AAPL

TSLA

AMZN

META

AMD

NVDA

PEP

COST

ADBE

GOOG

AMGN

HON

INTC

INTU

NFLX

ADP

SBUX

MRNA

Markets

NVDA

Prediction: The Nasdaq's Artificial Intelligence (AI) Supercycle Rally Will Outlast the Skeptics. 3 Best Growth Stocks to Own.

April 22, 2026 — 04:50 am EDT

Written by

Geoffrey Seiler for

The Motley Fool->

Key Points

- Nvidia remains the king of AI and has positioned itself well for the future.

- AMD offers products and services well-suited for the newest hot AI markets, inference and agentic AI.

- An elongated memory supercycle would make Micron a big winner.

- 10 stocks we like better than Nvidia ›

Artificial intelligence (AI) skeptics believe that spending on data center infrastructure is about to peak. More than $700 billion is expected to be spent on AI infrastructure this year, which is more than the gross domestic product of all but two dozen countries.

That's a massive amount of money, but the companies leading the charge on this spending have shown no signs of letting up. Most have publicly said they see AI as a once-in-a-generation opportunity and would rather risk overspending than underspending. And while investors like Michael Burry have questioned their returns, top tech companies have all said the AI chips they are buying have longer useful lives than assumed and that they are getting strong returns on their investments.

Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »

Image source: Getty Images.

This includes Alphabet, which last fall said it has 100% utilization of its eight-year-old chips. Meanwhile, neocloud operator CoreWeave has said that its Nvidia (NASDAQ: NVDA) A100 chips, introduced in 2020, are all fully booked, while noting that it had some 2022 H100 chips become available that it was able to book for 95% of their original price. This helps put a hole in the bear argument, and is a strong indication that the AI infrastructure supercycle could last a lot longer than skeptics think.

With that in mind, let's look at three of the best AI stocks to buy today.

Nvidia: Still the king

Even as challengers mount, Nvidia is still the king of AI infrastructure and a top stock to own. The company continues to have a wide moat in AI model training, as most original foundational code was written on its CUDA software platform. Meanwhile, Nvidia has set itself up to also be a major player in the areas of inference and agentic AI.

The company's partnership deal with Groq gives it language processing units designed specifically for inference, while it's also designed its own central processing units and data processing units. Together with its networking portfolio, it can now offer end-to-end rack solutions for specific AI tasks, like training, inference, and AI agents. This evolution will help make Nvidia a continued AI winner well into the future.

Advanced Micro Devices: Made for inference and agentic AI

While Advanced Micro Devices (NASDAQ: AMD) largely missed out on the AI model training wave, the company is very well positioned for both inference and agentic AI. It's greatly improved its ROCm software solution, and its upcoming MI450 chip looks particularly well designed for inference. It uses a chiplet design that packs in more memory, making it particularly useful for this task. It also has deals with both Meta Platforms and OpenAI to deliver them both 6 gigawatts worth of GPUs, which will be a revenue driver.

The company's biggest opportunity, though, may be in high-performance CPUs. AMD is the leader in this space, and it's currently designing CPUs with high core counts to specifically handle agentic AI. Servers managing AI agents will need much higher CPU-to-GPU ratios, and there is already starting to be a supply shortage for CPUs. Its acquisition of ZT System is also allowing it to offer full-rack systems for inference and agentic AI as well. This all positions AMD to be a huge winner in the coming years.

Micron Technology: Riding the memory wave

Worries about when the memory supercycle will end have kept Micron Technology (NASDAQ: MU) shares at a very low valuation, with a forward P/E of just 4.5 times fiscal 2027 analyst earnings estimates. However, this cycle could be set to last for much longer than many pundits think. For optimal performance, GPUs and other AI chips need to be packaged with high-bandwidth memory (HBM), and as demand for these chips continues to soar, so will demand for HBM. Meanwhile, Micron and its two Korean competitors in the space are starting to lock in three- to five-year HBM contracts.

On top of that, HBM can require more than three