A market hooked on a handful of mega-cap stocks is a precarious place to be.
By Wolf Richter for WOLF STREET.
The Russell 2000, which tracks the smallest 2000 stocks in the Russel 3000, fell 1.6% on Friday, was in the red year-to-date, was back where it had first been in January 2021, and was down 18% from its all-time high in November 2021 (data via YCharts).
But the Nasdaq Composite eked out the fifth record close in a row on Friday, and was up 9.1% from its previous-cycle high in November 2021 – having surged by 73% from the big-trough low in December 2022. Just another day in a hyper-precarious stock market where the entire world has whipped a handful of huge multi-trillion-dollar stocks into frenzy (all data here via YCharts).
The S&P 500 Index is now governed by a junta of three stocks with a combined market cap of nearly $10 trillion (Nvidia, Apple, and Microsoft) and another three stocks with a combined market cap of $5.4 trillion (Amazon, Alphabet, and Meta). Those six stocks combined have a market cap of $15.2 trillion.
All of the 503 or so stocks in the S&P 500 combined have a market cap of about $45 trillion. Without those six stocks, the S&P 497 would have a market cap of $30 trillion. This entire market lives and dies by these six stocks. A fund that reflects the S&P 500 has a third of its holdings in just six stocks. Forget diversification?
So the S&P 500 was just a hair in the red on Friday, and a hair off its all-time high, up by 13.3% from its previous-cycle high on January 3, 2022, with a big trough in between.
The S&P 500 Equal Weight index is not weighted by market capitalization. It contains the same stocks as the S&P 500 index, but all stocks weigh the same within the index. The purpose is to see if the performance of a small number of outliers with huge market capitalizations is driving the overall S&P 500 index and is in effect hiding what is happening to the rest of the market. And that’s the case.
The index fell 0.7% on Friday, is down 3.8% from its all-time high on March 28, and is right back where it had first been on January 4, 2022.
Year-to-date: For the year 2024 so far, the Russell 2000 (red) is down 1%; the S&P 500 Equal Weight (maroon) is up 3.4%, while the S&P 500 (purple) is up 13.9% and the Nasdaq Composite (light blue) is up 17.8%.
So that’s a very funny-looking split in the stock market:
The Mag 7 is now misnomer. Tesla fell out of the group. Its market cap plunged by 56% from its all-time high to $557 billion, less than half the market cap of Meta, which is the next smallest in line with a market cap of $1.28 trillion. But to revel in nostalgia, we’ll keep Tesla in the chart below (green).
What happened to Tesla – a 56% drop from the high, and a 75% drop at one point – can happen to any of the Mag 7.
In 2023 and year-to-date in 2024, Tesla is up 44%, still a massive increase in normal times, but that’s the smallest increase of the Mag 7.
The Mag 7, in 2023 and year-to-date in 2024:
- Nvidia [NVDA]: +802%
- Meta [META]: +319%
- Amazon [AMZN]: +119%
- Alphabet [GOOG]: +101%
- Microsoft [MSFT]: +84.6%
- Apple [AAPL]: +64%
- Tesla [TSLA]: +44%.
Five of these Mag 7 are the biggest stocks in the universe in terms of market value. The entire world is now invested in them. They massively move the needle, and they move the needle even when the rest of the market is pulling against them. Having the big indices – and portfolios – hooked on just a handful of stocks is a very precarious place to be. But that’s what it is.
Enjoy reading WOLF STREET and want to support it? You can donate. I appreciate it immensely. Click on the beer and iced-tea mug to find out how:
Would you like to be notified via email when WOLF STREET publishes a new article? Sign up here.
I spoke with someone in the know who told me that most NVDA chip sales are to Microsoft, Amazon, Apple, and Google. Around 10-15% are to everyone else.
This reminds me of the scenario in 2000 when everyone was buying Cisco equipment because of the year 2000 changeover. People were buying because they figured they might as well upgrade instead of patch for year 2K and then buy later. So in early 2000, orders just went off a cliff, because everyone who wanted to buy, had already bought, and no one would be buying big orders for 2 to 3 years.
At some point Microsoft, etc. will have all the AI chips they need, and there is no way to remotely sustain the stock price buy selling to the remaining 10-15% of the market.
What bothers me so much about this market is the ability for people who are connected to the orders and can tell how much and when (insider trading, whether labelled that or not) to make a lot of profit with little risk, and everyone else is at the mercy of the big boys.
Am I looking at the situation correctly?
What do people think?
Finding out where you stand is not very comforting.stress builds.
Us humans are a strange lot, in my next life I’d like to be a potato bug, when I get stressed out I could roll into a a little ball and chill out.
By international standards, the US concentration of 1% of stocks making up 33% of the main indexes is not especially high.
In terms of intelligence, we are basically at par with the potato bug the Lady bug and the stink bug. AI will be so advanced we will be but the human bug, plodding along is search of reason and meaning.
But in the mean time, comb your hair, put on your shoes and lay down your bets.
Chips may be overtaken by more powerful ones, but they depreciate very slowly. This is being typed on a 2013 Acer comp. Obsolete? Does the job.
Investment idea: buy all old comps, remove boards and sell to Russia. In the days of the USSR they are said to have bought 500 Chatty Cathies just for the chip. Now everything has chips.
