The battle for AI supremacy took a wild twist. Ever since Chinese AI startup DeepSeek unveiled its new R1 model – which rivals the best U.S. bots like ChatGPT – AI stocks have been on a wild roller coaster ride.
In fact, the news wiped out $600 billion of market cap in “king” chipmaker Nvidia – the biggest single-day drop in market history.
The AI breakthrough also sparked fears that US companies will pull back on their extreme AI spending.
So, the questions I’m going to tackle today are…
What does DeepSeek’s breakthrough mean for the industry as a whole? And does it mark the end of the big bull run in AI stocks?
Let’s get into it…
What’s so special about DeepSeek?
On January 20, DeepSeek released a new AI model called R1 that it claims boasts a comparable performance to OpenAI’s o1-mini model for ChatGPT. And it’s achieved this using inferior NVIDIA chips, called H800s.
And this prospect of more efficient AI models that don’t need Nvidia’s most advanced chips was enough to obliterate its share price.
What’s more, DeepSeek claimed its AI model was built in about two months and cost just $5.6 million to train. My first instinct when I heard the costs was not to believe it. After all, you can’t trust everything that comes out of China.
Nevertheless, whatever the costs were, DeepSeek has found ways to train and operate AI models using less electricity and less computing power than
“Made in America” AI.
And it could prove to be a game-changer!
Researchers, developers, and smaller start-ups who – unlike Microsoft or Amazon – don’t have tens of billions of dollars to throw at AI development, can now harness cheap AI to boost their productivity.
What does it mean for American AI stocks?
As the costs to build and train AI models drops, more players will start creating more and more AI models.
Sure, some may use less sophisticated chips from Nvidia and its rivals. But high-end chips will still be in demand from those who want the best performance possible. Either way, demand for chips will only grow!
Additionally, if AI does get cheaper, then all the US AI companies will simply be able to ramp up the output of their AI on the same hardware they have now and exponentially scale it up. Lower costs mean less expenses and more profits and cashflows for the mega tech giants and shareholders.
Lastly, machine learning uses 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 as well as more data centres.
Simply put – the fears of DeepSeek’s disruption to American AI players are overblown. So, I’m doubling down and scooping up AI innovators at cheaper prices.
One thing is clear: The US won’t sit idly by and let China dominate AI supremacy. So, expect the AI war to heat up even more in 2025.
Not a subscriber to Money Morning?
You can get free daily recommendations like these with Money Morning eletter. Just sign up here.
