It’s been hard to find a good deal on the stock market lately. Since the big tech hyperscalers like it Microsoft, Meta platforms, Amazon, Alphabet, adzeand Apple announced that they will spend nearly $700 billion on capital expenditures this year, artificial intelligence (AI) stocks have seen epic selling pressure.
Still, smart investors look closely at macro themes. The ongoing selloff in the tech landscape is primarily hitting software stocks the hardest. As spending on AI infrastructure increases, growth can be found in companies positioned to absorb large-scale budget spending.
Where to invest $1,000 right now? Our team of analysts just revealed what they think they are 10 best stocks to buy right now when you join Stock Advisor. View stocks »
Let’s dig into two AI chip stocks that look like great long-term buys as the AI infrastructure revolution accelerates.
Image source: Getty Images.
On the surface, $700 billion worth of AI investment probably makes you think twice Nvidia and Advanced microdevices are the AI semiconductor stocks to buy right now. After all, Nvidia’s Blackwell GPUs and upcoming Rubin GPUs were so popular that the company’s backlog was estimated by management to be in the hundreds of billions of dollars.
Some developers are investing in ways to migrate away from incumbents, however. Both Meta and Alphabet are designing their own custom AI accelerators — known as MTIA and TPU chips. Broadcom(NASDAQ: AVGO) helps both companies with their custom silicon solutions.
Another way Broadcom is benefiting from growing data center construction is through networking. When hyperscalers buy chips from Nvidia and AMD, they buy GPUs in the hundreds of thousands.
To connect these clusters, you need to connect the GPUs with routing and switching equipment. Broadcom’s Tomahawk and Jericho chips are the gold standard when it comes to stitching AI fabrics together and keeping GPU clusters processing data at ultra-low latencies.
The theme here is that Broadcom is quietly positioned to benefit alongside its peers as increasing amounts of AI capital are allocated to critical networking equipment and move beyond a sole focus on purchasing GPUs.
For almost three years, generative AI has been a core function of large language models (LLMs) — chatbots like OpenAI’s ChatGPT, Alphabet’s Gemini, X’s Grok or Anthropic’s Claude. Growing AI infrastructure doesn’t mean building more chatbots, though.
Hyperscalers have their eyes on far more sophisticated applications — autonomous vehicles, agentic AI, and humanoid robots, to name a few. These use cases require more intensive development than building a basic LLM.
Because of this, AI workloads are now growing at an unprecedented rate as major tech companies continuously train and infer their next-generation models. In essence, compute capacity is no longer the primary pain point for bringing new AI-based products to life. GPU clusters must be supplemented with memory and storage solutions. Micron technology(NASDAQ: MU) specializes in high bandwidth memory (HBM) chips.
To show how critical HBM has become, consider that prices for dynamic random access memory (DRAM) and NAND chips are expected to increase by as much as 60% and 38%, respectively, in the first quarter alone. Given that Micron has already sold out of its HBM inventory for 2026, the company should be able to wield enormous pricing power as memory walls become the new bottleneck along the AI chip value chain.
In the anatomy of an AI data center, the GPU architectures designed by Nvidia and AMD can be considered the brains for the server rack. The special chips from Broadcom and Micron represent both the nervous system and the memory that helps the brain function.
Hyperscalers no longer focus on buying GPUs in bulk for experimental work. Rather, the growing AI infrastructure signals that big tech now has its eyes on efficiency and scale. As development bottlenecks shift from capacity constraints to data movement and storage, both Broadcom and Micron are in a lucrative position to capture incremental capital spending.
Data by YCharts.
Against this backdrop, I see Broadcom and Micron as the new players powering the AI infrastructure revolution over the next few years. Smart investors might consider doubling down on both stocks right now, as their respective valuations look reasonable amid the ongoing tech selloff.
Before buying stock in Micron Technology, consider the following:
The Motley Fool Stock Advisor the analyst team has just identified what they think they are 10 best stocks for investors to buy now… and Micron Technology was not one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004… if you invested $1,000 at the time of our recommendation, you would have $415,256!* Or when Nvidia made this list on April 15, 2005… if you invested $1,000 at the time of our recommendation, you would have $1,151,865!*
Now, it’s worth noting Stock advisor the total average return is 892% — a market-crushing outperformance compared to 194% for the S&P 500. Don’t miss the latest top 10 list, available with Stock advisorand join an investor community created by individual investors for individual investors.
See the 10 stocks »
*The Equity Advisor returns as of February 20, 2026.
Adam Spatacco has positions in Alphabet, Amazon, Apple, Meta Platforms, Microsoft, Nvidia and Tesla. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Apple, Meta Platforms, Micron Technology, Microsoft, Nvidia and Tesla. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy.
Top Stocks to Double Up on Right Now was originally published by The Motley Fool
