You can tell these guys know nothing about LLMs or how they’re provided. I love how they show OpenRouter’s graph of token usage as if it speaks for usage across the board. DeepSeek looks like the king because people who use Anthropic and OpenAI use them on either a direct basis or AWS Bedrock …
And the bar chart for token costs, really? As if that’s information? Their sources are the API docs ffs. If they had at least modeled something to estimate token costs that would be interesting, but showing the public prices and calling it research is dumb.
toomuchtodo · 2026-06-29 14:42:08 UTC
Apollo Group has $1T assets under management, I believe them over most folks on HN. Their argument that the Mag 7 is burning up all of their free cash flow is factually accurate, as the returns are not materializing for the investments being made (ie underperformance).
re-thc · 2026-06-29 14:48:20 UTC
> Apollo Group has $1T assets under management, I believe them over most folks on HN
If only that's a measure of if they know how to time the stock market. The real measure is how much they earn not if their sales or marketing department manages to get more customers.
davidpapermill · 2026-06-29 14:51:35 UTC
The Mag 7 spending around $700B on capex this year, expected around a trillion next year.
That's more than their combined FCF, and they're borrowing to bridge the gap.
Danox · 2026-06-29 15:24:59 UTC
One is not spending anything out of the ordinary up until this point in comparison to their competition. But the Apple Silicon design group might start spending some money on something practical, something that might begin the typical Apple long-range process of replacing two or three existing companies within their current supply chain, but that’s nothing new over the last 25 years.
toomuchtodo · 2026-06-29 15:33:30 UTC
It is a bit unfair to lump Apple in with everyone else in the Mag7, as their strategy is arguably reasonable and prudent for the current position of the hype circle. Apple Silicon was a spectacular decision and investment. Everyone else has lost their damn mind.
petesergeant · 2026-06-29 14:53:37 UTC
> Apollo Group has $1T assets under management, I believe them over most folks on HN
Ah, the old _argumentum ad giletum Patagoniae_
serf · 2026-06-29 15:00:01 UTC
in pecunia veritas , an old HN favorite.
roboror · 2026-06-29 15:05:08 UTC
If we're doing money=smart, the mag 7 control almost $4T in assets and cash, so wouldn't you trust them even more?
toomuchtodo · 2026-06-29 15:11:14 UTC
Not based on the evidence of how they allocate capital. Meta wasted $80B on the Metaverse, for example. Apollo allocates capital as their day job. Mag7 allocates by vibes.
> Apollo allocates capital as their day job. Mag7 allocates by vibes
Lol what in the world? Is running a trillion dollar corporation anything other than allocating capital?
You can argue that Meta made a poor capital allocation decision with VR and perhaps continues to do so with AI.
toomuchtodo · 2026-06-29 15:26:30 UTC
I have an opinion based on the data, yours may differ. Favorite this thread for when the music stops. It has before (1999-2000, 2007-2008), it will again.
triceratops · 2026-06-29 15:29:45 UTC
I'm objecting to your categorization of the job of running one of the Mag7s as something other than "capital allocation". The CEO of Meta or Amazon or any of these other companies allocates capital all day long. They do it differently than an asset manager like Apollo for sure, and sometimes make worse decisions. But it's in the same category.
staticman2 · 2026-06-29 17:15:57 UTC
Vanguard has 12T under management so you believe any PDF they put out 12 times more?
toomuchtodo · 2026-06-29 17:26:52 UTC
Vanguard holds and manages index funds and are, broadly speaking, not active managers. Apollo is a sophisticated, active capital market participant|manager.
I mean, I'd assume this deck was trying to sell something. But the relevant comparison is to stonk-bros on HN LARPing as hedge fund managers, not Vanguard.
Which, to be fair, Vanguard has earned a good deal of trust from me on the passive investment side of the equation, although I don't think it's meaningful to make it linearly proportional to the total size of their managed assets? I'm not even sure how I'd operationalize that in reality.
Having actual skin in the game on getting an answer right is generally a sign of credibility, though.
That said, without context I'm not really drawing any conclusions from this.
pj_mukh · 2026-06-29 14:44:15 UTC
"I love how they show OpenRouter’s graph of token usage as if it speaks for usage across the board."
