Post New

Anyone w NVDA positions?

Whats everyone doin? Pretty insane parabolic move here. Could retrace to 100 or even 50 in a heartbeat while still in a “uptrend”

Holding tight. Let winners ride
Selling calls?
Selling out/trimming?
Replacing stock w calls (ITM to harvest extrinsic w a downdraft, Mungofitxh idea or OTM to “lock in” more profits but more likely to bleed extrinsicvalue)
Buying puts?

Would love to hear what you all think

Yes, since 1/2h before your post I initiated a NVDA position: $130 June 2026 Puts.

I have a position in my quant portfolio. It has been pleasant to watch. I'll keep it till the quant formula says to ditch it, which I check only every two months. Broadly speaking, this particular screen will have it replaced only when there is another high-sales-growth Nasdaq type firm with better price momentum. I expect I'll give back some of the profits, maybe most, but (statistically) not all.

An interesting little snip from a recent copy of the Economist:

"Are the pros any better? Only up to a point. Last year Mr Imas and colleagues published a paper on the buying and selling choices of 783 institutional portfolios with an average value of $573m. Their managers were good at buying: the average purchase, a year later, had beaten the broader market by 1.2 percentage points. But they would have been better off throwing darts at the wall to select which positions to exit. After a year, sales led to an average of 0.8 percentage points of forgone profit compared with a counterfactual in which the fund selected a random asset to sell instead.

"Unlike retail traders, the pros were not clinging on to losers. Yet neither were they making selling decisions analogously to how they make buying ones: by choosing the asset adding the least to their risk-adjusted return and offloading it. Instead they used a simpler heuristic, disproportionately selecting positions where relative performance had been very bad or good, and exiting those. As a result, they were throwing away two-thirds of the excess returns their skilful buying had won them."


The optimal strategy probably depends a whole lot on your intended investment horizon. But, just maybe, "yikes it's up a lot in a short period" might not be something you want to react to.

Jim

ADX top ten holdings, that's enough tech for me.

https://www.adamsfunds.com/funds/diversified-equit...

You sure see a lot of weird effects when the tails of a few stocks dominate the dog of the market.

"After the close Friday, the XLK ETF will be rebalanced to drop Apple's 22% share down to 4.5% and increase Nvidia's 5.9% share up to 21.1%, based on Bloomberg estimates...
All told, Friday's rebalance should force $12.7 billion in Apple stock to be sold and $11 billion of Nvidia to be bought. That's close to the dollar amount of Apple shares that trade any given day, and about one-quarter of the dollar amount that Nvidia trades daily. In other words, these are material amounts."

from
https://finance.yahoo.com/news/nvidias-surge-revea...

Jim

ADX top ten holdings, that's enough tech for me.

Same top 10 holdings as SPY and QQQ except for Adams Natural Resources Fund, Inc.
Top 10 are 37% of net assets.
30% in tech

SPY, top 10 are 34%. 32% in tech.

CAGR Return, Jan 1993 to Jun 2024:
ADX 9.04%
SPY 10.37%

ADX ER 0.61%
SPY ER 0.09%


Looking at the chart, they both go up and go down at the same time.
https://testfol.io/?d=eJytjzFPw0AMhf%2BL5xvCwnAz6s...

Index fund wins again.

After the close Friday, the XLK ETF will be rebalanced to drop Apple's 22% share down to 4.5% and increase Nvidia's 5.9% share up to 21.1%, based on Bloomberg estimates.

All of this stems from Great Depression-era investor protection laws, which require that indexes limit the concentration of individual stocks to earn the label "diversified."
...

Briefly stated, there are four companies — Nvidia, Apple, Microsoft, and Broadcom — that overrun the critical 4.8% threshold for individual names in a diversified index. And because they collectively exceed 50% of the entire index by weight, the weights of the smallest members are reduced according to a formula until all of the legal thresholds are respected.


Amazing. So the rule says that if the top 4 are over 50%, at neutral weights, you reduce the smallest ones of the top 4 to 4.5%?

It seems so - here are the rules:

6. The sum of the companies with weights greater than 4.8% cannot exceed 50% of the total index weight. These caps are set to allow for a buffer below the 5% limit.
7. If the rule in step 6 is breached, rank all companies in descending order by FMC weight, and reduce the weight of the smallest company whose weight is greater than 4.8% that causes the step 6 breach to 4.5%. This process continues iteratively until step 6 is satisfied.

https://www.spglobal.com/spdji/en/documents/method...

