The biggest thing happening in the stock market this month is the massive tailwind behind artificial intelligence stocks. The biggest hyperscalers -- Microsoft, Alphabet, Amazon, and Meta Platforms -- have announced plans to spend as much as $650 billion this year on AI investments, including capital expenditures for chips and outfitting data centers.
While stocks in those companies have slipped a bit as investors absorb the effect of those massive spending commitments, I think now is the time for investors to look hard at AI infrastructure companies. These five names will be among the biggest beneficiaries of the global AI buildout.
Image source: Getty Images.
1. Nvidia
Nvidia (NVDA 1.59%) deserves the top spot on any list of AI stocks to buy. Its graphics processing units (GPUs) are the backbone of data centers, as they are bundled by the hundreds to perform highly complex AI tasks. Nvidia's GPUs are the current gold standard for developing, training, and running AI platforms.
Nvidia is scheduled to report quarterly earnings on Feb. 25, and I'm expecting a big jump in the stock afterward. Its last report, covering the fiscal third quarter of 2026 (ending Oct. 25, 2025), showed record revenue of $57 billion, with $51.2 billion from its data center segment.
NASDAQ: NVDA
Key Data Points
All indications are that demand for Nvidia's Blackwell GPUs remains strong, and its next-generation Vera Rubin chips are expected this year. Nvidia is set for another strong year.
2. Taiwan Semiconductor Manufacturing
Taiwan Semiconductor Manufacturing (TSM 4.64%), best known as TSMC, is in the catbird seat right now. Nvidia and other chipmakers design the chips, but they don't have the capability to manufacture them. For that, they turn to a foundry, and TSMC is the foundry of choice for top chipmakers. It produced more than 11,800 products using 288 separate processes in 2024.
TSMC specializes in making powerful chips with small transistors. The smaller the transistors in a chip, the more efficient and powerful it is. As recently as 2023, TSMC got more than half of its revenue from chips larger than 7 nanometers, but in the fourth quarter of 2025, chips that were 3nm or 5nm accounted for 63% of its shipments.
3. Nebius Group
We've already covered chips, both designing them and manufacturing them. Now let's turn to AI data centers, and for that, my top pick is Nebius Group (NBIS 2.92%). Nebius is a Dutch company that builds data centers to provide full-stack AI cloud platforms that developers and hyperscalers alike can use to train and run AI programs.
Nebius uses Nvidia GPUs in its data centers and is rapidly expanding its operations. It had 220 megawatts of power connected for its data centers at the end of 2025, but plans to have between 800 megawatts and 1 gigawatt of connected power by the end of this year.
NASDAQ: NBIS
Key Data Points
Nebius also has contracts worth up to $19.4 billion with Microsoft and $3 billion with Meta Platforms to provide AI computing capacity. As it expands, you can expect it to land even more.
4. Digital Realty Trust
If you want another way to invest in data centers, consider Digital Realty Trust (DLR 0.45%). This differs from Nebius Group in that it's a real estate investment trust, or REIT, which means that it's required to return at least 90% of its taxable income to shareholders in the form of dividends.
Digital Realty has more than 300 data centers in 50 metropolitan markets, and counts Nvidia, Amazon, Microsoft, and Alphabet among its customers. Its revenue was $1.6 billion in the fourth quarter, up 14% from a year ago. Earnings per share was $0.24, versus $0.15 the previous year.
On top of that, Digital Realty Trust pays a dividend yield of 2.8% -- something that you won't get from Nebius.
5. Credo Technology Group
We've covered chips and data centers, but we haven't covered how those data centers make chips work together. For that, you need Credo Technology Group (CRDO 1.73%).
NASDAQ: CRDO
Key Data Points
Credo makes what it calls Active Electrical Cables (AECs), which are considered better than conventional copper cables to connect clustered chips in data centers. AECs use signal processors in the wiring that help move data faster and more efficiently.
Earnings for the second fiscal quarter of 2026 (ending Nov. 1, 2025) showed revenue of $268 million, up 272% from a year ago. The company is expecting revenue between $335 million and $345 million in the next fiscal year.
