I keep saying the US stock market, and quite possibly its economy, will crash, and this chart explains why.
*Foreign investors put $163 billion into US stocks in June* – headline.
If you want an indication of stock market volatility, that is it.
Wow, is all I can say. This is beyond hype and bubble territory: this is economic insanity driven by loose money pursuing profit unrelated to any real-world activity.
Does that create the potential for the bubble to burst? You bet it does.
US stock market is worth about $50 trillion.
0.3% is economic insanity. Ho Hum.
It sounds like spud has Doctor Evils’ grasp of numbers: “Why Make Trillions, When We Can Make… Billions?”
He’s been predicting bubbles in stocks / AI / crypto for years. No doubt he has written posts ready for when he is “right” and so he can claim the title of worlds most prescient potato.
Strangely the bubble he’s missed is the one in government spending which will soon burst
I’m one of those foreign investors. What is the prescient potato predicting for the UK economy? Does he think we’re going to do better than the US? Because I don’t.
That’s not the concerning chart (and I hate to say that Spud might be right). It’s the Buffett Indicator, the P/E ratio of the S&P 500. It’s how much of the value of it is about “AI” stocks.
You have ChatGPT and various “AI” services. These are losing money like sailors in a brothel on shore leave. These are running on top of Nvidia chips and other chips. These are hosted by various companies like Amazon, Microsoft, Oracle. This has not only led to growth of these companies, but also an assumption that the line continues. So, Nvidia have a P/E of 60. Microsoft have a P/E of 37. Oracle have a P/E of 48. Amazon are on 35. This isn’t being driven by predictions for more gamers, Microsoft Office, Oracle DB or Amazon flogging more USB cables. It’s all about AI chips, Azure, Oracle Cloud and AWS.
If the AI companies float and don’t make a profit, or run out of dosh, everyone figures out that robotaxis are as efficient as the Mechanical Turk, and down falls AI. You have a glut of GPUs. You get a fall in demand for these hosting companies. We could then be looking at many of these companies halving in market cap. That’s a 15% hit to the value of the index.
Then there’s the effect of Robin Hood apps and ETFs. These stocks have been heavily traded by retail investors. Who are also buying into S&P 500 ETFs because they perceive them as easy money, “line goes up”.. The impact of the 30% of the market in tech hasn’t just raised tech, but investments in the rest of it. You buy an ETF whose growth is being fuelled by this bull, you’re also putting some money into McDonalds, Ford and Wiliams-Sonoma.
Tech falls, it corrects to where it should, that’s a 15% market drop. Now, if ETF people who think “line goes up” start panicking and dumping those, you’ve got a full-blown crash. Not sure about the second part, but “AI” is going to fail. It’s a matter of when, not if.
Well, yes, but 15% is not a market crsh, that’s a market correction. And a 15% fall in the S&P most certainly isn;t the beginning of the end times sorta stuff that Spud is claiming it is.
…the end times sorta stuff that Spud is claiming it is.
Evidently, the fool sees himself as a Prophet of Doom – economic, environmental, political – for capitalism….I suspect he has messianic fantasies.
So, $163B in June – is this unusual? Or is this a typical month? And, assuming it’s unusually high, perhaps the headline is “European investors flee economic uncertainty at home, seek safety in US markets”?
Western Bloke
Great analysis -It’s like the ‘Dotcom’ boom from the early 2000s all over again, although with more money involved.
Just goes to show the old PT Barnum maxim still holds true.
Regarding Murphy he’s like a stopped clock on this issue. I think there will come a point when he’s proven correct but I don’t think it’ll be in his lifetime.
AI doesn’t just train itself on all the nonsense on the internet, it also trains itself on the user. Because it is programmed to maintain engagement it becomes a sycophant.
So if a retired accountant with learning difficulties uses AI the AI will quite quickly work out how to supply the nonsense required for a post. Probably write the comments below the line as well.
Esteban,
It’s mostly about people investing in tech companies which they perceive will be doubling money soon. There’s been a lot of money into ASML who are a Dutch company that make chip equipment for the same reason. Their share price has more than doubled in 5 years. Or there’s TSMC that make chips in Taiwan whose price has trebled in 5 years. People on various investing subreddits like wallstreetbets and valueinvesting talk endlessly about these companies alongside Tesla, Microsoft, Google etc. No-one is talking much about Build-a-Bear or Walmart
European stocks are doing fine. VEUA ETF is up about 12% this year. FTSE 100 is up 11%. That compares with about 15% with the S&P 500. Because the FTSE 100 is mostly a global index. Companies like Shell , BAT, Rio Tinto. FTSE 250 is more UK exposed. Even then, it’s a much lower P/E to reflect that, so might be good value.
