Semiconductor stocks have been pummeled recently, opening up a nice potential opportunity to buy these stocks while they are down. However, not all semiconductor stocks are created equal, so investors should be selective.
Let’s look at one chip stock I’d buy, one I’d continue to hold, and one I’d sell.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a “Double Down” signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same “Total Conviction” signal is flashing for a company 1/100th the size of Nvidia. Continue »
Buy: Nvidia
If there is one chip stock to buy in this sell-off, it’s Nvidia (NASDAQ: NVDA). Not only is the company the AI infrastructure leader, but its stock is also incredibly cheap and continues to grow rapidly.
Nvidia’s biggest advantage is that it has created a wide moat with its CUDA software platform for AI model training. It is the dominant player in the space because, many years ago, it planted CUDA in places doing early AI research, making its proprietary software the main platform where most original AI code was written. But this wasn’t the only prescient move Nvidia has made over the years; its 2020 acquisition of Mellanox laid the foundation for its networking platform.
CEO Jensen Huang’s ability to steer Nvidia toward where the market is headed makes the stock an attractive long-term buy, and the recent “acquisitions” of Groq and SchedMD continue to demonstrate his foresight. With the semiconductor sector sell-off taking the stock to a forward P/E of 15 times fiscal 2028 (ending January 2028) analyst estimates, this is a stock to run out and buy.
Hold: AMD
Advanced Micro Devices (NASDAQ: AMD) is one of my favorite chip stocks, but its valuation is not nearly as attractive as that of Nvidia today. Still, the company has a massive growth opportunity ahead, which is why I want to keep holding onto the stock.
AMD is currently riding two powerful trends. The first is inference, where the company has started to make inroads. Inference tends to be more about fast memory access than raw computing power, and its chiplet design can be packaged with more memory. The company boldly made some large graphics processing unit (GPU) deals with OpenAI and Meta Platforms, giving up equity stakes in the process, but this has led to new deals with Anthropic and Microsoft, which will use its GPUs and Helios rack systems to handle some of their inference.
The second big tailwind for AMD is agentic AI. The company is a leader in server central processing units (CPUs), and demand for high-performance CPUs is exploding as AI agents proliferate. The GPU-to-CPU ratio is projected to go from 8:1 for training to 4:1 for inference to 1:1 for agentic AI, leading AMD to forecast that this will be a $220 billion market in the coming years. The company is already making high-core CPUs specifically for agentic AI, which should help it maintain its lead in the space.
