Walter,
Nvidia’s Forecast 12m forward PEG is 2.43.
You can see all the valuation ratios at this Nasdaq page.
Regards
James
1 Like
P123 has NVDA long-term PEG at 1.04.
1 Like
Walter,
Given the signficant difference, I will trust the PEG published by Nasdaq which is more conservative and less a bargain buy.
Perhaps Marco can take a look into this and confirm that P123 PEG is correct.
Regards
James
2 Likes
Dear all,
Nvidia just put an intraday all-time high and overtake Saudi Aramco as the 3rd most valuable company globally. (after Microsoft and Apple).
Here is the latest snapshot of the Top 10 asset by market capitalization.
Regards
James
1 Like
It’s an interesting feedback loop. Google, Microsoft, Meta, etc. take the cash generation off their existing businesses and announced they’re going to pour it into H100s. The public gets really excited about the future of AI and starts bidding them up. Then NVDA reports blowout revenue off all these announced H100 orders, and the public gets even more excited about AI and starts bidding up Google, Microsoft, Meta once again before they’ve demonstrated that generative AI is even a good business for them.
No doubt building datacenters is going to be amazingly profitable for the next few years. Im onboard with AGI is going to be a big thing, but I’m less sure it’s necessarily going to be a good cash generating business model. We’ve seen with things like streaming and electric vehicles that the future turns out to be capital ex intensive race to the bottom for most players, and often worse than their current business models. At some point … 2025? 2026? The non chipmaking Mag7 will have to show their hands and demonstrate that all this capex is getting translated into real earnings. And what if AI actually eats up their current cashcow businesses that allow them to pour capex into GPUs and datacenters in the first place? If it kills advertising based search and social media models, I’m all for it.
1 Like
Dear all,
This in an interesting table comparing all the AIs currently available on the net.
Regards
James
AI is now smarter than the average human
1 Like
Different formulas. They use next 12 months EPS, we use NextYEPS for short term growth (so up to 2 years out) and it’s adjusted by the current quarter. This is the formula:
@y_till_ny:(Eval(QtrComplete=4, 2 , 2 - QtrComplete * 0.25))
@pegst:(PEExclXorTTM / gr%(NextFYEPSMean ,EPSExclXorTTM, @y_till_ny))
PEG is notorious for many different version. I much rather look at a chart. NVDA’s PEG is sitting close to the 15 year low PEG. Last time PEG was this cheap they were selling boatloads of GPU for crypto mining. And btw, how lucky are these guys? Two use cases that just fell in their lap that had nothing to do with their original business.
Either way, PEGST of 2 seems like a good time to sell, or $1,372 (the high Price Target is $1400 
)
2 Likes
Nvidia conference could trigger rally in AI coins
Cryptos | 03/14/2024 09:52:46 GMT
- Nvidia’s recent earnings releases have kick-started the AI narrative in the crypto market, recording huge gains for AI tokens.
- OpenAI's SORA announcement pushed WLD even higher.
- WLD may record more gains if OpenAI announces ChatGPT 4.5, ChatGPT 5, or Sora 1.5 around Nvidia's GTC 2024.
Nvidia will hold its first in-person GPU Technology Conference (GTC) since 2019 from March 18 to 21. The semiconductor firm's earnings have been a crucial factor propelling the price of AI-related tokens, and the possibility of any announcements during the upcoming event has the potential to lead to another bull run.
Nvidia earnings kick-start AI narrative
AI tokens were some of the major beneficiaries of Nvidia's Q4 impressive earnings report released in February. The GPU giant reported a 769% increase in net income, sending revenue to more than $22 billion and beating analysts’ expectations. This comes from the increasing global demand for AI technologies, which has boosted sales for its chips by wide margins.
While the AI tokens couldn't directly benefit from these earnings, Nvidia's report boosted investors' confidence in their viability. Their 30-day price change shows huge gains of around 200%.
[
](https://editorial.fxstreet.com/miscelaneous/Screenshot (6)-638460064880630754.png)
AI Cryptos Category
With Nvidia's GTC 2024 around the corner – March 18 through March 21 – investors are hoping for a similar impact. Considering this is the largest AI conference of the year, many traders expect it to be a catalyst for AI tokens.
The conference will begin with a keynote speech from Nvidia's CEO Jensen Huang, who will discuss the company's latest innovations and future plans. Expectations are that AI tokens with strong fundamentals may benefit largely from the announcements and workshops at the conference.
In anticipation of the conference, protocols like NEAR Protocol (NEAR), The Graph (GRT), and Worldcoin (WLD), among other AI tokens, have already posted large returns of 36.1%, 33.8%, and 29.2% on a week-on-week basis respectively before experiencing a slight correction.
Also read: Week Ahead: Crypto markets set to tip in favor of gaming and AI tokens
OpenAI's SORA announcement pushed WLD even higher
The importance of such events driving investor interest can be seen in the surge of Worldcoin (WLD) after Open AI unveiled Sora's text-to-video AI tool in February. The blockchain-based digital identity project spearheaded by Open AI CEO Sam Altman has risen by about 250% since Sora's debut.
WLD, trading around $3 on February 15, skyrocketed in price a few days after the announcement. It has since broken into the top 90 cryptocurrencies, accumulating a market capitalization of over $1.4 billion as of the time of writing.
What's next for WLD? ChatGPT 5 or Sora 1.5?
