Chip stocks have been crashing over concerns that the infrastructure building boom for artificial intelligence (AI) is not sustainable.
However, when some of the most brilliant minds in business are emptying their coffers to build AI data centers at a frantic pace, you can bet that they are getting strong returns on these investments and not just fruitlessly throwing money away, as some of the bears would have you believe.
Let's look at three semiconductor stocks to buy into this sell-off.
Nvidia
Nvidia (NVDA -3.55%) has been the biggest AI infrastructure winner, and it remains one of the best-positioned semiconductor stocks. Its CUDA software platform, where most foundational AI code was written, gives its graphics processing units (GPUs) a wide moat when it comes to AI model training.
Meanwhile, its move earlier this year to acquire Groq and its language processing units (LPUs) to handle the decoding phase of inference helps give it a strong offering in the inference market, as well.
NASDAQ: NVDA
Key Data Points
The company also recently signed a large deal with SK Hynix in which the Korean memory maker will supply it with high-bandwidth memory (HBM), a vital component that needs to be packaged with AI chips. Since HBM is severely supply-constrained, this access gives the company a big edge.
With the stock trading at a forward price-to-earnings ratio (P/E) of just 15 times the analysts' consensus for fiscal 2028 (ending January 2028), it is a top growth stock trading at a cheap valuation. I would be a buyer while it's on sale.
Image source: The Motley Fool.
Advanced Micro Devices
The faster they go up, the faster they go down. Shares of Advanced Micro Devices (AMD -5.51%) have shot up this year on the back of its strong growth prospects, but they have been hit particularly hard in this recent chip sell-off. Still, the stock's price action doesn't impact the huge growth drivers the company has ahead.
AMD is just starting to see growth surge on tailwinds from inference and agentic AI. The company now has large deals in place with OpenAI, Meta Platforms, Anthropic, and Microsoft for its next-generation GPUs and Helios Rack systems to help run inference for these hyperscalers.
NASDAQ: AMD
Key Data Points
Management paid a hefty price to secure deals with OpenAI and Meta in the form of warrants for up to 10% of its stock, but it did open the doors for future growth and helped establish it as a strong No. 2 player in the space. These deals should drive tremendous revenue growth in the coming years.
At the same time, the company has a huge opportunity with agentic AI and its central processing unit (CPU) business. AI agents will require more CPUs in data centers, and AMD is a leader in this space. It now sees this market ballooning to $220 billion in the next few years.
With AMD having some of the best growth prospects in the industry, now can be a great time to buy the stock following its pullback.
Broadcom
Another semiconductor stock with huge growth ahead is Broadcom (AVGO -2.78%). The company helped Alphabet co-develop its tensor processing units (TPUs), and it is now riding that huge tailwind.
This has also led other hyperscalers to ask it to help them develop their own custom AI chips. Broadcom sees this producing more than $100 billion in sales for fiscal 2027 (ending October 2027), with Citigroup projecting its AI business could grow to $180 billion in fiscal 2028.
NASDAQ: AVGO
Key Data Points
Broadcom has also smartly locked up HBM supply with Samsung Electronics, which likely means hyperscalers looking to produce custom AI chips at scale will have to turn to it for help. And its custom chip business also feeds into its fast-growing data center networking segment.
Following a sell-off in the stock, it now trades at a forward P/E of 19 based on fiscal 2027 analyst estimates, making this a great time to buy while it's cheap.