Remember the glut of memory chips and all the accusations of dumping, i.e. selling below cost of production. Moving to the present, the limit of waver thinness is close if not at hand. It is going to get much harder to make last year’s chip obsolete. The mass producers of chips could run into a new glut.
Agree that at some point, chip demand will plateau, and it may be abrupt.
Selling below cost is not dumping but rather selling in a foreign market for less than domestic market. This happened and was proven with South Korean memory chips in the 80s. You can sell at a loss and perfectly legal just not always a solid business plan.
…chase that market share first, adjust profit margin accordingly as it grows (until it doesn’t…).
may we all find a better day.
AI requires a lot of chip horsepower and it has barely begun. NVDA isn’t Cisco in 1999. It’s IBM in 1979. There will be an exponential increase in AI chips over the coming decades. NVDA may or may not be the one to own the market. Right now, it is head and shoulders above everyone else in technology. But if it gets used to high margins and the subsequent high costs that brings, it may end up the IBM of 1999.
Google makes it’s own AI chips (“tensor”) that are specialized hardware and has for years. All AI work will eventually go this way for efficiency. If NVDA is smart, they’ll start designing those for sale.
Reminds me of THGTTG, fire up an enormous use of datapower using vast resources, just to get bs answers.
There is no AI, it’s just a bigger script than previously and an enormous waste of resources for most takes.
That said, this BS could go on for years still.
The first mover will probably be burned if there is actually any money to be made in AI. The next generation chips will be twice as fast and use less power and the next guy’s costs will be much less. That’s the problem with buying chips when you can’t immediately make money on it.
I don’t know about insider timing but…
1) I do think that the market-weighting of the SP 500 (kinda a bit weird if you think about it) tends to greatly amplify the overvaluation/over-concentration of a small number of players in an index whose entire purpose is supposed to be diversification.
2) Auto-deposit 401k boom/fleeting demographic sweet spot then play into #1 to add fuel to the overvaluation/over-concentration fire.
3) Massive, self-serving but dangerously contingent stock buybacks then backstop the ramshackle structure created by #1 and #2.
It is pretty unhealthy but utterly toxic when coupled with #4…
4) DC interprets any significant fall in major stock indices as a Keynesian “suboptimal” mandating a four alarm fire of interest rate strangulation/money printing (which DC is horny for at all times anyway, in the service of its own interests).
So you get 20 years of conservative, reliable investment slain on the altar of truly goofy equity overvaluation (the once and future ZIRP).
I think your friend is correct, but I don’t see it as a concern. Disclaimer: I am long NVDA.
The big four are piling resources into “AI” because they see it as a threat to their locked-in customer bases. Chat GPT was a very rude awakening.
NVDA is in a stellar position because they own their own chip fabs. You can’t just jump into making chips. It takes years, billions in investment, and a raft of Ph.D. level expertise. You have to have a Ph.D. in something like solid state physics just to be the overnight guy on call if something breaks at the Intel fab. Chip manufacturing equipment breaks all the time, in byzantine ways.
NVDAs competition won’t come from some upstart. Such a company would have to get fab time from TSMC, who will not hold up Apple or Samsung runs for any reason. See: carmakers sucking hind tit two years ago.
If there is to be competition in the space it will have to come from the big processor companies: Intel and AMD. They have the fabs and technology in place.
Disclaimer: I am also long AMD.
I’m not familiar that Nvidia or AMD have their own fabs. Do they? I don’t think so.
This is wrong in so many subtle ways that it sounds like an AI response.
Nvidia does *NOT* own their own fabs. TSMC produces Nvidia GPUs. Ironic source: ChatGPT.
drg1234,
“NVDA is in a stellar position because they own their own chip fabs.”
No, they don’t “own their own chip fabs.” Nvidia designs the chips, and others manufacture them, primarily TSMC (Taiwan Semiconductor Manufacturing Company).
I hope that this misconception isn’t the reason you’re long NVDA. I hope you have valid reasons to be long Nvidia.
I know how it all evolved, especially the technical part…was part of it…and of course “economy of scale” notion driving small biz out…..I watched that happen……..thinking kinda fades away there, as I never joined the stock market crowd….get a bit more on small CDs now, though, I doubt if that or fixed COLA income beats inflation.
Downsizing is my “job” and avocation…….finding all the BS I DON’T NEED and trying to avoid what required/desired* shit I can.
*like TV/internet/own transport
And “stupid” flip phone…..which cost $60 extra to buy….most all exotic “smartass” phones were FREE.
That’s actually publically available information. You don’t have to be “in the know” to find out about that. Many news sources have released that information including Bloomberg, CNBC, etc. You forgot Meta though. Mark is spending lots of his Zuckbucks on their GPU’s.
“ Am I looking at the situation correctly?”
No, you should not be bothered but be happy you found a way that can give you financial advantage
You know and understand what is distorted and from history you also know what will happen next
Nvda earlier was riding high on the crypto BS and is now riding high on the AI craze. Good for them but they need to find another trick soon before purchase orders get less urgent (because customers have bought their stuff and buying more only gives diminishing returns). They have very high margins but competition is catching up. It always does
Google,Microsoft are riding high on expectations as well but sooner or later the calculation of the hourly cost of using their cloud stuff will swing back to customers having more in house. Apple, their problem is those iPhones last too long and improvements are only marginal
And sooner or later the bidding will be overshadowed by profit taking and more capital exiting because the upside is smaller than the downside risk
Don’t take your eye off the ball which is the future of AI. Specifically, will the dweebs give us permission too predict what they are likely to do based on a recorded history of what they always do.
