By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
While Arm (ARM) has shown strong growth and potential in AI and data center royalties, the panelists express concerns about its high valuation, margin compression risks from entering hardware manufacturing, and the cyclical nature of semiconductor demand. The shift towards agentic AI and Arm's entry into the CPU market are seen as potential growth vectors, but the execution risks and competition from open IP like RISC-V are significant.
Risk: Margin compression risks and capital expenditure requirements from entering hardware manufacturing, as well as the cyclical nature of semiconductor demand and potential market share erosion from RISC-V architecture.
Opportunity: The potential growth from the shift towards agentic AI and Arm's entry into the CPU market, as well as the possibility of offsetting licensing pressure with higher-margin software services.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Arm Holdings (NASDAQ: ARM) has more than doubled year to date and sits at a market cap above $300 billion. There's still a big gap between its current valuation and the $1 trillion milestone, but AI tailwinds could potentially push the stock to this benchmark within a few years.
A vast intellectual property (IP) portfolio tied directly to AI infrastructure and surging demand for central processing units (CPUs) has propelled the stock, and these catalysts could extend the rally.
Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now, when you join Stock Advisor. See the stocks »
Almost all of Arm's revenue comes from its IP portfolio. The company designs key components for smartphones, laptops, AI data centers, 5G, and other high-demand applications. Artificial intelligence has been the major revenue driver recently.
Overall revenue reached $4.92 billion in its fiscal 2026, which was up by 23% year over year. It represented the company's third consecutive fiscal year of more than 20% revenue growth.
Arm makes royalties from various tech giants that are selling AI chips, including Nvidia (NASDAQ: NVDA). Companies will pay Arm for its intellectual property so they can produce high-demand products.
Arm's business model resembles a LEGO instruction booklet. Without that booklet, you can't build the LEGO displayed on the box. Since the instruction booklet covers products critical to AI infrastructure and consumer devices, Arm can charge a high premium to companies that want to use it legally.
Although Arm has made almost all of its revenue from licenses and royalties, the company recently said it would enter the hardware industry and create its own CPUs.
The company's vast intellectual property has established it as a leader in the industry. That will make it easier for Arm to penetrate the CPU industry with its own hardware. Arm made this announcement in March and already has more than $2 billion of customer demand across fiscal 2027 and fiscal 2028.
Arm expects data centers to become the largest part of its business. Licensing and royalty fees will continue to operate in the background, but CPUs are expected to be the major catalyst for future years.
In that case, Arm Holdings has a real shot at becoming a $1 trillion company within a few years. The demand for agentic AI certainly works in the company's favor, since AI agents need more CPUs. That has shifted the CPU-to-GPU ratio in favor of CPUs.
Currently, the CPU-to-GPU ratio is between 1:4 and 1:8, according to TrendForce. That means a data center usually has four to eight GPUs for each CPU. The future ratio is expected to shift closer to 1:1 or 1:2.
That means CPU demand can quickly quadruple. Arm is well-positioned to ride that momentum. If it does, and the stock follows suit, Arm would be worth more than $1 trillion before 2030. It's a tall order for a stock already priced at 60x trailing sales, but still a plausible growth story.
Before you buy stock in Arm Holdings, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Arm Holdings wasn’t 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’d have $371,519! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,281,302!
Now, it’s worth noting Stock Advisor’s total average return is 892% — a market-crushing outperformance compared to 206% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of July 24, 2026. *
Marc Guberti has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Arm Holdings and Nvidia. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"Arm's path to $1T requires flawless execution on unproven CPU hardware margins and continued AI hype, making the current 60x sales valuation look more like hope than a base case."
Arm (ARM) has undeniably ridden the AI wave, with 23%+ revenue growth for three straight years and a shift toward higher-margin data-center royalties. The move into its own CPUs, backed by >$2B in committed demand for FY27-28, plus the expected CPU:GPU ratio swing from 1:4-8 toward 1:1-2 under agentic AI, supplies a credible growth vector. At 60x trailing sales and ~$300B market cap, however, the stock prices in heroic execution and multiple expansion to reach $1T by 2030. Missing context: Arm still derives the vast majority of revenue from licensing/royalties; hardware margins and competitive moats versus custom silicon from hyperscalers remain unproven.
The article glosses over Arm's heavy dependence on a handful of customers (including Nvidia), rising geopolitical risk around China, and the real possibility that hyperscalers continue building their own ASICs, capping Arm's CPU upside and leaving the 60x sales multiple vulnerable to any slowdown in AI capex.
"Arm’s transition from a high-margin IP licensor to a hardware manufacturer threatens its profitability profile and fails to justify its current 60x sales valuation."
Arm Holdings (ARM) is currently priced for perfection, trading at an astronomical 60x trailing sales. While the shift toward agentic AI potentially increases CPU-to-GPU ratios, the article conflates Arm’s licensing dominance with the risks of hardware manufacturing. Entering the hardware space introduces massive margin compression risks and capital expenditure requirements that the current premium valuation does not account for. Furthermore, the $1 trillion valuation thesis assumes linear royalty growth, ignoring the cyclical nature of semiconductor demand and the potential for RISC-V architecture to erode Arm's market share in the long term. ARM is a high-quality asset, but at this multiple, the margin of safety is non-existent.