I'm not sure what was said during what looks like a deck of a presentation? I'm hoping it wasn't this, because that's an obvious misfire.
wolframhempel · 2026-06-29 14:40:50 UTC
This seems healthy to me. A market where returns are less dependent on seven mega-cap names is probably more stable, not less. If earnings growth broadens out and capital starts flowing back toward quality and free cash flow outside the obvious AI winners, that should reduce concentration risk and make the whole market less fragile.
throw0101d · 2026-06-29 14:44:39 UTC
Historically stocks that had a good run then tended to underperform:
> […] Since 1926, the median ten-year return on individual U.S. stocks relative to the broad equity market is –7.9%, underperforming by 0.82% per year. For stocks that have been among the top 20% performers over the previous five years, the median ten-year market-adjusted return falls to –17.8%, underperforming by 1.94% per year. Since the end of World War II, the median ten-year market-adjusted return of recent winners has been negative for 93% of the time. The case for diversifying concentrated positions in individual stocks, particularly in recent market winners, is even stronger than most investors realize.
I mean these stocks have been performers for decades. If you posted this 10 years ago you'd look really wrong.
Lerc · 2026-06-29 15:20:23 UTC
Yes, but when their run ends they tend to underperform.
Every time.
nvme0n1p1 · 2026-06-29 15:23:22 UTC
If a stock market observation has no predictive power, then it's worthless.
I look forward to your weather report too: "It's always sunny outside until one day it starts raining. Every time."
andrewstuart2 · 2026-06-29 15:25:40 UTC
Off topic, but I love your username.
tjwebbnorfolk · 2026-06-29 15:30:15 UTC
hey he hacked my computer his user has a home folder inside of /dev
tjwebbnorfolk · 2026-06-29 15:31:41 UTC
ye, the stock market isn't magic, it's just a collection of what people think. and people can be very wrong in a big way.
nradov · 2026-06-29 15:37:04 UTC
Buddy I think you missed the joke.
EA-3167 · 2026-06-29 17:32:10 UTC
Not at all, if someone tells me that "This stock is historically likely to regress to and beyond the mean," it's information I can use to evaluate my risk tolerance. Just because a piece of information doesn't let you time the market like a psychic doesn't make it worthless, it's just not what you were looking for.
throw0101d · 2026-06-29 18:07:01 UTC
> I look forward to your weather report too: "It's always sunny outside until one day it starts raining. Every time."
I once ran across the comment that if you simply predict tomorrow's weather will be the same as today's you'd be correct 80% of the time. Not sure how true that is (can't find the source).
There are momentum indices. Momentum is actually a strange phenomenon from the perspective of the efficient market hypothesis because it does not have an obvious risk to balance its premium compared to other empirical factors.
ohyes · 2026-06-29 15:27:01 UTC
“When the stocks don’t go up they don’t match the market which generally goes up”
ertgbnm · 2026-06-29 15:27:36 UTC
Once you lose, you have lost. Ok, but how does that help us predict when something will lose?
nradov · 2026-06-29 16:41:04 UTC
Yes, it is predictive. But only retroactively.
eternal_braid · 2026-06-29 15:33:12 UTC
Many people who replied to you seem to have missed your joke. I appreciated it.
Lerc · 2026-06-29 15:49:54 UTC
It is my curse.
Years ago, My daughter's science teacher said that school should teach a love of learning.
I replied 'I thought the point of school was to make productive worker units in society'
And while he explained to me why I was wrong I was thinking to myself 'great, now he thinks I'm a terrible person'
It seems I deadpan too effectively.
rightbyte · 2026-06-29 17:46:58 UTC
Poe's law apply to real life too.
bluGill · 2026-06-29 18:02:15 UTC
Productive workers in society need to learn new things all the time. I can't think of any career that hasn't changed in my life. I recall a garbage man (sexism probably wasn't required even then, but I never recall females) hanging off the back of the truck while the driver drove to the next house - the driver today needs to know how to operate the arm on the truck that lifts my can. Fast food used to be cooked within 10 minutes of when it was thrown, now they obviously are keeping things warm for a lot longer.
pkilgore · 2026-06-29 15:55:21 UTC
well I laughed
c22 · 2026-06-29 16:09:39 UTC
Sometimes people bring me things that are broken 'cause I like to fix stuff. They always say "it was just working!"
davedx · 2026-06-29 16:11:04 UTC
Tautologies 'R us
nkmnz · 2026-06-29 20:02:04 UTC
Predictions are difficult, especially about the future.
fittingopposite · 2026-06-30 00:16:11 UTC
That's a nice tautology
throw0101d · 2026-06-29 17:44:56 UTC
> I mean these stocks have been performers for decades. If you posted this 10 years ago you'd look really wrong.
