NVDA近期基本面强劲（垄断定价、订单有保障、现金流充裕），但面临硬件周期性、循环融资和高估值的长期风险；质疑当前增长是否可持续，或是否会像过去的芯片周期一样下滑。

If you own shares of Nvidia, AMD, or Palantir, you are told to hold on, buy on the dip, and strengthen your positions because a multi-billion dollar wave is coming.

Next is Nvidia stock . Unlike Palantir, this is not a debate about whether the business activity is real or not . Everyone knows that this business is real. The question is, what are you paying for it?

The first argument for the rise is that they currently have monopoly pricing power . Nvidia no longer just sells graphics processing units. They sell the complete package that everyone needs. Central processing units, networks, software, all of that. Their revenues per gigawatt of data center power have increased from $18 billion with Hooper to $25 billion with Blackwell, and to $40 billion with Vera Rubin. So, instead of lowering prices with the competition, Nvidia actually raises prices in every cycle.

The second argument for the rise : the order is guaranteed. Nvidia generated $96 billion in revenue last quarter, a 106% year-over-year increase. They projected revenues of $108 billion for the next quarter. The administration stated that they face supply constraints until 2028. This means that demand is so high that they cannot manufacture chips quickly enough . Their revenue stream is essentially guaranteed for the next eighteen months.

The third argument for the rise, and this is what I like to hear, is the cash flow machine. Nvidia generated $ 21 billion in free cash flow in a single quarter. It has repurchased $26 billion worth of shares and still has $100 billion remaining in its buyback program . They print money like a software company and use it to reduce the number of shares, which raises earnings per share even if the share price is fixed. Guys, remember, as an investor, I don't want them to buy overpriced stocks, but we'll get to that later.

And one more thing about Nvidia before we move on to the negative arguments. This news just came out, literally. Nvidia acquires "Hugging Face" for $13 billion . The name "Hugging Face " sounds like the kind of name Bill Belichick might come up with for any social media company. If you don't know what " Hugging Face" is, it is basically the largest open platform in the field of artificial intelligence. Three million AI models, 500,000 datasets, and more than 18 million developers use it. It is where the world of open- source artificial intelligence lives. Now, here's why that's important. Hacking Face had already turned down a $500 million investment from Nvidia last year because they did not want one company to have too much control. And now, Nvidia is buying the company outright. Jensen Huang said it's about deploying open AI models in factories, hospitals, farms, classrooms, and everywhere we can reach. But let's be realistic about what this also does. It gives Nvidia control of one of the largest distribution channels in artificial intelligence. More open models mean more people need chips to run them. And on whose chips will they operate? This is Nvidia's largest acquisition ever, roughly double what it paid for Mellanox in 2020.

May I appeal to the CEOs of these tech companies ? Stop trying to be too impressive in your choice of names. They just gave the thing a name like "Joe". Or the name " Frank". These are the names of our software. It's unbelievable to me. " Hugging face". What a really stupid name, "Hugging Face"? Oh my God, that deal is what turned Nvidia from a chip company into a data center company . This deal could be just as important.

The first negative argument is that devices are cyclical in nature, almost always. No hardware company has ever escaped this . Once Microsoft, Meta, Google and Amazon have finished building their AI data centers , they move from build mode to maintenance mode. These chips last for years. Once the world’s largest data centers are filled with Blackwell and Robin chips, Nvidia’s revenues will not only stabilize, but could plummet.

The second negative situation is the problem of circular financing. Nvidia uses its balance sheet to fund startups, support leases, and guarantee loans, so that those companies can buy more Nvidia chips. Think about it. Nvidia literally funds its customers to buy its products. This shows up as revenue on the income statement, but if those startups fail to generate profits, the whole cycle will collapse. This is an artificial request.

Here's the third negative scenario, the $5.4 trillion problem. Nvidia is a $5.4 trillion company and is Now, guys, Tim just asked a great question . He says, what's the difference between Nvidia doing this and a Kohl's credit card or a salesperson saying, " Hey, you can pay me later." That's a great question, and I'm sure many people are asking it.

