NVIDIA Corporation (NVDA-Q)

Fantastic earnings growth. Just bounced off 200-day MA (a good support level), and that's where he added recently. Paying about 32x forward PE for 35% expected growth, so the PEG ratio is reasonable. 

DeepSeek news concerned some investors, but does it make the products that NVDA does with the same broad customer base? Have to see over coming quarters and years.

His team's fundamental analyst likes it a lot, with a strong buy. Bottomed in 2023, then an uptrend, now has been basing over the last couple of months. With Monday's DeepSeek volatility, it tested its key 40-week MA. It's chopping around that level now.

If it remains above that level, the medium-term trend is up. You can nibble here, but he'd prefer it to come back up and take out recent highs and on strength. If we get a multi-week close below that level, strongly suggests a bigger corrective phase. Off the top of his head, the next support level is around $90 or $94.

Investing is tough, and when a group of university students came visiting his shop, and they all put their hypothetical $$ to work in NVDA, that has him very concerned.

There are two scenarios after DeepSeek rocked the tech world earlier this week. One is that DeepSeek requires fewer computer chips to do the job, therefore eroding demand for NVDA's chips and decreasing their revenues. So, NVDA shares are overpriced and will decline. The other story is that maybe we're not getting the full story, that maybe the cost of developing DeepSeek was a lot more than publicized, and may been subsidized by the Chinese government. Independent research company, SemiAnalysis, questions the official cost and pegs it as much higher. If the facts are not true, selling NVDA this week was a mistake. Consider that Meta's Zuckerberg, Elon Musk and Oracle are paying full price for their huge orders for NVDA's chips and that they must have done their due diligence before ordering those chips. They must have known about DeepSeek already. He thinks the Semianalysis story is spot on.

There are two scenarios after DeepSeek rocked the tech world earlier this week. One is that DeepSeek requires fewer computer chips to do the job, therefore eroding demand for NVDA's chips and decreasing their revenues. So, NVDA shares are overpriced and will decline. The other story is that maybe we're not getting the full story, that maybe the cost of developing DeepSeek was a lot more than publicized, and may been subsidized by the Chinese government. Independent research company, SemiAnalysis, questions the official cost. If the facts are not true, selling NVDA this week was a mistake. Consider that Meta's Zuckerberg, Elon Musk and Oracle are paying full price for their huge orders for NVDA's chips and that they must have done their due diligence before ordering those chips. They must have known about DeepSeek already. He thinks the Semianalysis story is spot on.

(Stock split June 10, 2024) Still in the top 10 in his fund. Imagines that the Blackwell chips may take a bit of a back seat to the Hopper chips. His 12-month target is $175. Some Wall Street analysts have brought their targets down, but none are below $155, so still a very decent runway.
 
Still has a monopoly on the accelerator chips. Still a must-have in a portfolio. Companies involved in the large-language models, like META and GOOG, will continue to buy the Blackwell chips.

There are a lot of unknowns here, after the DeepSeek revelation Monday. He fears that chip order cancellations are trying to be priced into the stock, but doesn't know if it's finished. When it is, the stock is a buy. He needs more information. Too unclear.

NVDA tanked 17% on news of China's DeepSeek stealing the AI crown from ChatGPT, faster and cheaper. All AI-related stocks, including energy plummeted as the Nasdaq slid over 3%. Buy this weakness or sell? Honestly, he doesn't know. It's confusing. Analysts haven't had time to digest the DeepSeek news. How many NVDA chips does DeepSeek need to work?  Will NVDA's numbers come down if customers freeze and reassess their chip orders? Will things turn into the internet pause of 2002, which turned out to be a collapse? Or not? He has no view on DeepSeek, because he simply doesn't know enough. He sold some shares before today's news when shares were much higher to right-size his portfolio. However, the hardest thing is to wait until you know more instead of taking a knee-jerk action. 

Difficult to know exactly what valuation will be in a few years. Revenue growth has been strong, but the P/W multiples is also sky high. Would be a good stock to hold for the long term. Could see volatility in the next 1-2 years as growth plateaus. 

Probably the most crowded stock in the world, and people have made massive fortunes. Still hasn't reached target price. Huge beat with revenue up 97% YOY, thirst for chips. Trump's $500B AI announcement probably extends demand beyond 2028-29.

The danger with this stock is not owning it. A 27% growth rate, trading at 31x -- not as compelling as it was, but still has a PEG of almost 1. Try to buy lower, but better to buy now at $145 than not own at all.

Yes. Stock's rallied nicely, but backstopped by tremendous growth in sales and earnings. Most is due to its far-and-away leadership position on AI chips, which is still in early innings. Not inexpensive, but the sector has a huge runway for the next 3-5 years.

The Fed heavily influences tech stocks and this sector may underperform. Also, this trades at 55x forward PE, rich. The chart shows an uptrend, though. You don't want to see it fall through its last peak from mid-2024. He can't argue against this trend, but is cautious because NVDA has been too good. Currently, shares are consolidating, which is normal, but you don't want to see the chart break down. 

 A great, innovative company. They built the best mousetrap, but MSFT, Amazon, Meta and Google make up 40% of their revenues. So, those 4 companies must grow at the same pace, which he doesn't think they can. Ask: How much can their customers keep increasing spending? Also, those companies are developing their own chips, while AMD is nipping at their heels. Also, a problem with the new Blackwell chip is that not all data centres can't use them, because the chips may need liquid cooling, and not air cooling; not all data centres use liquid cooling. The valuation remains high. Revenues can be flat and its multiple may fall from 30x to 15x.

High beta, not for everybody. 52-week high followed by a 6% decline yesterday. But makes sense as part of a portfolio for a growth investor.

Denominator (PE multiple) keeps moving higher faster than the numerator (price). So it's actually fairly cheap. Looking at lots of free cashflow for 2025. Leader in AI and gaming technologies. Data centres, deep learning, breakthroughs in image and speech recognition. 

90% market share in the AI GPU segment. Earnings growth rate a staggering 50%, giving it a 0.7x PEG ratio, probably the cheapest mega-cap tech name out there. Yield is 0.03%.

Loves the company and CEO, but the valuation is not cheap (he's a value investor). Yes, NVDA is setting the industry standard in chips. Remember that hardware is, unlike software, a huge gross margin business. As in the past, no one company can maintain a hold on an industry, so eventually margins will come down. You must have a strong believe that Jensen Huang can maintain this lead and that data centres' build-out ca last long term.