AI-linked companies are now >10% of global market cap

ustonapc · February 18, 2024, 8:06pm

Dear all,

On Wednesday, Nvidia will report their Q4 2023 earnings results.

Since their last earnings report, on November 21st, the stock has gained an incredible 45% in value.

Over these 3 months, the stock has added $600 BILLION in market cap.

In other words, Nvidia has added an average of $10 billion PER TRADING DAY since their last earnings report.

Most AI companies now use Nvidia A100 and H100 chips.

Can Nvidia continue its historic run? Based on Marco’s logic @marco (if I understand correctly from his other post), it is time to sell as everyone piles in.

However, I think we should see the Q4 results before deciding. Currently, the options market is pricing in a massive 11% swing in NVDA stock post-earnings.

Nvidia is currently just ~12% away from becoming a $2 trillion company.

Regards
James

The top 10 assets in terms of market capitalization as of 2/16/2024

ustonapc · February 21, 2024, 10:52pm

Dear all,

Breaking : NVIDIA SHARES EXTEND AFTER-HOURS GAINS; LAST UP 8%

NVDA has an incredibly impressive quarter
EPS: $5.16 vs $4.60 est.
Rev: $22.1B vs $20.41B est.
Adj. Gross Margin: 76.7% vs 75,4% est.

Not a good idea to short NVDA or buy puts right now.

Regards
James

WalterW · February 21, 2024, 11:14pm

Grabbed this from twitter;

FCF: $11.246B vs $1.739B (Up 546.69% YoY)

Amazing.

ustonapc · February 22, 2024, 5:35pm

Dear all,

Nvidia shares jumped 15% this morning following the 8% rise in after hours yesterday.

This is an update from CNBC.

Regards
James

Nvidia shares pop 15% after AI-fueled bumper earnings

Published Thu, Feb 22 20244:05 AM ESTUpdated 2 Hours Ago

Key Points

Nvidia shares jumped 15% on Thursday morning, a day after the chip giant posted bumper earnings that beat Wall Street estimates, driven by excitement over artificial intelligence.
The company reported revenue of $22.10 billion for its fiscal fourth quarter, a rise of 265% year-on-year, while net income surged 769%.
Nvidia chips are used to train the huge AI models such as those developed by Microsoft and Meta.

In this article

NVDA+99.92 (+14.81%)

Jensen Huang, president of Nvidia, holding the Grace hopper superchip CPU used for generative AI at the Supermicro keynote presentation during Computex 2023.

Nvidia shares jumped 15% on Thursday morning, a day after the chip giant posted bumper earnings that beat Wall Street estimates.

The U.S. tech giant reported revenue of $22.10 billion for its fiscal fourth quarter, a rise of 265% year on year, while net income surged by 769%, as the company continues to see a boost from excitement over artificial intelligence.

Nvidia chips are used to train the huge AI models such as those developed by Microsoft
and Meta.

Nvidia sees no signs of slowing. The company forecast its revenue in the current quarter will hit $24 billion, way ahead of estimates.

“Fundamentally, the conditions are excellent for continued growth” in 2025 and beyond, Nvidia CEO Jensen Huang told analysts on Wednesday, adding to the bullish sentiment around the stock.

Nvidia’s Data Center business, which includes the company’s H100 graphics cards that are used for AI training, posted sales of $18.4 billion in the fourth quarter, representing 409% year-on-year growth.

The positive outlook from Nvidia prompted a round of broker upgrades Thursday. JPMorgan raised its price target on Nvidia’s stock from $650 to $850, while Bank of America Global Research hiked its target from $800 to $925.

Nvidia’s stock closed at $674.72 on Thursday. The shares were under pressure ahead of the earnings report as traders took profit and investors were concerned that Nvidia might not be able to hit lofty expectations.

But its market-beating numbers dispelled those fears and also pulled other global chip stocks higher.

Victor1991 · February 22, 2024, 7:27pm

An interesting question in this context in my opinion would be:

“What kind of data and/or ranking system would a user for Portfolio123 need for a company like Nvidia to appear in the top ranks, while (still) ensuring consistent and robust outperformence in a backtest over longer timeframes (15+ years)?”

Any time I try to target companies like that with the current available data/factors that P123 offers, my returns drop.

ustonapc · February 22, 2024, 7:34pm

Victor,

If you believe in the Magnificent 7 as Bloomberg/CNBC now calls them and they will continue to drive Nasdaq and S&P 500 higher from today’s ATH.

You can try what I am doing and use momentum to buy these name
ticker(“tsla,amzn,googl,meta,msft,aapl,nvda”)
and make sure that they are not overbought.
rsi(5)<90 and rsi(14)<85

Regards
James

EDIT : Here are the buy rules for my “Magnificent 7” screen

plan_trader · February 23, 2024, 12:10am

Hey James,
And what do the results of your screen look like? Save me typing in all of that.
Cheers
Gary

ustonapc · February 23, 2024, 12:19am

plan trader,

I only started using the screen last Sept/Oct.

Here is the 1 year and 3 year backtest performance.

