DeepSeek AI: Is It Really A Threat to US Tech Giants?

Today, we’re talking about a company that’s causing quite a stir in the tech world: DeepSeek. Now, I know many of you are familiar with the likes of OpenAI’s ChatGPT, but trust me, this is a different beast altogether, and it’s shaking things up in a major way. As an Indian investor, I’m always looking for opportunities and, frankly, understanding the potential disruptions that can impact my portfolio. So, let’s understand DeepSeek from that lens.

#1. What makes DeepSeek so special?

Let’s understand the DeepSeek’s Secret Sauce.

It is efficient.

It’s not just another AI chatbot, it’s a game-changer for a couple of very important reasons:

#1.1 The Efficiency Marvel

Think of it like this – imagine two cooks making the same dish, but one uses a fancy, expensive oven and a ton of ingredients, while the other creates the same gourmet meal using a basic stove but with much less resources.

That’s DeepSeek in a nutshell.

They’ve developed their AI models, particularly the “R1 model”, with significantly less computing power.

I would like to specifically mention that, DeepSeek AI model works on fewer of those super-expensive Nvidia chips that everyone in AI seems to be hoarding. That is why in yesterday’s trading session alone, the Nvidia stocks lost $600 billion in market cap (read).

While giants like OpenAI are using 16,000 chips or more to train their models, DeepSeek is achieving similar results with only around 2,000 chips.

This is a huge leap in efficiency.

They’ve basically figured out a smarter way to train their AI, making it far more cost-effective.

This is significant because the general belief until now was that to get good at AI, you need to throw money and to run it will expensive hardwares. DeepSeek is challenging that notion.

#1.2 Open Source Philosophy

Unlike OpenAI, who is very secretive about its technology, DeepSeek is adopting an open-source approach.

Think of Android or WordPress, which allows anyone to download their code and then modify and build on it, like DeepSeek’s open-source AI. In contrast, iOS or Wix, which is tightly controlled by Apple (it is closed-source), is like OpenAI, where the underlying technology is kept secret. This openness versus secrecy is a major difference, fostering innovation and adoption in one case and controlled access in the other.

[Note: I’ve specifically selected this example to highlight that benefit of an open source technology (like Android OS and WordPress) which gets mass acceptance. But iOS and Wix are also very successful products (one is an OS and other is a CMS) even though it is closed source.]

DeepSeek is openly sharing the underlying code, which is like giving the recipe of the dish to anyone who wants to make it.

What this means is that other developers and researchers can use DeepSeek’s technology, build on it, and create even more innovations.

It’s like saying, “Here’s how we did it, now go make it even better!”. This fosters collaboration and could accelerate the pace of AI development. In fact, a big part of the reason why DeepSeek has been able to create very powerful AI models is because of the open source ecosystem.

To put it in our Indian context, it’s like a small, local jugaad innovation that achieves the same result as a very expensive, foreign made product.

#2. Why are US Tech Stocks Trembling?

Now, why is all of this causing a panic in the US stock market?

I can understand the impact of DeepSeek on AI models like Open AI’s ChatGPT, but why shares of NVIDIA, Amazon, Microsoft, Google are falling?

Well, imagine you’ve invested heavily in a company that makes and sells those expensive ovens – suddenly, someone has invented a better and cheaper way of cooking. Your investment doesn’t look so good anymore, right? That’s the problem for some big US tech companies right now.

• The Nvidia Nightmare: Nvidia, the king of AI chips, has seen its stock price plummet. Their chips are the “gold standard” for AI training. If DeepSeek can achieve similar results without them, it raises serious questions about the long-term demand for these expensive chips. This sent shivers down the spine of investors. They realized that spending so much money on expensive chips might not be the only path to AI success.
• The “Big Tech” Shakeup: Meta, Microsoft, and Amazon have been investing heavily in building massive data centers packed with expensive, specialized chips – primarily Nvidia’s – to power their AI efforts. They were essentially betting that the only way to build powerful AI was to spend billions on these chips. Now, DeepSeek has shown that you can achieve comparable results with far fewer, cheaper chips. This makes investors question whether Meta, Microsoft, and Amazon have overspent on their infrastructure. Are their expensive data centers really necessary? If DeepSeek’s approach becomes the norm, will these companies struggle to compete with others who can develop AI at a much lower cost? This uncertainty and the fear of future losses are causing investors to sell off their stocks, leading to a drop in their share prices.

This is a good lesson for investors like us. We must not get carried away by the hype and always look for efficiency and scalability of the underlying technology.

#3. What Does This Mean for the IT Industry?

So, what does all of this mean for the tech industry as a whole? Here’s are my take:

• Democratization of AI: DeepSeek’s approach is a game-changer for the IT industry. The efficiency that they have shown basically means that you don’t have to be a giant tech company to create good quality AI. This has the potential to democratize AI. It will make it accessible to smaller companies and even startups. This can unleash a wave of innovation globally. This is very good for India because we can become a major player in the AI revolution.
• Shift in Investment Focus: Investors might start to shift their focus from companies heavily invested in hardware to those focusing on efficient AI software. This could lead to the rise of new tech stars who are good at the algorithms, rather than just the hardware. We might see a boom in software companies in the AI space.
• The Power of Open Source: DeepSeek’s open source approach emphasizes that collaboration is key. It highlights how innovation can be accelerated by sharing technology.
• A New kind of geopolitical competition: DeepSeek’s development is showing the world that the US’s lead in AI is being challenged. This could change the direction of global AI research. The US might feel the need to increase its spending in this area. This could lead to even more innovation. Experts are calling it a Sputnik moment for AI (read about what is a Sputnik moment here).

Conclusion

I see this as a major opportunity and a big wake up call at the same time.

The emergence of DeepSeek is not about “China vs. USA”, rather it’s about efficiency and accessibility.

For us, it’s a good reminder that we should not get overly reliant on just one particular technology.

We need to keep an eye on the global AI innovations and be ready to adapt. We also need to push our own capabilities in the software and algorithms space.

DeepSeek is a reminder that the tech world is not static. New players will emerge, and innovation can come from anywhere, not just Silicon Valley. We need to be adaptable and informed.

If you found this article useful, please share it with fellow investors or leave your thoughts in the comments below!

Have a happy investing.