If Arm successfully executes its hardware pivot, it could capture the entire value chain of AI compute, justifying the premium through vertical integration that rivals like Nvidia have already proven highly lucrative.
"ARM is priced for flawless CPU execution and market-share gains that haven't been proven, while the article conflates design wins with revenue and ignores that margin compression from hardware entry offsets much of the TAM expansion upside."
ARM's $300B valuation at 60x trailing sales prices in near-perfect execution of a CPU transition that hasn't yet materialized at scale. The article conflates *design wins* ($2B pipeline across two years) with *revenue reality*—that's ~$1B annually against a $4.92B base, or 20% incremental upside if fully realized. The CPU-to-GPU ratio shift (1:4-8 to 1:1-2) is real, but that's a *total addressable market* expansion, not Arm-specific demand. Nvidia, AMD, and custom silicon (Tesla Dojo, Google TPU) are all competing for that same shift. The article ignores execution risk: Arm has zero hardware manufacturing experience, zero supply chain, and zero track record shipping at scale. Licensing is 95%+ margin; CPUs are 40-50% margin at best. Margin compression is baked in.
If agentic AI truly does shift the ratio to 1:1 and Arm captures even 30-40% of incremental CPU demand, the TAM expansion alone justifies current multiples—and the stock could still run to $1T if licensing revenue compounds at 20%+ while CPU revenue scales to $5-10B by 2028-2029.
"Durable AI data-center demand and a higher-margin CPU business could drive a multi-year re-rating toward a $1 trillion market cap, but only if Arm sustains licensing momentum and monetizes its CPU push without degrading margins."
Arm's bull case rests on AI licensing tailwinds and the prospect of Arm-based CPUs generating new, high-margin revenue as data centers scale. Yet the gain hinges on a durable AI hardware cycle and Arm's ability to monetize a stronger CPU footprint without eroding margins. Risks include: licensing revenue that is cyclical and tied to smartphone/PC refresh cycles; capital-intensive CPU development with uncertain ROI; a GPU-dominated AI era that could keep CPU demand weaker than hoped; competition from open IP (RISC-V) and ecosystem shifts; and a valuation that already prices aggressive growth, making a meaningful re-rate fragile if capex or chip demand slows.
The strongest counterpoint is that licensing remains the backbone of Arm’s revenue and it could decelerate even as CPU pastiches lag in margin; a GPU-centric AI cycle, RISC-V competition, or slower data-center demand could compress earnings and keep the '1T' dream out of reach.
"Early Neoverse design wins at major hyperscalers provide nearer-term revenue proof points than the distant CPU pivot everyone is debating."
Claude's $1B annual pipeline math understates committed demand; the $2B is for FY27-28 only, implying acceleration beyond that. More critically, nobody has flagged Arm's Neoverse N2/V2 already winning sockets at AWS, Google, and Microsoft today—real silicon, not vaporware. If those hyperscalers ship at scale in 2025-26, the licensing+royalty flywheel compounds before CPU margins ever dilute the model.
"Hyperscalers building custom silicon are incentivized to minimize royalty payments, potentially decoupling Arm's architectural dominance from its revenue growth."
Grok, your focus on Neoverse sockets misses the 'moat' reality: AWS Graviton and Google Axion are specifically designed to reduce reliance on merchant silicon providers. By building custom Arm-based chips, hyperscalers are essentially 'in-sourcing' the value that Arm wants to capture as royalties. If these hyperscalers optimize their internal stacks to minimize royalty-bearing features, Arm’s revenue per chip could actually decline even as their architectural footprint grows. You are betting on volume; I am betting on margin erosion.
"Hyperscalers can't squeeze Arm royalties without abandoning Arm architecture entirely; the moat risk is defection, not margin compression within the Arm ecosystem."
Gemini's 'in-sourcing' thesis is sharper than I credited. But it conflates two different revenue streams: Arm makes royalties on *every* Graviton/Axion chip shipped, regardless of optimization. Hyperscalers can't minimize royalties without abandoning Arm ISA entirely—switching to RISC-V or custom. That's the real threat, not margin compression within Arm licensing. The question is defection risk, not royalty erosion.
"Royalties won't disappear with 'in-sourcing'—they morph into software/platform IP and services revenue, so margin risk remains even if hyperscalers build more of their own chips."
Claude, you stress margin compression and execution risk, but you underestimate Arm's software/IP monetization around CPUs. If hyperscalers license Arm cores while paying for optimization tooling, security libs, and verification IP, Arm can offset CPU margin declines with higher-margin software services. In-sourcing reduces royalties, but it doesn't erase them—royalties morph into platform/IP and services revenue. The bear case hinges on software monetization failing to offset licensing pressure.
While Arm (ARM) has shown strong growth and potential in AI and data center royalties, the panelists express concerns about its high valuation, margin compression risks from entering hardware manufacturing, and the cyclical nature of semiconductor demand. The shift towards agentic AI and Arm's entry into the CPU market are seen as potential growth vectors, but the execution risks and competition from open IP like RISC-V are significant.
The potential growth from the shift towards agentic AI and Arm's entry into the CPU market, as well as the possibility of offsetting licensing pressure with higher-margin software services.
Margin compression risks and capital expenditure requirements from entering hardware manufacturing, as well as the cyclical nature of semiconductor demand and potential market share erosion from RISC-V architecture.