Worldcoin is looking set for more growth after struggling in certain regions due to privacy concerns. With Sam Altman back on the OpenAI’s board, the firm’s products could see faster developments and quicker launches.
One potential event that may boost WLD's price is if Open AI accelerates GPT 4.5, GPT 5 or Sora 1.5 development to launch simultaneously with Nvidia's GTC 2024. A potential launch of any of these products may see WLD's price even raking in more gains than it had seen before and potentially break into the top 50 cryptocurrencies.
Investors should conduct their own research before taking positions in any of these AI tokens, as crypto exchange Coinbase has cautioned that some of their increase may be fueled by mere speculation.
1 Like
Dear all,
Jensen Huang from Nvidia is now the 17th richest person in the world with $83.1 billlion. CZ from Binance just adds $8.42 billion today and ranks 26th with $48.3 billion according to the Bloomberg Billionaire Index.
Regards
James
1 Like
Dear all,
Analysts are forecasting Nvidia's annual revenue will jump to $131 billion by 2026, more than double the $61 billion recorded in the fiscal year that just ended.
Regards
James
2 Likes
Just don't touch anything like that
ZGWZ,
I get your general point and agree 100%. But for some lame reason my port bought NVDA and your post made me curious as to how it turned out (I had not bothered to look until now). I understand there is no greater point to this one-time anecdotal example. But I did not over-ride my port's recommendation and I was just curious:
No fees because the transaction is edited to the real price with zero commission at Fidelity.
Alert: Publication bias. I.e., a negative result (or loss) might now have been posted (published).Or maybe I would have to show my general agreement. But I truly get there is no broader point here unless it is to say I never override my port's recommendations. 
Not sure what was happening 6/9/24 to trigger a buy. It did not seem to persist whatever it was. This is the only NVDA transaction for this port.
Jim
Actually, I disagree. Looking at companies and eyeballing their financials or their other characteristics can be a great way to come up with new creative ideas to build new factors.
Basically, by zoning in on companies that have been doing well over an extended timeframe - or more specifically, by zoning in on the underlying characteristics of their success, one can come up with new potential strategies that can be applied in general to other companies (e.g. by incorporating those characteristics in a new factor). These new generalized ideas can then be backtested and checked for robustness.
Of course, one might come to the conclusion that nothing is to be found (i.e. the company's price has been elevating based on structural speculation and not based on the underlying business dynamics). Or one might come to the conclusion that something is indeed to be found, but which later (after a solid backtest) turns out to not work in general. That doesn't mean it is not a useful tool to come up with new ideas though!
Best,
Victor
I would use a systematic search to try to discover them, rather than individual attempts. But with dan's sheet, I don't even need to do that at the moment. His factors are very good.
And, studies have shown this to be the inefficient way to go. Machine learning models that learn from stocks that have historically risen a lot don't have a high return-to-risk ratio, whereas models learning from stocks that have historically fallen much more identifies common factor patterns. A better way is to study poor performing stocks and then find out if their contraries might be useful.
Edit: In other words, instead of learning from the successes of NVDA, we should learn from the failures like FFIE, AMC, BA, INTC, MAXN, SMCI, AILE, IEP and others.
1 Like
This is a Bayesian thing—at least in part. As an analogy, a test for breast cancer may detect 100% of breast cancer cases. But if the test is positive 50% of the time when the patient does not have breast cancer the test is completely useless for screening. Although it could be used as a rule-out. If the test is negative you could be sure that there is no breast cancer.
Your method looks at both the positive and negative results when screening for good stocks.
For Bayesian methods specifically, one would look at the probability that a feature is present given that a company is successful and divide that by the probably that a feature is present given that a company is unsuccessful. This his called the likelihood ratio or the Bayes Factor.
For a company successful just because it is lucky many features will have a likelihood ratio of 1. I.e, the feature had nothing to do with the success of the company. This is only evident if you look at the entire likelihood ratio including the probability that the feature is present given that the company has done poorly.
So your prior belief about how successful a company is likely to be gets multiplied by a Bayes Factor of 1 with no change in your prior belief.
While not a priority for me personally, P123 could consider adding Naive Bayes as an ML model in the future when P123 moves to classification methods.
Naive Bayes could be particularly effective when used as a red-flag in buy/sell rules. For instance, it could serve as a machine learning refinement of the Piotroski score. This approach would leverage Naive Bayes' ability to classify based on multiple categorical inputs, potentially enhancing the traditional Piotroski score. The probabilistic output of Naive Bayes could provide a more nuanced risk assessments, flagging stocks that exhibit characteristics historically associated with poor performance.
With features standardized using z-score one could also use continuous inputs (unlike the Piotroski example above) with Gaussian Naive Bayes
Summarising from positive examples is less effective for the same reason as stated earlier: stocks that go up big can be winning lottery tickets (like GME and TSLA) or they can be really good stocks. However, stocks that go down big generally do have serious problems and are therefore better suited to get useful results.
For example, it is more useful to draw lessons from the failures of SGMA than from the successes of NVDA.
Of course, I'm actually not prepared to analyse any individual stocks at all. If one is going to search for factors, one should do it systematically like dan's list (or more so). Otherwise that list is good enough.
1 Like
I can only agree with that. My point was that one can learn a lot by looking at individual companies to come up with new ideas. That of course also goes (or maybe even more so) for companies that have been doing particularly bad to find general characteristics that can again be backtested such as to avoid picking such companies within your own holdings in the future.