I do wonder how much of the hype around AI is NVidia and AMD moving from selling cards to crypto miners to selling cards to AI companies, once the two largest crypto coins moved to proof of work.
Tim,
“Well, yes, but 15% is not a market crsh, that’s a market correction. ”
I checked the ETF number and actually, even that won’t make a huge difference as it’s only $2tn of that $50tn. Even if everyone sold them, it wouldn’t make much of a dent.
“Even if everyone sold them”
But then somebody else would have to buy them.
Western Bloke,
The other issue, part with regard to nVidia, is that Google et al, let alone the Chinese, don’t really want to be paying cash away to them.
They have their own TPUs under development, and for computer vision models at the edge, you’ve been able to buy Coral (USB or PCIe, maybe M.2 as well) for a while now. Other TPUs may be available.
nVidia’s moat seems to be the interconnects at scale, and the S/W dev kit.
The cunning little yellow devils may have breached the first, and I don’t see the second lasting.
And there’s AMD and Intel as well. Might be some way behind, but still.
@Western Bloke
There is a market for this sort of AI stuff. I’m hoping in 10 years to have a fully autonomous car so I can maintain my mobility when I’m to old to drive. Taxis (preferably robotic) would be an option, but my home is too rural to be able to spontaneously travel. Taxis would need to be booked well in advance.
The second thing I will need about then is a domestic robot to tend to my wife and I. Bought new for £40K, or taking an analogy to the car market perhaps leased for £5K per year, might keep my wife and myself out of a care home for a further 10 or even 20 years. As care home bills for a couple are currently in the region of £13K per month this would very quickly pay for itself, though using robots in care homes could mean slashing their employment costs and so the bills passed on to the residents “should” (ROFL) drop too.
My hope is that AI and robotics will displace a lot of the minimum wage jobs that no one wants to do. We are currently “inviting” immigrants into the country to do these unpopular jobs. The great thing about robots in these roles is they don’t have children or relatives to pull in with them. Filling the jobs with immigrants just kicks the problem down the line whilst making everything a bit shittier now due to the strain of increased population.
Not however that any lawyer advocating “robot rights” needs to be strung up immediately.
$2trn of $50trn is 4%. It’s a fairly significant chunk of market cap.
QQQ, SPY plus IVV seem to be a serious chunk of that $2trn, unless they weren’t included in the number.
Add in all the other trackers, plus the mutual funds like FMAGX (39bln apparently) many of which are highly likely to be closet trackers, and.. total cash in trackers could be well north of 8% of the market.
For the closed end funds, well, they do like the odd swap here and there.
I don’t have a particularly high degree of confidence right now.
Also, it seems that the option funds have made a reappearance. They had their own little death spiral last time around.
It’s the old adage: in a gold rush be the one selling them shovels.
Andyf
Fully automated cars are probably a pipedream. The technology is achieveable – particularly in urban areas with embedded sensors in the road surface – but the problems are ethical: viz. the challenge of programming ADAS to make ethical decisions in unavoidable accident scenarios, with no easy answers to questions of whom to prioritize.
Even if fully automated cars appear I won’t be able to afford one so it’ll be Shanks’s for me, then confined to barracks.
@Theophrastus
There are already thousands of driverless taxis in service, likely to be millions by 2030, and they don’t need any embedded sensors in the roads.
The “ethical” dilemma is easily addressed. Given the choice of killing a pregnant nun pushing a pram over a zebra crossing or erasing a thug swearing at children, the autonomous programming does not attempt to weight their social values. It just avoids hitting both of them by not getting into a situation where hitting them is an outcome.
That bias to safety does not mean they are immune from accidents as their hesitancy does result in almost as many minor accidents as humans. Mainly due to human drivers not paying attention and driving into the back of them.
Ducky,
“The other issue, part with regard to nVidia, is that Google et al, let alone the Chinese, don’t really want to be paying cash away to them.”
For China it isn’t just about the money, but that they don’t want to depend on the Yanks who can ban exports. The money being thrown at RISC-V development isn’t just about saving money, it’s making sure they can get chips.
If Spud is prescient of financial cycles how come he’s living in an end of terrace in Ely and not the in a grand mansion somewhere? Making money would be so easy all he had to do was buy at the bottom and sell at the top.
Any fool can say: “This market is over-valued and a correction is inevitable”. The wise man asks: “When?”
As the old (but accurate) saw has it – the market can remain irrational longer than you can remain solvent.
“If you want an indication of stock market volatility, that is it.“
Complete and utter bollocks.