And Japan performed ridiculously well for over decades and then stagnated for decades after that, but it averaged out between the two periods:
> Ben Carlson: It's just a really long mean reversion. You got like 22% per year from 1970 to 1989 in Japan. Small caps in Japan did 30% per year for two decades.
> It's insane. The returns almost had to be poor after that. If you put them together, the boom with the bust, it's like almost 9% per year.
> It's kind of crazy. Over 50 years, the long-term worked. It's just that over that 20 or 30-year period, it didn't work so well.
Annualized 9% per year is pretty good: the S&P 500 has average 10% since 1957 (70 years). Is there anything preventing US equities from doing the same thing: great performance from 2010 until now, and then 10+ years of stagnation starting (theoretically) tomorrow. If you look at 2000s S&P 500 you got zero returns, and the only thing that would have saved a US domestic (only) investor was having a bond allocation:
This is why diversification is important. People talk about "US stocks" doing well, but have US industrials done better than non-US industrials? US finances or energy done better than non-US? Or are "US stocks" doing better simply because tech stocks specifically have done better? Perhaps a US allocation is really a tech sector play:
> Historically stocks that had a good run then tended to underperform
This is more of a mathematical axiom than a financial effect, because you're defining "underperform/overperform" with respect to an average that contains them.
gruez · 2026-06-29 17:24:30 UTC
>> Historically stocks that had a good run then tended to underperform
>because you're defining "underperform/overperform" with respect to an average that contains them.
Why is this true? For instance, if you're comparing the GDP growth of countries in the G7, why is it that one country (eg. US) can't consistently overperform year after year?
Shouldn't you look at the YoY change instead, to compare to stock returns ? Otherwise that's like comparing market cap, and then it is obvious that a big company tends to stay big.
gruez · 2026-06-29 18:04:50 UTC
>Shouldn't you look at the YoY change instead, to compare to stock returns ?
This might work for the G7 case[1], but not the US vs DRC case, where it's an obvious case of sloping up vs sloping down. Granted, the case is contrived, but the original claim was that it was an "mathematical axiom", so it should still hold.
[1] though even in the G7 sample, you can find counterexamples. If you switch to "relative growth" you can clearly see that italy has lagging since the mid 2000s, with no accompanying faster-than-average growth to make up for it. If the claim is that "Historically stocks that had a good run then tended to underperform", then surely the opposite must also hold?
throw0101d · 2026-06-29 18:01:55 UTC
> This is more of a mathematical axiom than a financial effect, because you're defining "underperform/overperform" with respect to an average that contains them.
Most stocks suck:
> We study long-run shareholder outcomes for over 64,000 global common stocks during the January 1990 to December 2020 period. We document that the majority, 55.2% of U.S. stocks and 57.4% of non-U.S. stocks, underperform one-month U.S. Treasury bills in terms of compound returns over the full sample. Focusing on aggregate shareholder outcomes, we find that the top-performing 2.4% of firms account for all of the $US 75.7 trillion in net global stock market wealth creation from 1990 to December 2020. Outside the US, 1.41% of firms account for the $US 30.7 trillion in net wealth creation.
> Four out of every seven common stocks that have appeared in the CRSP database since 1926 have lifetime buy-and-hold returns less than one-month Treasuries. When stated in terms of lifetime dollar wealth creation, the best-performing four percent of listed companies explain the net gain for the entire U.S. stock market since 1926, as other stocks collectively matched Treasury bills. These results highlight the important role of positive skewness in the distribution of individual stock returns, attributable both to skewness in monthly returns and to the effects of compounding. The results help to explain why poorly-diversified active strategies most often underperform market averages.