AI is a Trojan horse,
They’ll never “have all the AI chips they need”. Chips will continue to improve in performance and new silicon will replace old silicon.
The biggest threat to NVIDIA is the fact that they don’t control manufacturing, they design chips and have them produced.
All of their biggest customers have been working on designing their own AI silicon for multiple years already and will eventually shift a very large percentage of their current NVIDIA purchases into production of their own designs so that they control the entire hardware/software stack and save money in the process.
NVIDIA got really lucky windfall when ChatGPT threw everyone into a panic frenzy to add Generative-AI into everything prior to the major cloud providers completing their own in-house hardware designs.
I suspect NVIDIA’s revenue derived from AI hardware sales will drop 80%+ once Microsoft/Google/Apple/Meta make the shift to having their own designs manufactured for them directly. Given the extreme dollars currently being spent and the fact that they were already years along into the design process, I don’t expect it will take very long for this to happen.
No wonder CA home prices hold up in many of the areas except San Francisco . The large tech presence and the PE funding available for many of the AI customers is something that reminds me of 1999 and the internet. I hope this one is more stable. AI has the potential to bring about some terrific changes in this world and I for one love the potential for the technology. On a side note something that Wolf tracks is the forecasted drop in future oil demand. Primarily as a result of the rise in renewables and a shift to EV VS CE with different choices for Electric generator fuel requirements than CE required refined products. The drop in fuel this summer seems to have had a larger than expected drop in CPI though still high . This drove the Mag 7 even higher !
“AI has the potential to bring about some terrific changes in this world and I for one love the potential for the technology.”
The lack of operational product specifics and granular business use cases are always a big red warning light of a tech hype cycle.
…and AI has been oversold multiple times before…and had multiple, monumental recent absurd f*ck-ups, and Nvidia (always reminds me of invidious) is a supposed engineering enterprise run by a popped-collar, leather jacketed hype-meister who has recently taken to signing fan-girls t*ts.
With that level of steady, sober, serious application of effort, how could AI fail?
CAS127 – You sound jealous. I missed the run up also and am afraid to get in now because I don’t want to be the last one at the punch bowl. There will be more stocks like this in the future. There will always be the next big thing.
I have been running several AI models predicting the short term market valuation. The model tells me that I am nuts and the R2, the correlation with my prediction and the actual result has been declining in a manner familiar too any market bear individual, who expect a market failure.
“The drop in fuel this summer seems to have had…..”
It’s still Spring……
Astronomical summer in the northern hemisphere begins on the summer solstice (6/20) but meteorological summer started on June 1.
Summer officially grasps the wand on June 21, the very day of the equinox.
…unless one is south of the equator, where it is Autumn’s advent. The spacecraft continues to maintain its slightly-wobbly axial tilt while it revolves and orbits it’s vital star, regardless…
may we all find a better day.
…a thousand pardons, here, should read: “… WINTER’s advent…”.
may we all find a better day.
WOLF – I can’t find it, but has there been a time in history when the S&P has been this dominated by a few stocks?
Actually in most other countries markets the concentration is even heavier.
Mentioned multiple times as though it is a defense.
Reminds me of excuses for US’ macroeconomic pathologies (“We’re *almost* the least crappy!! Pay no attention to the other economies behind the curtain actually kicking our ass!! Stay sexy and stupid, America!!”)
Cas – mebbe a not-uncommon human tendency to eschew the love/need/knowhow of everpresent competition for a sense of entitlement after a period of success. History’s road is littered with the wrecks of societies that lost the plot of how they got to the dance (…or eventually got tired…)…
may we all find a better day.
Other countries have much narrower markets. The US has an extremely broad array of large public companies. But difference.
Other than cutting jobs, saving profits for companies and selling AI chips. Has there been a monetization of AI to justify these mind boggling stock prices. NVDa , can it keep a 600 percent growth rate when its big four customers are done gorging Toonces wants to know …
Microsoft monetizes ChatGPT through its corporate sales programs (Microsoft 365). Access was $20/user/month last I looked. MS knows a lot of mega corps are going to want their own so they will license the data and they will have their own NVDA plants. And that’s just for LLMs. Robotics (a bigger deal in my small mind) doesn’t even use the LLM technology. They use reinforcement learning and will need billions of chips worldwide, with constant replacement. I’ve gotten deeply into this stuff over the last year at the end of my career and start of my retirement. I’m doing my own little AI project for fun and maybe profit someday. It’s been an “O Sh!t” moment for me. The stuff works, still has lots of room for improvement, and has a million applications. Stocks are high. But they are for a very valid reason IMHO.
“so they will license the data”
*What* data for *what*?
It is amazing how all the “explainers” leave all the actual specifics (product operations revenue generating use cases/granular economics/etc) out.
Until the automagical AIs can be counted on not to tell me that George Washington was a Navaho lesbian werewolf…I think I’ll hold off on buying the hype or letting an AI do my accounting reconciliations.
