NEW YORK, January 2, 2026, 07:07 ET
- Data-center “pick-and-shovel” suppliers such as Vertiv and Arista have outgrown Alphabet and Microsoft over the past three years, but trade at richer valuation multiples.
- A separate investing column highlighted Alphabet, Tesla and IBM as long-duration ways to play AI beyond chipmakers.
- Barron’s flagged power supply as a potential bottleneck that could test the AI trade in 2026.
Investors are starting 2026 by shifting attention from headline AI model makers to the physical plumbing that keeps data centers running — power, cooling and networking — as valuations rise across the sector.
The debate matters now because the AI boom is increasingly tied to the pace of data-center buildouts. That puts more focus on the companies that sell the hardware and systems needed to deploy and run AI at scale.
It also raises the risk that constraints in the real world — especially electricity supply — could dictate how far and how fast AI spending can go. Barron’s this week called 2026 a “make-or-break year” for the AI trade and warned that power supply may ultimately “apply the brakes.” 1
In a Jan. 2 analysis, The Motley Fool compared recent growth at big AI “hyperscalers” — large cloud companies that operate massive data centers — with key suppliers that support those facilities. It said Vertiv, which makes industrial cooling and electrical systems used in data centers, grew trailing-12-month revenue 70.4% over three years, while Arista Networks, a maker of network switches and related infrastructure, grew 92.8%. 2
By comparison, the same analysis put three-year trailing-12-month revenue growth at 37.3% for Alphabet and 44% for Microsoft. It also said Vertiv’s net income rose 1,250% over the period, versus 148.2% for Arista, 107.2% for Alphabet and 55.5% for Microsoft.
The numbers come with a valuation catch. The analysis said Microsoft and Alphabet traded around 30 times forward earnings, while Vertiv and Arista traded at about 41 times and 46 times, respectively — a premium that reflects investor expectations for the data-center buildout to continue.
“Pick-and-shovel” is investor shorthand for companies that sell equipment to a boom — like selling tools during a gold rush — rather than betting on which software or model wins. For AI, that increasingly includes cooling gear, electrical systems and networking that keep servers online and connected.
A separate Motley Fool column, also carried by Yahoo Finance, highlighted three large-cap names it framed as long-term AI exposures: Alphabet, Tesla and IBM. 3
In that column, Alphabet was pitched as a “high floor” AI stock because its core advertising engine throws off cash that can help fund AI spending. It said Google and YouTube advertising generated $74.1 billion in revenue in the third quarter and pointed to Alphabet’s cloud platform and first-party data as advantages. 4
The same column cast Tesla as a higher-risk robotics bet tied to CEO Elon Musk’s push into AI and humanoid robots, including Optimus, while noting the company’s vehicle business has struggled in recent quarters.
IBM was framed as a rarer dividend-oriented AI play, citing a dividend yield of about 2.2% and a 29-year streak of annual dividend increases. It said IBM is positioning itself around hybrid cloud — combining on-premise computing with public cloud services — and that analysts expect IBM earnings to grow at a high-single-digit annualized rate over the next three to five years.
