A new Chinese AI model called DeepSeek is getting a lot of attention.
Some say it might collect personal data, and others believe it could change how we use AI.
One big talk point is that DeepSeek uses a special method called “chain-of-thought” to answer questions. This means the AI works step-by-step, much like how we solve a math problem.
Last week, investors sold off AI stocks in the belief that DeepSeek will be cheaper to run than current models.
Stocks of AI pick and shovel plays such as Arista Networks, Vistra Energy and Vertiv Holdings sold off sharply.
I didn’t buy the hype.
While early tests show that while DeepSeek may save energy during training, it uses more energy when answering questions.
For example, one test found that a long answer from DeepSeek used 41% more energy than a similar model from Meta.
When examining many questions, DeepSeek used nearly 87% more energy because it often provides much longer responses.
This extra energy use comes from the way DeepSeek thinks through questions.
So DeepSeek isn’t the lowest cost provider that Mr. Market assumed it would be.
The rise of chain-of-thought models like DeepSeek might lead to more advanced AI, but companies will also need to manage higher energy costs.
This is a key trend to watch as AI continues to grow and change. And here’s why…
How DeepSeek Is Changing the Game
Despite Mr. Market’s gloom last week, DeepSeek will accelerate the adoption of AI.
That’s why I continue to be bullish on AI stocks.
(Note: For my AI recommendations, click here to see how to unlock my American Prosperity Report portfolio.)
DeepSeek claims that it can build AI models for a fraction of the current cost.
I’ve found when something is too good to be true, there’s always a catch.
The $5 million they reported for building the model left out a whole bunch of other costs. My sources tell me the cost was closer to $1 billion, which is where the rest of the industry is at.
But here’s the thing: When AI does becomes cheaper and easier to use, many companies will adopt it — not less. It’s called Jevons’ Paradox.
Source: Jono Hey, Sketchplanations
It says that when something becomes more efficient or cheaper, people end up using it more.
Here are a few examples:
- Energy-Efficient Cars: When cars got better gas mileage, driving became cheaper. So, people started driving more, which meant they still used a lot of gas.
- LED Lighting: LEDs use less energy and cost less to run. But because they are so cheap, people use more lights or leave them on longer.
- Public Transportation: When buses and trains improved, more people chose to travel, which did not reduce traffic problems.
The same idea applies to AI.
As AI becomes more efficient and affordable, businesses and individuals will use it in many new and exciting ways.
This means more jobs in technology, more exciting products and more ways for us to solve big problems.
Why AI Is a Mega Trend
The history of technology shows that when something becomes easier to use and more affordable, its use grows rapidly.
I see the same unfolding with AI…
- More Efficiency: Just like energy-efficient cars or LED lights, AI will make work and daily life more efficient. This means companies can do more with less effort.
- Lower Costs: As AI tools get better and cheaper, more companies will use them. This can lead to big savings and increased profits.
- New Opportunities: With companies like DeepSeek leading the way, there will be many new jobs and innovations. The world of AI is just getting started, and there is so much potential for growth.
AI is still in the early innings.
As more companies invest in AI, including innovative companies like DeepSeek, we can expect even more breakthroughs.
AI will change our world in amazing ways. The future is bright, and the potential is huge.
Let me know what you think about AI’s future at Insider@ProsperityResearch.com. I’d love to hear your thoughts.
Regards,
Charles Mizrahi
Founder, Alpha Investor