Over the past week, we’ve had a front-row seat to Wall Street’s “react first and think later” mentality – and it’s taken a toll on technology stocks overall and artificial intelligence stocks in particular.
First, President Trump’s announcement about Stargate last Wednesday triggered a lot of excitement about the future of AI in the U.S. In fact, the tech-heavy NASDAQ soared 2.4% higher on Wednesday in the wake of the announcement.
We covered the details of Stargate in a Market 360 last week.
But yesterday’s sharp pullback in technology stocks wiped out those gains – and then some. NVIDIA Corporation (NVDA), for its part, lost 17% yesterday. That’s a staggering $589 billion loss in market cap – the largest value destruction for a single stock in market history.
The reason? A Chinese AI company called DeepSeek.
So, in today’s Market 360, I want to touch on what happened with DeepSeek and explain why it has the tech world in such a tizzy. Then, I’ll give you my two cents on all of this and share how investors should prepare for what’s next.
What Is DeepSeek – and Why Does it Matter?
On January 20, a Chinese AI company, DeepSeek released a new AI model called R1 that it claims is nearly on par with the U.S.’s AI models. In fact, DeepSeek claims the R1 model has a comparable performance to OpenAI’s o1-mini model for ChatGPT. What’s more, it utilizes inferior NVIDIA chips, called H800s, that are compliant with the export bans that were put in place by the Biden administration.
Oh, and the company claims it was built in about two months and cost just $5.6 million to train.
As a result, DeepSeek is now the number-one app downloaded from the Apple Inc. (AAPL) App Store.
Now, the tech world began to put all of this together late last week. And it rocked Wall Street as well as hammered AI stocks yesterday.
The fact is the entire roadmap of how AI is supposed to evolve is being questioned. As a result, some investors are wondering whether the U.S. tech sector will win the AI race.
And there are two primary reasons why…
- If DeepSeek’s AI models perform similarly to American AI models at a fraction of the cost, then it means American programmers and engineers missed something. And it throws into question the U.S. lead on AI development.
- DeepSeek has stated that it uses far fewer chips for AI training. A report released in December showed it only used a cluster of more than 2,000 NVIDIA chips to train its V3 model, while five times that number is typically used for similar AI training models.
Both raised concerns that NVIDIA was overvalued and triggered questions about the necessity for more data centers and an electricity grid buildout if AI doesn’t take as much computing power as originally anticipated.
Personally, I think these concerns are a bit overblown. Let me explain…
There’s Still a Lot We Don’t Know…
We are still in the dark with a lot of this, folks.
First of all, these claims need to be verified.
Second, as I discussed in a Special Market Podcast with my Growth Investor subscribers, folks are failing to consider the nature of machine learning.
Yes, AI can be as simple as optimizing various packets of data – which it sounds like that’s what Deep Seek is focused on – or it can be much more complicated.
I suspect DeepSeek is simply not using non-correlated data to do their AI.
You see, you have to be careful when you build a model: You don’t want to put things in it that might not have any impact on the model whatsoever. You want to put restrictions and rules in place – and you definitely don’t want to feed it irrelevant data.
I suspect that’s all DeepSeek is doing.
But this suggests the possibility that the other big AI firms weren’t doing this. And that could be a problem if it is true. They might want to be more focused when they build their prediction models.
What This Means for the AI Buildout
Now, longtime readers may know that NVIDIA is one of the longest holdings in my Growth Investor service. We’re currently up by more than 3,000% on the stock – and I have consistently said that it is poised to dominate the next decade.
So, what does this mean for NVIDIA?
Remember, NVIDIA supplies the crucial hardware and frameworks for large-scale deep learning, including generative AI. But a lot of folks are questioning the case for NVIDIA moving forward. They argue that if we can do advanced AI computing for cheaper, then that will hurt the demand for NVIDIA’s chips.
I couldn’t disagree more. Now, whether it means NVIDIA needs to lower their prices remains to be seen. But, if anything, this means that more folks can build AI models, not less. And that’s good news for NVIDIA.
In fact, NVIDIA seems to agree. They responded to the DeepSeek news, saying it is “an excellent AI advancement” and that “DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely-available models and compute that is fully export control compliant.”
This suggests the possibility that DeepSeek “leveraged” something that was already there, but also that it will lead to more demand for its chips. Investors seem to be thinking more clearly again. NVIDIA was up roughly 9% on Tuesday, and many of the infrastructure-related plays are up as well.
Now let’s address the infrastructure plays.
Right now, AI is controlling an increasing share of the power grid. And the biggest victims of the DeepSeek news were companies involved in building out the data centers and the power grid.
But the point remains that machine learning utilizes a lot of computing power – and it’s going to need more, even if DeepSeek has found a more efficient way to do it. So, you will still need all the AI hardware and software and need to double the utility grid and build more data centers.
In my opinion, the notion that DeepSeek is going to reduce AI electricity demand is highly improbable. Moreover, the idea that methodical utility companies are going to abandon their plans to double grid capacity because of a Chinese software app is, frankly, ludicrous.
So, I view the pullback in these infrastructure/grid plays as a great buying opportunity.
The Trump Factor – and How to Respond
Folks, in my opinion, this is way overblown. I am highly confident that even if DeepSeek has a temporary AI advantage, its success will not persist, because few developments are more motivating to Silicon Valley than a wake-up call.
So, in case there’s any question, under no circumstances, is NVIDIA about to lose its market dominance due to a Chinese software app. In fact, the deep dive many of these stocks experienced on Monday gives you a chance to load up on NVIDIA and the power-grid stocks.
A new AI economic war is breaking out, and I would not recommend that any investors sell NVIDIA and the companies helping to expand the utility grid due to possible Chinese propaganda.
Still, President Donald Trump called this news a “wake-up call” for U.S. companies to reassert their dominance over AI. Silicon Valley is expected to win the AI wars because it will enjoy the level of support that Chinese companies enjoy from the Chinese Communist Party, but within the context of capitalism.
Bottom line: President Trump has been very vocal about the U.S. dominating AI, and I don’t foresee him stepping aside to let a Chinese app derail the U.S. tech industry.
I have been telling my Growth Investor subscribers for months to prepare for Trump 2.0. I’ve said that as soon as Trump takes office, the U.S. economy will begin to hit the overdrive button.
I’ve prepared everything you need to know in a special presentation. Click here for all the details now.
Sincerely,
Louis Navellier
Editor, Market 360
The Editor hereby discloses that as of the date of this email, the Editor, directly or indirectly, owns the following securities that are the subject of the commentary, analysis, opinions, advice, or recommendations in, or which are otherwise mentioned in, the essay set forth below:
NVIDIA Corporation (NVDA)