I doubt it. The impacts of AI, whatever they may be, won’t come to fruition this economic cycle. And if Big Tech doesn’t demonstrate significant profits in the next year or so—which I doubt—attention will generally shift to the next bright ‘n shiny technology.
I’m not discounting AI altogether, just saying it’s way overhyped at the moment. Let’s see where we are at the beginning of the next cycle.
“attention will generally shift to the next bright ‘n shiny technology.”
I’m hearing great things about Dyson spheres.
Also, glitter-infused carrots that convey super-powers to humans.
CNBC will be covering both next week…
Not enough is the anthem of the protestors that have everything to lose and nothing too gain.
Funny how everyone stopped talking about Tesla as soon as its stock stopped going up
WTF are you talking about? Tesla gets more press and attention than almost any other company. Totally disproportional. Jeez.
In the article, where Wolf writes “What happened to Tesla – a 56% drop from the high…” he’s specifically referring to Tesla’s stock, not the company as a whole; that it’s about the stock is implied.
Same with my comment.
I am probably a country polk outlier, but I find the google AI additions to be an annoyance amd I am not sure why it is even appearing on facebook. My guess is that MSFT will find a way to profit by integrating AI for business needs while the rest just blow money with their own Chat GPT clones.
Agreed, I was just pointing out in a recent converasation with someone that most of the time when I search an answer on gogle the AI response is wrong. If not wrong completely then it is at least lacking critical context….but mostly just outright wrong. Yet another way for gogle to mislead the masses willingly or unwillingly. I almost cant even believe that a company so worried about search reaponses are allowing it to show up.
I also got my first “beta” AI response from Google to a technical question about Google Analytics, and it was BS, just gobbledygook. You’d think Google AI would know the correct answer about Google Analytics. But no.
Just to further this: Amazon has rolled out an “AI” function that replaces the search box that you used to be able to use to search through the reviews. It is beyond useless.
We need a name for the process of swallowing/hyping gobbledygook…
Gobbledygeekery?
Gobbledygriftery?
Stock gobblers?
The emperor is naked.
Googelygook?
(I couldn’t resist)
…don’t go dumpster-diving and expect to emerge unsoiled…
may we all find a better day.
The real danger is the input to the engines. If they are a product of available information then one would expect their to be the same biases as mainstream media, and the darker side is they can be programmed to just be that much more effective at propaganda. Thought CNN and MSNBC were extremely political and biased sources? This could be next level.
CNN and MSNBC may be biased but I can think of a few other “news” organizations that are off the charts in bias. They are actually run like components of campaign organizations. I wonder why you didn’t mention them?
Escierto,
There are about 3000 official media outlets in US alone so seemed too verbose to mention them. That isn’t even where the most manipulation is done as you have to look to big brother for that. A good example is the Pentagon who ran anti vax anti China campaign that was reported in last few days. How can an AI sort through biases and outright lies? Very tricky algorithm and of course they can be easily manipulated.
These “AI” products are only as good as the data they’re trained on. I consider it to essentially be a faster Google search. Chat gpt is often wrong but if I ask a few questions to it back and forth I generally get what I need, faster than if I’d just googled it. It also great for annoying redundant tasks like writing a cover letter. I copy and paste my resume and a job description and tell it to write me an xx length cover letter and mention xx for this job based on my experience. Typically I change a word or two because I wouldn’t call it completely natural language, but I get a custom cover letter in 3 min instead of about 30 min
Your AI-written cover-letter story was the core of an article in the WSJ some time ago about how HR people at companies and recruiters no longer read anything because they’re flooded with AI-generated resumes and cover letters because people send out a gazillion of them now that ChatGPT is available to the public. So this AI-generated stuffs gets fed into their AI system, from AI to AI. Obviously, that has been going on for years to some extent, but the public availability has driven this to an entirely new level.
MB – My son was actually just granted a scholarship as he was told his essay was an honest response and obviously not written by ChatGPT. The next generation is losing the ability to write and to think objectively as they are starting to rely heavily on this new technology.
In the future, there will be a bull market on all things authentically human. I.e. customer support where you can talk to a real person instead of AI (and you won’t care if your real person is in a different country at that point).
…and GOOD ‘customer service’ finally (and unbelievably) becomes something considered worth paying for by the buying public? (…and cleverly separated from a proper expectation of a vendor to provide same)…
may we all find a better day.
Indeed it’s already happening. Many customers tell me they’d rather buy from me despite higher shipping costs etc. because they don’t want to deal with Amazon.
MM – glad to hear that, hopefully the trend won’t swing back to a situation with heretofore relentlessly cheaper pricing from other sources becoming irresistible to your customers, good service or no (high general customer knowledge of the niche, especially when recreation-oriented, tending to favor online self-service)…best.
may we all find a better day.
My fingers are partly burnt to charcoal learning that you can only invest in the USA and presently only in IT and similar, which is where highest growth lies. FTSE is in death throes, socialist EU hates wealth, I don’t eat rice with chopsticks or particularly like curry, so no China or India, although Hong Kong gave me money until Xi crushed Jack Ma. Chips are the modern creation of the wheel and their development and application have a long way to go. Be There or Beware still holds, the commercial and private computers holding many trillions of dollars in cash are just waiting for sweaty little fingers to push their Buy Equities buttons, mine as well, at the slightest hint of a correction. No doubt that everything is overpriced, but Perfection of the Market also holds true, it supposedly knows what is happening 6 months in advance. Stretching out now, but interest rates are more likely to drop than rise, they have had their days for a while.