And this certainly can have a financial effect on your finances: having the "wrong" stocks in your portfolio (i.e., most of them) and not have the "correct" ones will mean a (e.g.) comfortable retirement or not.
energy123 · 2026-06-29 18:35:28 UTC
This is wrong. There is no lookahead bias in the factor.
energy123 · 2026-06-29 18:39:15 UTC
It's interesting, but the usual disclaimers apply with factor research
- Does it replicate internationally?
- Is it explained by another phenomenon such as the beta anomaly or small cap premium. Implication: large caps with high beta are already known to underperform, so this this isn't a new regularity.
patrickk · 2026-06-29 14:45:14 UTC
On slide 6, they list the Mag7 stocks (not defined in a foreword), but on slide 8 they list the free cash flow (FCF) of a somewhat different set of companies. Why not stick to the Mag7 FCF only? It muddies the waters.
anonu · 2026-06-29 14:49:02 UTC
the FCF slide is for hyperscalers - which excludes TSLA, NVDA
patrickk · 2026-06-29 15:44:56 UTC
Yes but the Mag 7 is the main focus, suddenly switching to a different set of companies is comparing apples to oranges. That slide belongs in an annex.
dualvariable · 2026-06-29 16:25:05 UTC
That slide should probably be the main focus, with the rest supplementary.
That is the slide that is likely to pop the economy.
Barbing · 2026-06-29 16:39:26 UTC
I was here when you called it!
Is that Oracle down there increasingly in the negative? & Amazon's free cash flow reaching zero?
sroussey · 2026-06-29 14:49:44 UTC
Apple is not burning its cash on buildout, which would make their graph less interesting.
enopod_ · 2026-06-29 16:23:48 UTC
True. This slide also struck me the most, but for the data it shows. Free cash flow for all the hyperscalers basically evaporated in the last couple of months, with Oracle in the minus. What does this mean?
re-thc · 2026-06-29 14:45:48 UTC
It is less that the Mag 7 is starting to underperform and more that a market correction is likely and coming very soon.
anonu · 2026-06-29 14:48:41 UTC
I guess "nothing lasts forever".
rvz · 2026-06-29 15:07:17 UTC
Of course. The rest 'n vest people at FAANG companies will tell you its lasts forever 6 years ago. Now they are scared for their jobs.
This is why we are seeing a correction at those companies, perks and free food going away with constant layoffs and all time low morale.
Now the party is at Nvidia. But I will tell you that that will not last forever either.
Espressosaurus · 2026-06-29 15:23:54 UTC
Everything reverts to the mean in the long term.
Maybe even the state of unprecedented relative peace we had enjoyed.
guluarte · 2026-06-29 14:49:14 UTC
This feels like telecom "massive demand, bad returns".
If a good enough model can be swapped in every few months, the value moves away from the model and toward cheap inference. That is great for users, but not always great for returns on huge capex.
interestpiqued · 2026-06-29 14:53:54 UTC
The only mag 7 that had a shooter chance with models is Google rn so doesn’t seem like there being good enough models will be that negative for mag 7. Might even be better since they are the providers for the inference
bwfan123 · 2026-06-29 14:50:13 UTC
I like that the invisible hand of market is slapping the Mag-7 for capex which is the only way to discipline them. Investors are waking up to say: hey, you are spending all your profits on data-centers, where is the return for me ? But, it surprises me that there are vast pools of capital which we collectively call the "market" that makes these calculations, or maybe a simpler causal explanation is the missing stock repurchase bid. At some point, one of the hyperscalers (msft ?) will break from the pack and announce reductions to capex and increase stock repurchases to stem the decline.
ares623 · 2026-06-29 14:58:51 UTC
Why is Apple included though?
tomrod · 2026-06-29 15:03:24 UTC
Share repurchases also aren't great I guess?
TravisJamison · 2026-06-29 15:17:05 UTC
Markets generally love buybacks.
The market is both happy that Apple didn’t spend all its cash on AI build-out, but also at the same time angry that Apple is “missing AI”.
Not to mention the grumblings that Apple has peaked.
symfoniq · 2026-06-29 20:31:09 UTC
And did you see Apple’s recent price increases?
eitally · 2026-06-29 15:54:12 UTC
My perception is that the market lumps companies together into sectors whether they have similar business models or not. When one is punished, other associated firms tend to be, too. You see this in any industry (not as a guaranteed rule, but in general).