Regards
James

plan_trader · February 23, 2024, 1:34am

Thank you James. Pretty impressive , but I guess who knows what NVDA et al will do going forward.

marco · February 23, 2024, 3:05am

Happy to report I did not sell any of my NVDA prior to earnings. Not sure what I said anymore but I believed I implied that there’s nothing stopping the race for artificial general intelligence (AGI). It is the next frontier, like it or not. Something like that.

I was worried that there are starting to be many more competitors so I put a hedge in place (which also ended up making money ) .

Regarding competition I’m not sure why NVDA is selling so much more than the others. Is it because it also has the best software ecosystem which is much harder to attack? Don’t all the ML packages run perfectly fine with hardware other than NVDA? Or are they just able to make many more than the others. Also, is NVDA going to turn out to be the Intel of CPUs, and be taken over by an newcomer ?

We’ll see, but $1000 is a given now. I’ll wait.

Victor1991 · February 23, 2024, 9:06am

James,

That’s an imperssive Sharpe ratio. For me to be confident in a strategy however, I would want the returns to hold over the longer timeframes that I mentioned previously (15+ years). The strategy would also have to be without any lookahead bias (which in my opinion is present in the case we only filter for the ‘magnificent seven’.)

Let me know your thoughts

ustonapc · February 23, 2024, 9:31am

Victor,

The factors that are driving the market are very different now than it was 15+ years ago and I have basically given up on using a long backtest to support the validity of a stock strategy (ETFs are different).

For instance, Nvidia is just a mid-cap stock and there are almost no investment in the field of A.I. 5-10 years ago. Personally, I rarely look at the results of a backtest further than 5 years for stocks.

If you feel that there maybe a look ahead bias issue by chossing only 7 stocks, it maybe possible to put in a market cap above USD 500 billion instead.

This is the result of a 5 year backtest on the same screen with a USD 500 billion market cap but without specifying the 7 names.

Regards
James

Jrinne · February 23, 2024, 11:38am

James,

So if I understand from our email conversation when you did this “If you feel that there maybe a look ahead bias issue by choosing only 7 stocks, it maybe possible to put in a market cap above USD 500 billion instead.” you did this 5 years ago to before the backtest is done.

If I understood correctly, this certainly does reduce the look-ahead bias. I am not sure that it removed it entirely as this is intended to select the 7 stocks with knowledge that we have today but it is enough to make one want to look at this this generalized rule over multiple periods, I think. Perhaps with random numbers around 7 as the idea that 7 stocks is a good number may have been popularized just recently…**

Edit: @ustonapc you have focus on just 2 stocks in your screen, I understand now.

It is a rational theory, i think, to say that in this market environment large companies are able to consolidate control of a market growing and even getting some monopoly (or oligopoly) power by buying smaller companies, economies of scale, influencing regulations, having better patient attorneys (e.g., ChatGPT funded by Microsoft may be able to withstand the lawsuit claiming they should not have been reading the New York Times without paying the New York Times), or making their product interface well with other large products (Apple watches work better with iPhones).

Personally I would test the idea prospectively in different markets and different time-periods. Perhaps, including 2000 to see if the larger dot-com companies are the onset that did well going forward (was Amazon.com big then, and did some large companies fail? IDK)…

My $0.02.

Best,

Jim

WalterW · February 23, 2024, 12:37pm

Hi James,
Interesting use of Close(CTR-1), but isn’t that literally look-ahead? What’s the result of comparing Close(CTR) to Close(CTR+2)?

Here’s a sim that uses a similar look-ahead buy rule;

ustonapc · February 23, 2024, 12:43pm

Walter,

The use of Close(CTR-1) is inspired by another user post (sevensisters - Florence if I remember correcly) in the forum. (It has been working out of sample for me in different systems so far which is why I am keeping it)

I have not tried Close(CTR) to Close(CTR+2), pls let me know your findings.

Regards
James

WalterW · February 23, 2024, 1:05pm

Hi James,
It may be working, but I would disregard the simulated equity curve. Comparing yesterday’s close to tomorrow’s close is difficult to do in practice.

ustonapc · February 23, 2024, 1:12pm

Walter,

I think the formula (original from sevensisters) is structured in a way that using mutiple days (instead of relying on a single day) e.g. 7>3, 5>2 which explains why it works out of sample although there is still some look ahead bias.

Regards
James

WalterW · February 23, 2024, 1:26pm

The real test is to remove the look-ahead bias. That’s a quick edit.

ustonapc · February 23, 2024, 5:09pm

Dear all,

Nvidia share price just cross $800 and surpasses $2 trillion in market cap during intraday trading.

It will be interesting to see if it can stay above these levels at market close before the end of the week.

Regards
James

WalterW · February 23, 2024, 6:12pm

NVDA’s mean next year EPS* is $29. With a share price of $800, that’s a forward PE of 27.5. The trailing 1 yr EPS growth is over 500%. NVDA looks like a good value!

*from snapshot page, However, I’m having trouble reconciling that with p123 estimate functions. Buyer beware.