If you want a measure of volatility look at the VIX, it’s really healthy. Then go look at the futures, it’s really healthy until December when it goes flat, which may indicate a flat period for a while, a slow grind slightly up or a slow grind slightly down. Yes you have to look every couple of days but it’s being saying this story since it went back in to contango after the tariff correction. Short term there are no issues.
Jesus, Phil, really?
Take a look at what the VIX did back at “Liberation Day” or whatever the fuck it was called.
Then tell me how the bloody thing is built.
Yes, that was rather a snigger. VIX is what people currently think about the future. Obviously that balance of opinion doesnt include a crash because if it did the crash would happen now.
I’m referring to the Vix and the Vix futures. Two different things.
To clarify:
Vix is currently 14.4 = current volatility is low = immediately, Murphy’s statement is complete bollocks.
Vix futures are WELL in contango, near month is 14.2%, which is high which means markets go UP.
October 5%, November 1.6%, December 4.5 January 1.something. And on
The futures change daily and December on currently has improved from where it was.
It’s always the near month you want to look at and it’s very healthy. If it falls into backward action it gives a good 2-3 week lead on a correction as it did before the tariff correction and as it always does.
backward action = auto correct of backwardation
WB; yeah. I guess that you could say that the CCCP was able to spend a fair amount of time, feet up on the sofa, watching, taking notes.
About 25 years? Nixon?
Looks like they weren’t entirely idle, given the national security externality, but RISC-V development did languish. May be they reckon they properly understand the NSE by now? God only knows.
Open source, interesting they chose it, rather than get all nationalistic and build from ground zero.
Is it Malaysia or Indonesia that whacks stuff out at 25nm or higher?
Why does he think this is unrelated to real world activity?
The US stock market is up in general, GDP report says 3.3% increase, 2% inflation. This is just some people saying ‘hey, the US economy isn’t such a shithole anymore, we should try to get in before it gets expensive’.
I keep saying the US stock market, and quite possibly its economy, will crash . . .
*Foreign investors put $163 billion into US stocks in June* – headline.
=============================
Nothing messes up an economy like foreign investors.
Gamecock thinks AI valuation is irrational exuberance. A bubble. He doesn’t understand how it can be monetized. What is the business model? ChatGPT gives you information for free. How do you make money on “free?”
AndyF,
“There are already thousands of driverless taxis in service, likely to be millions by 2030, and they don’t need any embedded sensors in the roads.”
There are zero driverless taxis in service. There are cars that are automated and being monitored by a team in a nearby office telling the cars that it’s a green light or a red light, or to stop the car before it hits a child. That’s how it “works” where we know that the driverless Cannonball Run record is only 98.5% of the journey.
The whole thing reeks of an investor scam to me, a Mechanical Turk. We don’t know if this actually adds up until any of these companies are profitably operating taxis that are cheaper than human ones.
We also know that this is going to work better in a well-connected city, at low speeds, and with really boring roads, like LA. But even then a chunk of LA. You want to take a Waymo to Pasadena or Malibu beach, you can’t.
The other answer to taxis, is greater shared riding. And there’s a lot of work going into this. Uber have run some experiments on it. If there’s 2 people in Market Snodsbury that want to go to Droitwich at about the same time, the driver picks both up. You get something like 1/3rd off your fare for the small inconvenience.
Gamecock,
“Gamecock thinks AI valuation is irrational exuberance. A bubble. He doesn’t understand how it can be monetized. What is the business model? ChatGPT gives you information for free. How do you make money on “free?””
You can pay for a subscription to ChatGPT. Each level comes with usage limits. But the thing I heard is that they’re losing money on these. This stuff is expensive to process.
My guess is that OpenAI will go IPO in 6 months at some ridiculous number like $500bn. The hype has about peaked and so there’s not much point in burning more money.
Western Bloke/Gamecock;
Yeah, pay the subscription fee. That’s exactly the same model as Office or Adobe or whatever.
But, there are two types of search; type “bluetooth speaker” into Google, you’ll get all the sponsored shopping links plus all the hits from the likes of Currys and so on. Those sponsored links are obviously paid for.
Try “Jevons paradox”, and all those links vanish, not surprisingly. It’s difficult for Google (other search engines may be available) to monetise that type of informational search via advertising.
What you get instead is the Gemini summary on the topic. Which isn’t delivered on the product/service search. Gemini also shows a list of sites that it has (apparently) referenced in building the response.
So, if you are the author of a piece on Jevons Paradox (or whatever), and you want your articles to be shown as being used by Gemini, then you’d pay Google to have Gemini do exactly that. Ker-ching!
That’s what I’d do.
Of course, there’s the minor technical hitch that an author with intentions could pay to poison the well, but hey.
Subscription fees. Thanks.
AI is predicted to increase electricity usage in the US by 60%. That’s a shitload of subscriptions.