DaedalusII · 2026-06-29 16:38:40 UTC
because $100bn annual revenue, high FCF, high margin. Its a monster.
twoodfin · 2026-06-29 17:47:47 UTC
You can simultaneously believe that the hyperscalers are becoming capital-intensive long-term in a way that’s bad for their profits and that as a result will be raising the COGS of Apple’s business in a way that also hurts profits.
ambicapter · 2026-06-29 19:39:06 UTC
> increase stock repurchases
buy high, sell...whenever?
ctoth · 2026-06-29 20:20:40 UTC
> increase stock repurchases
I just can't wait to get back to when innovation meant financialization, can you?
Building things? Real Jobs for real electricians? Real buildings? Real compute? Why would anybody want that!
gandalfgeek · 2026-06-29 14:55:56 UTC
6 months ago when Mag7 was overperforming everyone was worried about it being too high a fraction of the S&P500.
andxor · 2026-06-29 15:27:08 UTC
Exactly. As usual people are bullish at the top and bearish at the bottom.
> The once high-flying "Magnificent Seven" are looking more like the Dreadful Seven.
> The why: Wall Street is growing increasingly impatient with Big Tech's astronomical capital expenditures on artificial intelligence, projected to balloon 70% to exceed $700 billion this year.
ianm218 · 2026-06-29 15:02:52 UTC
What has been the best way to determine return on the AI specific capex for hyperscalers?
I would naively expect Microsoft’s to be the highest since they are probably mainly just selling access to their capex through cloud since they aren’t seriously pursuing frontier AI, I’d imagine Google to be in the middle (selling TPUs, general cloud GPUs, Gemini, revenue lift on ads from better AI) but also spending heavily on infra to compete with OAI/ Anthropic, and then Meta to be on the low end since they are likely getting serious revenue lift from AI but not monetizing their models by API.
OMG - ridiculous to evaluate it based on the last 3 mo.
zerobees · 2026-06-29 15:12:11 UTC
I am invested in some of the companies that are downstream of the capital expenditures of Big Tech (e.g., COHR), so I have nothing to complain about.
I am really struggling to see what's the investment thesis behind Google valuation increasing 2x in response to AI, though. Assuming no magical AGI singularity, by the end of the day, they're still selling the same services, but the services have gotten more expensive for them to provide. Everyone was already using Google Search, but now, provisioning AI summaries on top of requires more compute. Everyone was already using Google Docs and Meet, but now, AI features cost Google more. Etc, etc.
The only place where they stand to make money is selling AI compute to enterprises. But with the current supply-chain challenges, the margins there are probably getting thinner.
solumunus · 2026-06-29 15:15:59 UTC
The market runs on memes, hype and fraud. Fundamentals haven’t mattered for a long time.
senordevnyc · 2026-06-29 15:23:58 UTC
I guess Google’s Q1 earnings of $62 billion were just hype.
That's trailing and mighty impressive. However when you decided to quote Q1 earnings. $62bn of Q1 earnings ends up at only $10bn. The LTM numbers are less relevant.
AI-related capex is one hell of a drug.
senordevnyc · 2026-06-29 18:47:47 UTC
The discussion here is whether Google can make money in the AI world. They obviously can. Whether they turn around and spend that money on capex to try and increase future earnings is irrelevant.
torginus · 2026-06-30 00:56:57 UTC
Yeah, but I think the recent explosion of both memory prices and stock prices of memory makers highlight that datacenter GPUs cost about 10x as much as the would in a healthy supply-demand situation. Gaming equivalents of datacenter GPUs (I know they are not the same, but those had a hell of a lot of compute and bandwidth) cost less than 1/10th when normalized to FLOPS.
So its scarcity pricing. If a paper were to come out tomorrow proving that additional compute has diminishing returns, there would be hell to pay in markets.
Comments
And the bar chart for token costs, really? As if that’s information? Their sources are the API docs ffs. If they had at least modeled something to estimate token costs that would be interesting, but showing the public prices and calling it research is dumb.
If only that's a measure of if they know how to time the stock market. The real measure is how much they earn not if their sales or marketing department manages to get more customers.
That's more than their combined FCF, and they're borrowing to bridge the gap.
