Arm puts a big number on its newest venture, but its stock falls
By Britney Nguyen
The chip designer says it has more than $2 billion in customer demand for its first-ever data-center CPU
Shares of Arm, led by CEO Rene Haas, were falling in Wednesday's extended trading.
Arm Holdings just divulged a new figure related to its plans to start making central processing units.
The company said Wednesday that it has more than $2 billion in customer demand for its first-ever data-center CPU stretching through fiscal 2028. Arm's overall data-center-focused business will soon be Arm's largest, according to the earnings materials.
The CPU venture has been a hot theme for investors, and while Arm shares traded meaningfully higher shortly after the report came out, they quickly turned lower, ending the extended session down 6.4%.
The British chip designer reported revenue of $1.49 billion for the fiscal fourth quarter - up 20% from a year ago, and ahead of the FactSet consensus of $1.47 billion. Arm's adjusted earnings of 60 cents per share also beat estimates of 58 cents.
The company's licensing segment saw revenue of $819 million in the March quarter, topping the $781 million that analysts tracked by FactSet were looking for. That segment's revenue was up 29% from the previous year.
Revenue for Arm's royalty business was $671 million, up 11% year over year, but below expectations for $690 million. However, Arm said its data-center royalties more than doubled from the previous year, supporting growth in earnings per share.
For the current quarter, Arm is forecasting revenue of $1.26 billion at the midpoint, slightly above the $1.25 billion that analysts tracked by FactSet were modeling. The chip designer expects adjusted earnings per share of 40 cents at the midpoint, also ahead of estimates for 37 cents.
Management disclosed on the earnings call that operating expenses are expected to step up sequentially in each quarter of the new fiscal year but the company "will be getting back to delivering that incremental margin by end of year" - the way it did prior to its heavy investment phase.
By 2031, CFO Jason Child thinks Arm's intellectual-property business could reach a 65% or so operating margin, while the more capital-intensive chip business could see a margin in the 35% range.
"How quickly we get there is probably going to be a function of how powerful revenue grows," he said.
The company has been a recent beneficiary of surging demand for central processing units, and in March, announced its entry into the booming market where it will eventually compete with its own customers.
On the earnings call, CEO Rene Haas expressed confidence that the company would have the largest market share by CPU type come the end of the decade. "We think it's a market that we can play [into] in a very large way," he said.
See more: Arm's stock shoots to best day in a year as an Nvidia-like chapter may be starting
The Arm AGI CPU was co-developed with Meta Platforms (META) and built for agentic AI, Arm said. That refers to software that can handle tasks autonomously.
"Data centers are expected to require more than 4x current CPU capacity per gigawatt as agentic AI scales, creating a market opportunity of more than $100 billion by 2030," the company said in its shareholder letter.
Arm licenses its instruction-set architecture to customers, including Apple (AAPL) and Qualcomm (QCOM), that develop chips for low-power devices such as smartphones and personal computers.
The company's footprint is now expanding to the data center, where Intel's (INTC) x86 architecture, used by Intel and Advanced Micro Devices (AMD), has historically dominated. Nvidia's (NVDA) Grace CPU, Google's (GOOGL) (GOOG) Axion CPU and Amazon.com's (AMZN) Graviton CPU are based on Arm's architecture, which is focused on energy efficiency - a crucial point for intense AI workloads.
The current agentic AI wave and shift of focus to inference, or running AI models after training, have shined a light on the importance of CPUs in an industry whose attention has so far been dominated by graphics processing units.
UBS analyst Tim Arcuri estimates that the total addressable market for server CPUs could reach about $170 billion in 2030, which "would translate to across-the-board upside" for Intel, AMD and Arm.
Arcuri said in a Monday note that the attach rate of CPUs to AI chips should grow about five times through 2030, citing conversations with experts. Intel CEO Lip-Bu Tan had noted on his company's earnings call that the attach rate of CPUs to GPUs has changed for inference workloads. While training workloads typically require one CPU per eight GPUs, the ratio for inference looks like one CPU for four GPUs, Tan said - and it could be moving toward parity.
Arcuri also sees "net new demand for standalone CPU-only server racks," which he thinks will likely be evenly split between x86-based CPUs and those based on Arm's chip architecture.
Both Intel and Arm will benefit in the near term from AI-driven demand for CPUs, Arcuri said, but in the long run, he sees Arm's instruction set capturing more growth, as it's more optimized to scale core count and throughput while prioritizing energy efficiency.
Arm's share of the server CPU market could reach between 40% and 45% of total units by the end of the decade, he added.
The stock is up more than 100% so far this year.
Emily Bary contributed.
-Britney Nguyen
This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
05-06-26 2031ET
Copyright (c) 2026 Dow Jones & Company, Inc.The articles, information, and content displayed on this webpage may include materials prepared and provided by third parties. Such third-party content is offered for informational purposes only and is not endorsed, reviewed, or verified by Morningstar.
Morningstar makes no representations or warranties regarding the accuracy, completeness, timeliness, or reliability of any third-party content displayed on this site. The views and opinions expressed in third-party content are those of the respective authors and do not necessarily reflect the views of Morningstar, its affiliates, or employees.
Morningstar is not responsible for any errors, omissions, or delays in this content, nor for any actions taken in reliance thereon. Users are advised to exercise their own judgment and seek independent financial advice before making any decisions based on such content. The third-party providers of this content are not affiliated with Morningstar, and their inclusion on this site does not imply any form of partnership, agency, or endorsement.
Popular
4 Stocks to Buy Before They Rise Further
2 Undervalued Stocks to Buy Before They Rebound
The 10 Best Dividend Stocks
12 Best Blue-Chip Stocks to Buy for the Long Term