My exact thoughts.
Just happens that the best managed and productive companies are the biggest ones. And technology, one way or another, is the future for a looong time. Corrections will happen of course.
But more tricky question is if small caps will catch up ever.
Biker
Have you identified your own contradiction?
How can small caps ever catch up if the best managed and productive companies are the biggest ones?
Ok. Then they will not 😉
Yup, everyone wanted a website in the late 90s or they’d be left in the dust. Thus the dot com bubble.
Penisland.com springs to mind.
We’re 25 years further along, and bricks and mortar still declines in the face of internet shopping… the dot com era signalled a revolution, not a generational evolution that spans the life of most people’s retirement portfolios,
Only five years ago we had AI revolution 1.9 with the whole deepfake thing.
Chatgpt got it a bit further by doing the same stuff as pictures but with words.
In the interim meta wanted you to buy a premium store-front real-estate on their main-street on their meta-verse.
Others wanted you to buy NFTs or blockchain to access all your stuff and buy stuff.
Nvidia wanted everyone to work on their projects using VR in the omniverse.
Let’s not forget Facebook was banging on the VR drum a decade ago and it’s still not become “iPhone 2.0”
Even Apple haven’t pulled that off.
Tech advancements are never-ending and supposedly always watersheds, and supposedly need front-running or you’ll miss the boat.
But time and again they’re not. They take decades to mature and during that maturing process quite often the leaders are the losers by the end.
But this time it’s different and we won’t see the massive flip flop in valuations?
The best managed and productive companies are the biggest ones? What a joke. I once made a recruiting trip to Wal-Mart in Bentonville. They paid for my plane tickets, a rental car and a nice hotel for a couple of days. When I got to their headquarters at the appointed time, no one there had any idea who I was or why I was there. My experience was typical.
Obviously I meant the big 3 or even the big 6.
Agree E: Had similar experiences a couple times, and also had the exact opposite when arriving for appt. with COO at smaller company.
COO had an unexpected call away, but receptionist expected me, and co owner met me at appt time — and hired me on the spot ( after interview ) with COO to negotiate salary, etc., when available.
Loved that company with ”hands on” owners, similar to other smaller companies worked for over the years.
In silicon valley if you are not a part of the clan you don’t get in. If you are a part of the clan, you get the job even before you have the token interview.
“but interest rates are more likely to drop than rise, they have had their days for a while.”
Nope. This has already been covered on this site repeatedly. You might want to actually read what the FOMC says rather than listening to journalists.
The last “dot plot” (6/12/2024) indicates that “interest rates are more likely to drop than rise,” according to the FOMC members.
Fig. 2 from https://www.federalreserve.gov/monetarypolicy/files/fomcprojtabl20240612.pdf
Interest rates could also rise or stay the same. This is just a forecast. But that forecast does indicate that FOMC members believe interest rates will likely drop.
“The last “dot plot” (6/12/2024) indicates that “interest rates are more likely to drop than rise,” according to the FOMC members.”
You’re ignoring the first derivative aka rate of change.
Notice how with each successive dot plot, the ‘longer run’ rate keeps getting higher.
People need to stop wishing for the rates to be cut. We are past the point where that would be beneficial and are now at the stage where the first rate cut is typically the canary in the coal mine and the signal that the next recession has already started. Enjoy the 3% inflation while it lasts and build some equity.
This stock market is blind to what will happen six months in advance, although you might possibly obtain an outlook with greater accuracy when you look below the surface as Wolf has done with his S&P497. Superb really.
@AB
I don’t need Wolf to know the details like in this article. I guess you lack your own research. I recommend for more balance and understanding.
You are 100% correct. You have taken a guess.
Sorry AB. After re-reading your comment I misplaced my response. It was more appropriate for others comments.
Looks like many commenters are only reading Wolf, but definitely good side if anyone relies only on one source.
Maybe I’ve read too much Science Fiction, but it kinda looks to me like one of the Mag7 will win the AI Arms Race and go on to control the world. I include Tesla, because Musk is heavy into AI.
Very enlightening observations as usual.
What problem in your life, or mine is AI going to solve?
I can see an app for a big retailer like Wally. The price/sales of thousands of items could be tuned hourly to max profit.
But it’s not like, say, electricity, that chills yr food, heats or cools yr house, and literally cooks yr breakfast. But it is being hailed as such.
Is it relevant to note that 2 of the recent Boeing crashes weren’t caused by humans, they were caused by auto systems interfering? The day before the first crash, when the auto system started forcing the stick down, a pilot in the back seat knew what it was and told the guys up front how to turn it off. A few days later. there was no one in the back seat.
@DK
I guess similar comment like 25y ago some saying about internet, can’t mow the lawn. 😉
The South Park special Panderverse was very much about this where handymen were the billionaires and accountants and programmers were hanging outside of Home Depot looking for work. Good angle on entertainment industry but Disney is an easy target!
Yup. The internet STILL can’t mow the grass or fix a toilet or clear trees that fell on the power lines that cut off the electric to its servers.
