The recent surge in AI stocks is truly incredible. I’ve noticed that since the launch of ChatGPT, the repricing of AI concepts across the entire technology stock market has reached what can only be described as an all-time high. Many companies themselves haven’t seen much growth in their performance, yet their stock prices have more than doubled. So what exactly is the logic behind this? I dug deeper into it recently and want to share some observations with everyone.

According to industry reports, global companies’ spending on AI solutions is indeed accelerating. In the coming years, the size of spending across the entire AI industry chain—from chips and infrastructure to the application layer—is expected to continue growing at a fast pace. What does this mean? It means it’s not only chip manufacturers that will benefit—companies across the entire upstream and downstream will also have opportunities.

When it comes to AI stocks, I think the most worth focusing on are a few core targets. NVIDIA is undoubtedly the biggest beneficiary of this AI wave. The company’s GPU chips have almost become the industry standard for AI training and inference. From the time ChatGPT emerged until now, NVIDIA’s stock price has risen by more than 10 times, and its market capitalization has reached the $4 trillion level. Why is it so strong? Put simply, AI development cannot do without computing power, and NVIDIA’s chip computing capability is absolutely leading globally. Not only do major tech companies need them—CEOs at big firms like OpenAI and Meta have also sought out NVIDIA’s leadership in order to secure chips.

In addition to NVIDIA, Broadcom is also a very interesting target. Although it’s not as well known as NVIDIA, it has almost monopolized the market in network communications chips. AI servers require high-speed network connections, and Broadcom’s ASIC chips and network switches are essential. In fiscal year 2024, Broadcom’s revenue reached $31.9 billion, and the share of AI-related products quickly increased to 25%. This clearly shows how significant the boost from the AI wave has been for Broadcom.

As NVIDIA’s direct competitor, AMD has also been playing catch-up these past two years. Although it still has a gap in GPU market share, AMD’s MI300 series chips are already not inferior to NVIDIA in terms of performance, and they cost half as much. As cloud service providers look for a second supplier source, AMD’s growth potential is actually quite large. Personally, I think that if AMD can keep working on cost and ecosystem, its future market share will increase noticeably.

Beyond these chip makers, major tech companies like Microsoft and Google are also investing heavily in AI. They are both major buyers of chips and developers of their own AI applications. This dual role puts them in a very strong position across the entire industry chain.

As for how to invest in AI stocks, my advice is not to focus only on a single stock. In addition to buying individual stocks directly, you can also consider allocating through thematic ETFs, which helps diversify risk. Dollar-cost averaging is also a good choice, because although AI is a long-term trend, short-term volatility is definitely unavoidable.

However, it’s important to remind everyone that there are risks in this space. First is industry uncertainty—AI technology is advancing so fast that even professional investors sometimes struggle to keep up. Second is the issue of valuation—many AI-related companies’ stock prices have already priced in too many expectations, and if growth slows down or policies tighten, the pullback could be quite severe. There’s also regulatory risk— as more countries pay attention to AI, regulations related to data privacy and ethics may become increasingly strict.

So when investing in AI stocks, you must have a long-term mindset. Don’t be frightened by short-term fluctuations, and don’t chase high prices. I’m optimistic about the long-term trend of AI, but when it comes to choosing specific stocks, you still need to be cautious. It’s recommended that you review your holdings on a regular basis and don’t put all your chips into a single company.