In other words, when Microsoft is $3.2t and Apple is $3.1t and Nvidia is $3.0t and Broadcom is $809b, you end up having MSFT at 22.1%, AAPL at 22.0%, NVDA at 5.9% and AVGO at 5.6%. Then when NVDA moves up from $3.0t to $3.2t and drops Apple to 3rd place, you get NVDA and MSFT with the >20% stakes and it's AAPL that has to be cut to 4.5%. And if NVDA drops back to number 3, you reverse all that? Seems like a rule dreamed up by amateurs, not the premier indexer S&P. Their brokers must be having a field day with commissions.

dtb

All told, Friday's rebalance should force $12.7 billion in Apple stock to be sold and $11 billion of Nvidia to be bought. That's close to the dollar amount of Apple shares that trade any given day, and about one-quarter of the dollar amount that Nvidia trades daily.

This suggests that the dollar amount of Nvidia shares trading daily is approximately 4 times the dollar amount of Apple shares trading daily. For 2 companies with approximately the same market cap (at least for the time being), that seems ... odd. Perhaps it reflects the nature of many of Nvidia's "shareholders."

Broadly speaking, this particular screen will have it replaced only when there is another high-sales-growth Nasdaq type firm with better price momentum.

High sales growth and momentum sounds like it could be quite the ride. Does this screen wholly make up the aggressive part of your mechanical allocation, or are there other screens that also make up the aggressive piece?

Broadly speaking, this particular screen will have it replaced only when there is another high-sales-growth Nasdaq type firm with better price momentum.
...
High sales growth and momentum sounds like it could be quite the ride. Does this screen wholly make up the aggressive part of your mechanical allocation, or are there other screens that also make up the aggressive piece?

It's around 6% of my current quant portfolio, so around 20-25% of the aggressive chunk which is itself around a quarter of the total.
Yes, it's a bit volatile, but the *average* is quite good, so it's intended as yeast for the overall loaf.

The backtest is pretty good, anyway. That doesn't mean all that much, but it's better than a bad backtest!
For example: six picks, equally weighted, hold 2 months and repeat, January 2000 to May 2024 inclusive. (2000 was not an auspicious time to start a Nasdaq momentum screen)
CAGR 19.8% versus S&P 500 total return 7.4%
First five months this year (not annualized), up 38.7% versus S&P 12.1%. It has held NVDA all year so far.
It's a wild ride, but then so is an index. Worse rolling year for the 6-stock screen was -36% versus -46% for the S&P 500 and roughly -60% for the Nasdaq 100 equal weight. The odds of a positive rolling year are about a tie with the S&P. As with many a quant screen, the problem isn't the wildness of the ride, it's the long stretches that it fails the "what have you done for me lately" test. For example, that hypothetical 6-stock portfolio went net nowhere from August 2015 through April 2020. Patience and faith are required in this field, and sometimes they aren't ultimately rewarded. That's why I run more than one screen, and that's why I strongly favour screens that have been working at least a little for several years after they were created.

I'm sure I'll be forgiven for not posting the exact screen details of this one, but it picks large Nasdaq-listed stocks from the Value Line 1700 set of firms based entirely on sundry momentum and sales growth measures, nothing else.

Jim

After the close Friday, the XLK ETF will be rebalanced to drop Apple's 22% share down to 4.5% and increase Nvidia's 5.9% share up to 21.1%, based on Bloomberg estimates.


Good thing they only readjust every trimester, because after a 3% drop today, NVDA is back down to #3... It would be funny if Standard and Poor's had to buy and sell $20b worth of stock every trimester to respect this crazy rule, with the top 3 (now MSFT, AAPL, NVDA in that order,) currently at $3.3t-$3.2t-$3.1t, respectively.

Weighting. Each index is capped market capitalization weighted. For capping purposes, the indices are rebalanced quarterly after the close of business on the third Friday of March, June, September, and December using the following procedures:
1. The rebalancing reference date is the second Friday of March, June, September, and December.
2. With prices reflected on the rebalancing reference date, adjusted for any applicable corporate actions, and membership, shares outstanding and IWFs as of the rebalancing effective date, each company is weighted by FMC. Modifications are made as defined below.
3. If any company has an FMC weight greater than 24%, the company’s weight is capped at 23%, which allows for a 2% buffer. This buffer is meant to mitigate against any company exceeding 25% as of the quarter-end diversification requirement date.

So if the markets close in 2 hours, with the 3 big companies still in the current order (MSFT, AAPL, NVDA), it would be ironic that they would be selling and buying all those $24b worth of shares to get AAPL down to 5% and NVDA up to 23%, even though AAPL is back to #2 and NVDA is back to #3.


6. The sum of the companies with weights greater than 4.8% cannot exceed 50% of the total index weight. These caps are set to allow for a buffer below the 5% limit.