The difference lies in this. Many of these companies that sell chips to startups either have no revenue or don't even make a profit. If those companies do not make a profit, who will ultimately bear the loss ? Suddenly, Nvidia will have to amend its financial statements to say: "Hey, do you remember this revenue we recorded previously? Instead of being 10 billion, it will now be 3 billion because that's just what we collected." This is the problem at hand.

So, guys, you've seen both sides of Nvidia, the positive and the negative. And here's the honest question. Can you do it yourself? Can you open Nvidia's data now, look at the actual financial figures, and decide whether you're paying a fair price or a stupid price? Can you take a $5.4 trillion company that's trading at 40 times sales and decide whether it will give you the return you need, or are you just relying on wishful thinking? Because that's the difference between investing and guessing.

Guys, think about this. What is the cost of buying Nvidia shares at the wrong price? Buying Nvidia or Palantir and paying an exorbitant price because a news headline excited you, and because everyone is shouting that these companies are growing like crazy. But guys, one bad decision about one stock could cost you thousands and thousands of dollars.

Okay guys, let 's review "Nvidia" here. Again, it's a company worth five and a half trillion dollars. This is the price. The company is valued at 5.5 trillion. So it's not as good a " Balantir" as it is, but its debt is still very low for a company that generated $127 billion in free cash flow last year. And guys, $193 billion in net income. This company literally prints money. It's as if nothing in this company is bad in terms of numbers. In terms of what it has achieved, returns of 73% on capital, and 55% for the past five years.

Here are the scary parts. The free cash flow rate and price-to- earnings ratio are only 28. So, I know that many people look at this and say it's great, and I think it really is great. But again, we have to ask the question: Will their revenues and profits be permanent from now on , or will it be like the rest of the chip world and its history, cyclical and then eventually declining?

Now , this company is much cleaner than Palantir. The only "X" marks we have here are the five-year free cash flow rate , and in their defense, their five-year free cash flow is $52 billion, while their one-year free cash flow is $127 billion. So, it has risen considerably . So, I'm focusing more on this one year. The same applies to net income, from 70 billion in the past five years to 193 billion.

Hey guys, can I show you how this company has grown? Because I think it's really important to understand that. These are the annual revenues for the past ten years. They generated 7 billion in revenue. In 2017, their total revenue for the entire year was 7 billion. They made $300 billion in the last year recently. What does that mean, guys ? They are now achieving in almost one week what they achieved throughout the whole of 2017. They are achieving it almost in one week . Isn't this amazing? As for net income, they achieve a larger net income , 193 billion. They achieved 1.67 billion. This is 150 times higher than it was in 2017. Absolutely unbelievable. And I ask the question again , is this sustainable?

Okay, let's see what the analysts think. Well, folks, analysts are predicting that earnings will rise from $4 per share to $20 per share, a fivefold increase over the next four or five years. As for revenues, they range from 213 billion to reach one trillion dollars. Again, a 5-fold increase over the coming years? Absolute madness. Absolute madness.

Okay guys, we're now in front of our stock analysis tool. What is the value of the company based on my assumptions? Once again , I want to reiterate that people should remember that I am looking at this from a 10-year perspective, including what I consider to be a decline as the cycle progresses. I still think they are bold projections, but I have set revenue growth rates of 12, 20 and 30% for the next ten years. I set a profit margin of 35, 45 and 55%, because as competition enters , and as chips enter a certain cycle, their profit margins will decrease. This may actually be high . I don't know. Next, the earnings multiple 10 years from now. I placed 18, 23 and 28. Again, my return is 9.5 percent . I pressed the analyze button. Boom. Guys, this is the amazing thing. I have a low price at 140, a high price at 1230, and an average price at 400. Look how wide these ranges are, and based on my average assumption, it looks like this could be a buying opportunity here. The question is, do you think these average assumptions will be their rates for the next ten years? This is the question we should ask ourselves about any investment we make, even when there are so many positive indicators as we see here.

nor is it a king like Nvidia. Unlike Nvidia and Palantir, AMD stock is trading at a much more reasonable forward price-to-earnings ratio . so they don't become entirely dependent on Nvidia.

Nvidia operates with a gross profit margin of 75%. This means that from each segment they sell, after paying all the direct costs associated with it, 75% of it goes as profits to cover general expenses and taxes.

AMD is trying to fight Nvidia at the top of the AI market,