Ah, the old _argumentum ad giletum Patagoniae_
https://finance.yahoo.com/sectors/technology/articles/mark-z...
https://www.nytimes.com/2026/03/19/technology/mark-zuckerber... | https://archive.today/iEGAj
Lol what in the world? Is running a trillion dollar corporation anything other than allocating capital?
You can argue that Meta made a poor capital allocation decision with VR and perhaps continues to do so with AI.
https://en.wikipedia.org/wiki/Apollo_Global_Management
Which, to be fair, Vanguard has earned a good deal of trust from me on the passive investment side of the equation, although I don't think it's meaningful to make it linearly proportional to the total size of their managed assets? I'm not even sure how I'd operationalize that in reality.
Having actual skin in the game on getting an answer right is generally a sign of credibility, though.
That said, without context I'm not really drawing any conclusions from this.
I'm not sure what was said during what looks like a deck of a presentation? I'm hoping it wasn't this, because that's an obvious misfire.
> […] Since 1926, the median ten-year return on individual U.S. stocks relative to the broad equity market is –7.9%, underperforming by 0.82% per year. For stocks that have been among the top 20% performers over the previous five years, the median ten-year market-adjusted return falls to –17.8%, underperforming by 1.94% per year. Since the end of World War II, the median ten-year market-adjusted return of recent winners has been negative for 93% of the time. The case for diversifying concentrated positions in individual stocks, particularly in recent market winners, is even stronger than most investors realize.
* https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4541122
Every time.
I look forward to your weather report too: "It's always sunny outside until one day it starts raining. Every time."
I once ran across the comment that if you simply predict tomorrow's weather will be the same as today's you'd be correct 80% of the time. Not sure how true that is (can't find the source).
Allegedly momentum investing does pretty well:
* https://en.wikipedia.org/wiki/Momentum_investing
(I'm more of an index guy myself.)
Years ago, My daughter's science teacher said that school should teach a love of learning.
I replied 'I thought the point of school was to make productive worker units in society'
And while he explained to me why I was wrong I was thinking to myself 'great, now he thinks I'm a terrible person'
It seems I deadpan too effectively.
And Japan performed ridiculously well for over decades and then stagnated for decades after that, but it averaged out between the two periods:
> Ben Carlson: It's just a really long mean reversion. You got like 22% per year from 1970 to 1989 in Japan. Small caps in Japan did 30% per year for two decades.
> It's insane. The returns almost had to be poor after that. If you put them together, the boom with the bust, it's like almost 9% per year.
> It's kind of crazy. Over 50 years, the long-term worked. It's just that over that 20 or 30-year period, it didn't work so well.
* https://rationalreminder.ca/podcast/412 (~4m20s)
Annualized 9% per year is pretty good: the S&P 500 has average 10% since 1957 (70 years). Is there anything preventing US equities from doing the same thing: great performance from 2010 until now, and then 10+ years of stagnation starting (theoretically) tomorrow. If you look at 2000s S&P 500 you got zero returns, and the only thing that would have saved a US domestic (only) investor was having a bond allocation:
* https://www.forbes.com/sites/advisor/2010/09/13/its-not-real...
This is why diversification is important. People talk about "US stocks" doing well, but have US industrials done better than non-US industrials? US finances or energy done better than non-US? Or are "US stocks" doing better simply because tech stocks specifically have done better? Perhaps a US allocation is really a tech sector play:
* https://ofdollarsanddata.com/should-your-portfolio-be-100-us...
This is more of a mathematical axiom than a financial effect, because you're defining "underperform/overperform" with respect to an average that contains them.
>because you're defining "underperform/overperform" with respect to an average that contains them.
Why is this true? For instance, if you're comparing the GDP growth of countries in the G7, why is it that one country (eg. US) can't consistently overperform year after year?
https://ourworldindata.org/grapher/gdp-per-capita-worldbank?...
Or if you want make it even more clear, you can construct a index consisting of two countries: a normal country (eg. US) and a basketcase (eg. DRC):
https://ourworldindata.org/grapher/gdp-per-capita-worldbank?...
This might work for the G7 case[1], but not the US vs DRC case, where it's an obvious case of sloping up vs sloping down. Granted, the case is contrived, but the original claim was that it was an "mathematical axiom", so it should still hold.