If they don't want companies with weights greater than 4.8% totalling to more than 50% of the total weight, it's because they think a 5% weight is too big, so they certainly shouldn't want 2 companies with 22% each (MSFT and AAPL). But they will not achieve that result at all by just switching out AAPL for NVDA. It is ironic that they will be doing all this buying and selling and will still end with 2 companies that add up to 43%, and in fact, by the time they do it, they will have just put the #3, and by far the most volatile of the 3, into a 21% stake, with MSFT staying at 22%! And if they don't want 10 companies with 5% each representing half the index (which would require an adjustment, under their rules, it would be simple to do so, without much buying or selling each trimester, by just capping the weight of any company at a lower number, like 4% or 3%.

dtb

Option is #1:

I have no idea whether NVDA is currently priced right. Their stock has come a long, long way ... but their future is very bright. (Worth twice all the oil companies combined?) If I was forced to guess, I'd say they'll be down a year from now. It feels a little like 2000 when shorting certain tech stocks was the correct trade ... but a bad long term strategy.

The important thing about NVDA, is to understand what it means for the rest of the market. Right now at least, they're a picks and shovels company. At some point, human level intelligence can/will be deployed into everything that moves or with a plug. Training that intelligence is a big business. (But it's nowhere near the value of the products it enables.) There won't be many companies with the capital and know-how to compete in this space. This is a long boring presentation on the economics of Robotaxi and Bot, two of the products AI enables:

https://www.youtube.com/watch?v=bSchsi1p_IM

There is a similarity between NVDA and TSLA. Both were misunderstood. NVDA slumped when people thought that they were a gaming or BTC play. The market figured out (after their CEO talked about it for years) that they're an AI company. Same will be true for TSLA, the market thinks they're a car company when actually they're an AI company ... maybe the biggest (not in revenue but in potential).

--------------------------------------------

Option #2, and in the spirit of trolling for rec's:

This whole AI thing is just a fad. Nobody will even remember it in ten years.

I have no idea whether NVDA is currently priced right. Their stock has come a long, long way ... but their future is very bright. (Worth twice all the oil companies combined?) If I was forced to guess, I'd say they'll be down a year from now. It feels a little like 2000 when shorting certain tech stocks was the correct trade ... but a bad long term strategy.

Sounds not crazy.

I would add this: the really big firms who are their clients are spending VAST amounts of money on their products. They have the means, opportunity, and motive to build an alternative themselves. Nvidia doesn't merely design chips, their software ecosystem is an excellent lock-in I'm told, but even that merely becomes a hurdle if someone very rich and motivated decides to find a way out of spending so much money. So...their lead may not last very well, meaning their moat may fade in a medium time frame. Nobody with an alternative likes to have a supplier making vast margins.

But the best way to think about the pricing of a business with a "high growth" price level is to think what has to happen for investment at current levels to turn out to have been a good one.
A possible scenario:
* They are quite profitable, so we can think about earnings rather than building it up from sales and margin trajectories.
* One source has estimated old fashioned earnings at $2.60 this year and $3.35 next year (put in whatever numbers you like).
* That's a growth rate of 29%.
* Imagine their impressive earnings growth rate fades so that each year is only X% of the prior year's growth rate, so we have a glide path of growth rates, giving us a trajectory of earnings.
* We add the assumption that no firm trades at a multiple over 19 forever, and so few manage it for very long that it can be ignored as a possible assumption.
* So, we have an earnings growth "shape" and a terminal multiple.
* I usually take the average EPS 5-10 years into the future, and estimated share price from that, to get a CAGR annualized from 7.5 year hold.
So, we end up with the following: to get a 9.0%/year return starting from here (nominal), you need each year's earnings growth rate to be about 93.3% of the prior year's earnings growth rate (all earnings nominal).
i.e., this series:
EPS growth 2024-2025: +29%
2025 to 2026: +27%
+25%
+23%
+22%
+20%
+19%
+17%
+16%
+15%

So, having worked backwards, this makes it a simpler question: to believe that the current price is a good place for a medium-long term hold, getting you you sort of have to find that trajectory to be sufficiently certain to be believable to allow a reasonable margin of safety -- or at least more certain than other investment alternatives available to you.

Note, none of this is a forecast. I'm just suggesting a way of working backwards that I have sometimes found useful...what HAS to happen for the current price to be a decently good one?