The thing about AI is how vulnerable it is to having its plug pulled.
Funny how truly poor the security is at all those power gobbling data centers.
Nick – your first paragraph is prescient, modern speed of communication seems to becoming integrated as even more of a component of the modern general retail marketplace. The frenzy found in the commodities pits may look like a placid Sunday picnic, at least for awhile…
may we all find a better day.
Nick Kelly,
“What problem in your life, or mine is AI going to solve?”
AI has already started to reduce costs by replacing people with AI. This started happening years ago. It’s like automation in factories, but at a higher level. People who write reports, meeting summaries, analyses, news stories, fake news stories, basic software code, etc. are being replaced. This happened years ago. There are many other areas where similar things are happening — taxi drivers, for example. Waymo’s robotaxis are doing just fine in San Francisco and other cities, a lot better than human drivers. And they’ll replace human drivers and cut the costs for the operator over the long run.
So the problem AI has been “solving” for now is costs for the operator — rather than generating new incremental revenues (except in companies that sell the hardware, software, and services for AI).
Sure, it might create new problems (such as shitty work), but humans are known to do that too.
“AI has already started to reduce costs by replacing people with AI. This started happening years ago.”
After leaving legal practice 20 years ago, I was able to make a very good living as a freelance legal translator. Since around 2020, AI translations have displaced many a translator, even those with specialized areas of expertise who, if they are still in business, now spend most of their time editing.
‘People who write reports, meeting summaries, analyses, news stories…’
I know that kind of writing, although it is mostly still from humans. A lot of the boilerplate re: interest rates could be regurgitated by AI or anyone with marginal literacy. It reminds me of student essays where the object is to somehow get to required word count, so it just repeats stock phrases.
Insightful writing, not as easy. Then there is talk of the coming, shudder, AI humor or novel.
Where I think a lot of legit money will be made with AI is animation that looks real. Maybe the movie career of Elvis is not over. A great danger of this tech will be producing video where people appear to be saying stuff they didn’t really say.
AI flunked two tests for me already. In one of them Wells Fargo ATM machines started putting political crap on their screens when you tried to use them for routine transactions. I made a complaint to the office manager to take that crap off and stick to processing banking transactions. She promised a prompt response. I got a letter in the mail a few days which was processed using AI. The letter didn’t even answer my concern about the ATM machine and had a boiler plate response saying they would make the human tellers more responsive to customer. AI’s grade was ‘F’ in handling this simple request.
In the second example of how AI failed me, several months ago a car was parking next to a fire hydrant in the street right next to my house. I was going to call the authorities and have the car towed but wanted to check to see if it was legally parked or not. A dude at the golf course volunteered to use his AI system to find out the answer. He used AI and it came back with the wrong answer. AI said the car was parked illegally if it was less than 10 feet of the hydrant which it was. It was about 4 feet from the hydrant. I double checked with the local police who told me that it was legal if it was 3 feet or more. AI got it wrong again.
garbage in; garbage out; it makes no difference if the so called intelligence is human or artificial. The only folks interested in AI are the business owner class, who yesterday sang your workingman praises and will fire you tomorrow if they can get a bot to replace you. Hell, they already tried it with india. Just hire those software developers. Disney anyone? AI is a fad. Today.
“What problem in your life, or mine is AI going to solve?”
As much as I resist AI on the customer service chat I’m in charge of (boss keeps bringing it up), the truth is >50% of questions I get could be answered with a simple google search.
There’s a reason letmegooglethat dot com exists.
What you get for free with AI currently via ChatGPT or Gemini is nothing compared to what enterprises are getting. Adobe has some sick products. Their Firefly product is amazing.
Why is Firefly in any way ‘amazing?’
Wolf has been using AI for a least a year now. He has trained it to filter comments and respond appropriately with RTGDFA when necessary and no one has even noticed. :)
WolfChatAI®
Re: forcing ‘down’ should be forcing control ‘forward’ to force plane’s nose down. Although it’s no longer a ‘stick’ the shudder as the system fought the pilot is called a ‘stick shaker’.
Great article per usual. Please allow me to address a few issues with my personal observations. Regarding AI, please note that I make no claims to be an expert, but I am a statistician and statistical programmer who has a pretty good understanding of gradient descent and other AI technicalities. And the opinion I am stating here is shared by many AI experts in the field. Take a look at Gary Marcus for one.
1) AI is essentially three things: A) Faster processing speeds. B) Large data sets obtained by scraping the web and all the garbage that goes along with it. C) Statistics. Not any new statistics, mind you. Just the same old stuff dressed in fancy clothes. I remember Marvin Minsky talking about how amazing AI was going to be when I was a kid, and I am 70 years old. This AI stock bubble is going to burst, just like the Internet bubble of the late nineties. Meanwhile, enjoy your Nvidia stock. But please remember that the chip industry has always been a “boom and bust” industry. That means you should invest during the bust and sell during the boom. In other words, buy low and sell high, the exact opposite of what most investors do because they get caught up in the latest fad which is now AI mania taking over the world.