[1] though even in the G7 sample, you can find counterexamples. If you switch to "relative growth" you can clearly see that italy has lagging since the mid 2000s, with no accompanying faster-than-average growth to make up for it. If the claim is that "Historically stocks that had a good run then tended to underperform", then surely the opposite must also hold?
Most stocks suck:
> We study long-run shareholder outcomes for over 64,000 global common stocks during the January 1990 to December 2020 period. We document that the majority, 55.2% of U.S. stocks and 57.4% of non-U.S. stocks, underperform one-month U.S. Treasury bills in terms of compound returns over the full sample. Focusing on aggregate shareholder outcomes, we find that the top-performing 2.4% of firms account for all of the $US 75.7 trillion in net global stock market wealth creation from 1990 to December 2020. Outside the US, 1.41% of firms account for the $US 30.7 trillion in net wealth creation.
* https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3710251
> Four out of every seven common stocks that have appeared in the CRSP database since 1926 have lifetime buy-and-hold returns less than one-month Treasuries. When stated in terms of lifetime dollar wealth creation, the best-performing four percent of listed companies explain the net gain for the entire U.S. stock market since 1926, as other stocks collectively matched Treasury bills. These results highlight the important role of positive skewness in the distribution of individual stock returns, attributable both to skewness in monthly returns and to the effects of compounding. The results help to explain why poorly-diversified active strategies most often underperform market averages.
* https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2900447
And this certainly can have a financial effect on your finances: having the "wrong" stocks in your portfolio (i.e., most of them) and not have the "correct" ones will mean a (e.g.) comfortable retirement or not.
- Does it replicate internationally?
- Is it explained by another phenomenon such as the beta anomaly or small cap premium. Implication: large caps with high beta are already known to underperform, so this this isn't a new regularity.
That is the slide that is likely to pop the economy.
Is that Oracle down there increasingly in the negative? & Amazon's free cash flow reaching zero?
This is why we are seeing a correction at those companies, perks and free food going away with constant layoffs and all time low morale.
Now the party is at Nvidia. But I will tell you that that will not last forever either.
Maybe even the state of unprecedented relative peace we had enjoyed.
If a good enough model can be swapped in every few months, the value moves away from the model and toward cheap inference. That is great for users, but not always great for returns on huge capex.
The market is both happy that Apple didn’t spend all its cash on AI build-out, but also at the same time angry that Apple is “missing AI”.
Not to mention the grumblings that Apple has peaked.
buy high, sell...whenever?
I just can't wait to get back to when innovation meant financialization, can you?
Building things? Real Jobs for real electricians? Real buildings? Real compute? Why would anybody want that!
https://finance.yahoo.com/markets/article/magnificent-7-stoc...
> The once high-flying "Magnificent Seven" are looking more like the Dreadful Seven.
> The why: Wall Street is growing increasingly impatient with Big Tech's astronomical capital expenditures on artificial intelligence, projected to balloon 70% to exceed $700 billion this year.
I would naively expect Microsoft’s to be the highest since they are probably mainly just selling access to their capex through cloud since they aren’t seriously pursuing frontier AI, I’d imagine Google to be in the middle (selling TPUs, general cloud GPUs, Gemini, revenue lift on ads from better AI) but also spending heavily on infra to compete with OAI/ Anthropic, and then Meta to be on the low end since they are likely getting serious revenue lift from AI but not monetizing their models by API.
Amazon $200B
MS $190B
Alphabet $175B-$185B
Meta $115B-135B
https://finance.yahoo.com/news/apple-lazy-ai-strategy-could-...
I am really struggling to see what's the investment thesis behind Google valuation increasing 2x in response to AI, though. Assuming no magical AGI singularity, by the end of the day, they're still selling the same services, but the services have gotten more expensive for them to provide. Everyone was already using Google Search, but now, provisioning AI summaries on top of requires more compute. Everyone was already using Google Docs and Meet, but now, AI features cost Google more. Etc, etc.
The only place where they stand to make money is selling AI compute to enterprises. But with the current supply-chain challenges, the margins there are probably getting thinner.
Focus on cash/cashflow.
AI-related capex is one hell of a drug.
So its scarcity pricing. If a paper were to come out tomorrow proving that additional compute has diminishing returns, there would be hell to pay in markets.