Personally I don't immediately see that as a sensible baseline assumption, primarily because I think they will see some tough competition at some point. But I do currently hold a tiny bit of the stock for the short term as a momentum play. Even if it's a bubble (which it may not be), a bubble is a terrible thing to waste. The thing about bubbles is, they generally burst a lot more slowly than most people realize. As long as you know firmly in advance that you're not going to continue investing as if the bubble were still happening long after it has plainly ended, there is sometimes some good money to be made. You don't have to nail the top, you mainly have to know that you're going to stop that kind of investing when it's time.

Jim

This guy has been good at spotting trends and getting out.
Grok:
Jason Camillo is known for his unconventional approach to investing, which he refers to as "information arbitrage." He looks for "game-changing things in real-life, that are having game-changing impacts on real companies." This involves identifying trends and opportunities before they become mainstream and investing aggressively when he spots these high-conviction, material changes. He then exits the trade when financial media starts talking about the trend he spotted, indicating that the opportunity may be over.

Following his strategy, I'd say it's time to exit NVDA.

[Jason Camillo] has been good at spotting trends and getting out.
Following his strategy, I'd say it's time to exit NVDA.

So has he noticed the trend of the market hitting new high after new high and gotten out? The financial media has been talking about this for quite a while.


This involves identifying trends and opportunities before they become mainstream and investing aggressively when he spots these high-conviction, material changes.

Which is exactly the same thing that EVERY market guru claims to do.

I think he is just talking his book.

i.e., this series:
EPS growth 2024-2025: +29%
2025 to 2026: +27%
+25%
+23%
+22%
+20%
+19%
+17%
+16%
+15%

I've been thinking along the same lines, slowly declining growth in profit as time passes. But when I look at the numbers, not necessarily these, any numbers, it seems kind of absurd to me. These types of numbers literally assume that there will be no recession anytime in the next 10 years, and that none of these big companies (that create these profits) will reduce capex over that period. The likelihood of that happening is pretty close to zero.

Maybe the numbers will be something like +30%, +25%, +25%, -3%, -20%, -2%, 0%, +15%, +17%, etc instead?

Maybe the numbers will be something like +30%, +25%, +25%, -3%, -20%, -2%, 0%, +15%, +17%, etc instead?

Indeed. The trajectory might be anything.
My own approach uses the average real EPS 5-10 years out. Those numbers could be all the same, or wildly varying. I sort of assume there will be a bad year or two in there, but don't try to model it explicitly. Since we don't know what they'll be, assuming a smooth trajectory is as good as any other assumption--you can at least get a feel for whether that trend seems plausible.

It's pretty surprising to think about, but the earnings in the next few years don't really matter much for the value of a share. (gasp!) They're useful mainly for trying to estimate the longer term, and evaluate how well the business model is already working.

I use a terminal multiple approach not because the later earnings don't matter, they dominate in fact. I stop there because I don't think there are any companies predictable enough past the decade mark. If you can't predict that far, you can't predict that their growth will still be above average, so you might as well assume that their valuation multiples will be relatively ordinary by then. Since you have your valuation multiple to use, you don't need the list of later earnings. I assume it will always be something in the teens, unless something has gone horribly wrong.

Jim

It looks like the rebalance happened. And it wasn't as extreme as some people feared.

Last week the top weights were something like
MSFT 22%
AAPL 22%
NVDA 6%

I looked just now (6/28/24) and the weights are
MSFT 11.6%
AAPL 10.5%
NVDA 9.8%

Followed by
AMZN 6%
META 3.9%

Here's a good (too long) discussion of humanoid robots:

https://www.youtube.com/watch?v=AC004kwZO-Q&t=800s

TL;DR: Convergence of different technologies are enabling an orders of magnitude change in the cost and capabilities of robots. Look to the smart phone for an analogy only this is more on the scale of electricity or fire. Humanoid will be the dominant embodiment because so many different skills are enabled with our biological form and the world is already designed around us.


This, and products like this, is why Nvidia is valued at 3T. Current earnings are instructive only in that they establish them as having the capital and skill to execute on their business plan. The products they'll be selling in ten years do not exist today.

The only way to properly value a company like Nvidia is to guess at what they'll be selling, estimate how much of that they'll be selling, hypothesize what the margins are on those products and then take a wild-assed guess at what the multiple will be on those earnings. When I've done this kind of analysis before I break it down into a good and bad outcome, give each an equal weight to arrive at a value ... and then discount it back to today. (Not an easy task, requires multiple napkins.) Be prepared to spend a lot of time and energy. Even then, NVDA belongs in the "too hard" pile for 99% of us. That's where it's been for me.

The other investing methodology is the "Jesus take the wheel" strategy. Invest in a great CEO, which Nvidia certainly has. Still, would have been better to figure all this out before the recent 10X.

Post New

Berkshire Hathaway FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.