Please believe me when I say I mean this in the best possible way…
I told you so.
I told you to be careful about Nvidia (NVDA). That its ever-rising share price was based on too much hype. And that you should never put all your faith and trust in a single stock.
Here’s what I said in the pages of The Wide Moat Letter back in November:
Nvidia’s sales are based on the $10 billion to $20 billion (per quarter) of [capital expenditures] that the hyperscalers are spending on its chips right now. That might not be sustainable long term. Right now, it’s unclear as to whether or not the Big Tech spenders are going to generate attractive [returns on investments] on their massive AI budgets. And if they don’t, their investors will force them to pull back on spending. If that happens, Nvidia’s sales are going to slump, resulting in significant multiple compression… and falling share prices.
Monday’s market action proved me right.
The Big Tech behemoth slid a whopping 17% on the news that it might just not be the grand master of the AI universe after all. Startup Chinese rival DeepSeek, it turns out, has its own AI product that apparently works amazingly…
Without Nvidia’s pricey chips…
On less energy with less data demands…
At less than 6% of the training cost.
If that’s all true, it’s no surprise that DeepSeek-R1 became the most-downloaded free app through Apple’s (AAPL) App store in the U.S. on Monday. That meant it dethroned the Nvidia-chip-run OpenAI.
All told, Nvidia lost $588.8 billion in market value on the news. That’s “billion” with a B.
Look, I’m not gloating. There’s no joy in seeing a stock get hit that hard knowing there are people behind it panicking.
But there are some very important lessons to learn here, if you didn’t already take them to heart.
We never want to fall in love with an investment – even one with Nvidia’s potential. A company might be awe-inspiring in its groundbreaking potential. Or it could have the most amazing history of safe and steady returns.
Either way is no guarantee of future results when, as we saw on Monday, anything can happen. That’s why we need to keep our heads on straight.
With that goal in mind, let’s consider two practices to keep your emotions in place while you invest.
Lesson No. 1: Don’t Believe Everything You Hear
For the past three years, we’ve been told both directly and indirectly that AI is the end-all, be-all of everything.
It’s shaping the present. It will reinvent the future. And there’s no going back, they all said.
To be fair, Monday’s market activity didn’t prove any of that wrong. But it did show there was a flaw somewhere in the story “they” were selling.
Nvidia, it turns out, isn’t invincible.
Yahoo Finance, one of the many financial news platforms that have happily hyped Nvidia all these months, noted in Tuesday’s Morning Brief how “Monday’s rout highlighted the fragility of the current bull market and its reliance on Big Tech’s big hopes – and that risks to it can come suddenly and without warning.”
So very, very true.
Now, it did go on to immediately assure that Big Tech is still on fire, and it has a point there, too. After all, despite Nvidia’s plunge, it’s still up over 2,000% in the last five years.
All the same, I think it missed a valuable opportunity to highlight the reality that fire isn’t something to play with.
That’s what Jason Thomas, head of global research at influential investment firm Carlyle, was driving at when he appeared on Opening Bid yesterday (another Yahoo publication).
You can make your projections and have your best estimates about what’s going to happen on the top line. You can make your assumptions about competition. But then you have the question of the valuation: This is an issue of what is the expected return here.
And, at this point at least, it seems like that expectation was too high.
Or as billionaire investor Ray Dalio said on the same podcast last week – before this drama went down – Big Tech companies like Nvidia were priced for absolute perfection. And that’s a dangerous place to be.
Lesson No. 2: Don’t Believe Everything You Hear
Same lesson. Different angle.
Just like you don’t want to invest on hype alone, you also don’t want to sell on drama alone. Neither practice tends to end very well.
As we all saw, Nvidia investors freaked out yesterday. So did data center-specific real estate investment trusts (“REITs”) like the ones I wrote about last week.
Digital Realty (DLR), for one, the largest data center in the world, dropped almost 12%. And it fell again yesterday.
Which, to some degree, makes sense. If there really is a less data-intensive chip out there than what Nvidia’s been pitching… then we don’t need as much room to store data. Which means we don’t need an explosion of new data centers.
Yet does the DeepSeek announcement mean these tech companies are dead in the water?
I already criticized that Yahoo Finance piece for being too optimistic. But now I have to call out the other side as well who are predicting that this is the end of America’s AI growth story.
The Chinese have beaten us, they say. Customers are all going to turn away from Nvidia and opt for DeepSeek products instead.
That’s it for the data-center bull run.
I’m not trying to mock this position. There are some very intelligent and respected people taking it, including Gary Wojtaszek, the former CEO of now private data center REIT CyrusOne. He’s now on the board of GDS, a Chinese data center, and he believes American AI is now on the decline.
“What [DeepSeek] did was impressive,” he told me on Tuesday. “They supposedly built something much cheaper than the competition. For billions of dollars less. It’s almost as good as ChatGPT” while using “older technology chips.”
And he’s right that, at face value, this is disruptive news. Yet I can’t help but think there’s more information to process before acting on it.
That’s why I kept digging.
Two Different Viewpoints on the AI Situation
Another very intelligent person I spoke with about the DeepSeek dilemma was someone you might know of: Nick Ward, one of my right-hand men here at Wide Moat Research. And he thinks Monday’s sell-off was:
… likely a big overreaction.
Maybe the Chinese firm figured out a way to train AI in a much more efficient way… but maybe not. (Do we really trust data that comes out of China?)
I wouldn’t be surprised if this was some sort of PsyOp by the People’s Republic of China in an attempt to scare off U.S. AI investment so that our gap doesn’t widen.
Ultimately, while I see the value in both perspectives, I have to say I side with Nick much more. As I said before, we need to look at available data closely before making any big decisions – both past and present.
Past information has repeatedly shown that China has no problem exporting lower-quality products. So who knows how well DeepSeek’s chips will work long term.
Yet, believe it or not, the fact that they exist at all is excellent for a company like Nvidia. It’s been flying high for far too long and, frankly, it needs the reality check that competition provides.
Nvidia will now have to get more innovative and consumer conscience – including about pricing – than ever before in order to stay relevant. A little bit of humility can go a long way.
The same applies to data centers. On the one hand, I’m quite confident they’re still going to be necessary. But it’s about time their stock prices had a reason to fall back to Earth and their management a reason to remember that anything can happen.
It’s that kind of mindset that keeps both companies and investors alike growing over the long term… not just for a few years.
Regards,
Brad Thomas
Editor, Wide Moat Daily
P.S. If you did read last week’s article about data centers, please note I’m not ready to recommend buying new shares just yet. As I stressed above, I want to properly evaluate their current numbers before I give any “all clear” signals here. Caution might cost us a few dollars here or there, but what we save in the end more than makes up for the “loss.”
MAILBAG
What practices do you try to follow to keep your emotions in place during investments? Write us at feedback@widemoatresearch.com.