2) An astute reader at Wolf Street made an interesting comment about the math of market caps. Since I am a mathematician, it caught my eye. For any given major stock like these tech giants, they have many shares outstanding. Yet only a sliver typically trade on any given day. But that sliver of trading, among those who want to sell and buy, affects the entire market cap. I live in a small retirement complex of nice townhomes in Massachusetts. If one of us sells our home, then everyone’s home valuation changes, right? But why should my home affect everyone else’s? What would happen if everybody went to sell at the same time? The market value would probably collapse. The mathematical point being that market cap is based on a tiny sliver of trading, and can change very dramatically and very quickly. Beware. When sentiments shift, and a bigger sliver wants to sell, that market cap and share price can drop quickly.
3) By any sane metric, the valuation of these mega-cap stocks is scary. They are priced to perfection, so let’s take a look at them:
Apple – Growth prospects limited in the two largest countries in the world. China and India want to build their own smartphone market.
Tesla – Most profitable plant was in Shanghai. Meanwhile, China is quickly dominating the EV world with their own cars. Tesla keeps trying to reinvent itself for good reason, because they are facing a huge flood of EV competition.
Meta / Facebook – Will they keep growing their fan base? Doesn’t seem likely.
Nvidia – Already talked about. The chip industry is currently in a “boom” cycle. Beware.
The Rest – They all fall into the same AI mania which I already addressed. Let the buyer beware. This includes Amazon. They make the vast majority of their profits from AWS which is Amazon Web Services.
Well stated.
“Meta / Facebook – Will they keep growing their fan base?”
No.
N=1, but literally none of my spouse’s 7th grade class use Facebook. They all say it’s for old people, and use Tiktok instead.
Meta’s value is probably more in Instagram than Facebook going forward, the youngsters are still using that. But yes, the very young are using Tiktok.
Microsoft: Control of the market from the beginning. It was never about a good OS but pushing out competition.
Can someone explain their tactics?
P.S. Please don’t get me wrong. It’s not that these are bad companies. Many of them are excellent, cash-rich, and successful. It’s just that they are currently over-priced. You should buy them when they are under-priced which could happen any day now. Market timing is an impossible task. Ask the Fed if you don’t believe me. Their track record of predicting stuff isn’t that great.
“You should buy them when they are under-priced…”
The challenge with megacaps is you’re competing with every institutional investor and their brother.
The microcap space is great because there’s less competition, ergo easier to find deals.
My main argument in this AI charade, is that globally, supercomputing processing capabilities have been following Moore’s Law since the early 90s. Supercomputing problem solving by government and university think tanks have been commonplace for decades, yet here we are today, with AI as a novel game changer that will magically transform the human race.
The algorithmic functioning that became known as AI or machine intelligence has been around since 1960.
The advent of this current nuclear bomb explosion in AI Bubble Mania started with ChatGPT and then industry wide panic to find ways to hype Siri and Alexa functioning into a semiconductor arms race.
This race of malfeasance and misappropriation is a war between billionaires that is weaponizing capitalism and distorting if not destroying global economic stability.
The absurdity of Nividia market cap, along with Mag Five in such a short time frame, has exploded so fast that the concept of antitrust violations haven’t even been pondered, as their share values to new levels of dominance. This too big to fail status creates profound economic instability, as computer algorithms purchase greater amounts of shares that are increasingly becoming illiquid. That’s a threat to national security imho, but apparently this junta, this collusion by this AI cartel has blinded everyone from seeing that AI is simply a faster way to sort recipes and family photos.
The further absurdity in this circus, is the amount of companies racing to create efficient ways and processes to steal consumer data in efforts to steal market share from each other.
When this ridiculous tulip mania finally implodes, the amount of sellers trying to escape through the same door will be fascinating to watch!
Since this will probably be a deleted comment, I’ll add another part to my rant:
“The infinite monkey theorem states that a monkey hitting keys at random on a typewriter keyboard for an infinite amount of time will almost surely type any given text, including the complete works of William Shakespeare.”
If enough monkeys use AI, it’s likely there will be various versions of Shakespeare interpretations that will essentially be hallucinated compilations — but sadly these monkey based think tanks will never have the ability or capacity to extrapolate their “thinking” into a process of creative innovation, to be able to dream up new things, versus regurgitating a finite amount of scrambled data points.
Agree extra one:
”The algorithmic functioning that became known as AI or machine intelligence has been around since 1960.”
GIGO ruled then, does now, and will continue to do so.
Clearly the case with ”AI” so far, eh!
What does everyone think about the Buffet Indicator (US Public Stocks Market Cap divided by US GDP)? It is like Price to Sales ratio for the entire US stock market.
Quite high (not as high as some episodes but very close).
Also of concern is the price to sales ratio of the Mag 7 (or 6 or 8 or whatever).
I know someone is going to say, “but AI, MMT, the Fed, ZIRP, FOMO, this time is different…”
I have noticed that when Buffet comes out and says the market is just too damn high (Oct 1999, June 2021, and a few months ago) there seems to be about a 6 month lag before things start going off the rails.
I would like perspectives on how people are thinking about valuation and what they are actually seeing as far as cracks forming (or not).
Ok, my take 😀
The reason is that the biggest components of the 500 are technology companies. And these quality companies tend to have high P/E. And this is new and becoming the new normal….
I know that this sounds cheesy but you asked for the reason LOL
Also think about Buffet own diversification. From top of my head: 60% of ONE stock, the one biggest in the 500. Also why he missed the semis boat, among other misses. He is a good investor but not Oracle. No one is.
