score:2)
So are you shorting them?
Re: (Score:2)
Who are "my friends"? Are you doing conspiracy theories?
And there is no problem shorting something after telling it will crash. There is a problem shorting it after NOT telling it will crash but knowing over unofficial channels it will. Here every little news site tries to predict a crash, shorting it would not get you accusations but credibility.