My brother-in-law who knows less than nothing about the stock market even said the stock market was overvalued last year. My brother’s first wife knew even less. She made me a lot of money in the dope stock era as I was all in shorting dope stocks knowing she had a big long position in one of them.
AI or no AI, the valuations have become beyond ridiculous.
It’s like people have forgotten how much a TRILLION is.
I would bet that most people if asked, would not even know that a trillion contains 12 zeros (1,000,000,000,000.)
Here are a couple of amusing facts–
If you stacked a Trillion dollars in THOUSAND DOLLAR BILLS, one on top of the other, it would reach an altitude of 65 MILES. (That is higher than where Jeff Bezos’ spaceship goes.)
If you could travel at the speed of light, 186,000 miles per second, it would take you 62 DAYS to travel one Trillion miles.
Tech stocks are in major bubble territory. It is irrefutable. Having followed the financial markets for more than 25 years, I won’t be foolish enough to say it won’t get even bigger before it collapses -especially since it’s an election year- but it WILL COLLAPSE.
That is a certainty.
Yep, for me the basics of a trillion being a thousand billions, so since a billion is a thousand millions, a trillion is a million millions. And the debt per citizen and per taxpayer puts it in pretty understandable terms I think.
///
@Wolf, is this money running into a corner?
I have not seen a more scary chart in your blog for a looong time.
///
How can the biggest 3 stocks alone have ‘values’ in excess of $10 trillion when the total assets of the Federal Reserve are only around $7 trillion?
No one in America gets arrested for rigging the stock market. It all started with greenmailing then progressed to ponzi status. Every underhanded trick in the book knowing all the crooks and shysters have total immunity. If the middle class goes broke all the bag holders are gone and the rich will have no one to sell their overpriced shares to.
The daily moves of the russell 2000 look like someones trying to use a defibrillator on a corpse, I’m guessing the market makers or big fund managers like blackrock and vanguard trading it back and forth like bitcoin whales. “Clear!” *zap* pump it 1.5 percent and hope retail thinks its a breakout. Then they see no ones buying and sell it back to the partner whale. They tried that with the recent fed announcement and noone took the bait.
MW: Big Tech — not the Fed — is driving the stock-market rally.
I saw that headline too this morning (MarketWatch) and thought what a stupid ignorant headline.
No, not big tech. Crazed speculators and investors around the globe are driving the stock market by all piling into Nvidia and other big tech until they run out of stuff.
Meanwhile, the Fed has been sitting in its corner quietly doing QT for two years so that these people will eventually run out of stuff.
Wolf,
What are your thoughts about the Buffet Indicator and how that relates to current market dynamics?
If you look at the indicator over about an 80 year period, you can actually see the impact of QE. The impact of QT is much more limited (so far).
Have you thought about linking the Buffet indicator to liquidity analysis?
‘People who write reports, meeting summaries, analyses, news stories…’
I know that kind of writing, although it is mostly still from humans. A lot of the boilerplate re: interest rates could be regurgitated by AI or anyone with marginal literacy. It reminds me of student essays where the object is to somehow get to required word count, so it just repeats stock phrases.
Insightful writing, not as easy. Then there is talk of the coming, shudder, AI humor or novel.
Where I think a lot of legit money will be made with AI is animation that looks real. Maybe the movie career of Elvis is not over. A great danger of this tech will be producing video where people appear to be saying stuff they didn’t really say.
This was a great article. Northman Trader just did a great podcast with Adam Taggart from Thoughtful Money breaking down the technical charts of the current market setup and it’s pretty bullish from here; it wouldn’t be surpring to see some wild all time high’s from these mega caps. That being said, he did post some precarious charts of everything under the hood; the amount of names under their 50 and 200 day will the market edges ever so higher.
“wouldn’t be surpring [sic] to see some wild all time high’s [sic] from these mega caps”
I think the current all time highs are already surprising.
indeed. there’s nothing anyone could say that would convince me that current mag7 valuations are even remotely reasonable, even assuming the rosiest projections regarding ai and whatever else.
Been following Northman Traders works for quite some time and he is a technical analyst TA fully detached from fundamentals and valuation.
Every TA would say the same thing: trendline is all bullish and momentum is your friend.
if enough people believe that, it creates a positive feedback loop, and becomes a reality.
for a while anyway.
IMHO, ai adoption will be hampered by ai ignorance, and as more cases are presented to demonstrate ai ineffectiveness, fewer people will be amazed by its proposed magical abilities to transform human evolution.
Within two years, ai will be viewed as an anomaly — Nostradamus
See: “ Although LLMs are designed to produce fluent and coherent text, they have no understanding of the underlying reality that they are describing. All they do is predict what the next word will be based on probability, not accuracy.”
MW: The AI ‘bubble’ has helped the U.S. stock market dominate the world. What happens if it bursts?
The ai clown machine didn’t work out at MickyD
“ Fast food chain McDonald’s is giving up on its AI-powered drive-thru system, following a string of hilarious yet infuriating failures that resulted in unlucky customers ending up with over $250 worth of chicken McNuggets and unwanted packs of butter.”
lol
Your blog is a breath of fresh air in the crowded online space. I appreciate the unique perspective you bring to every topic you cover. Keep up the fantastic work!