By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
Panelists are divided on ARM's future, with concerns about execution, competition, and regulatory risks outweighing optimism about AI demand and licensing moat.
Risk: Internal silicon development by hyperscalers and regulatory/export control headwinds
Opportunity: Potential upside from AI demand and licensing moat
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 plc (NASDAQ:ARM) is one of the Best Up and Coming AI Stocks to Buy Now. The company completed its IPO in 2023 with a market valuation of roughly $54.5 billion and was considered one of the largest IPOs of the year. Since then, the company has gone on to become a prominent AI company with a current market valuation of $366.276 billion.
Recently, on June 4, Mizuho raised the firm’s price target on Arm Holdings plc (NASDAQ:ARM) from $425 to $500 and maintained an Outperform rating on the shares. The firm noted that agentic AI tailwinds are accelerating for Arm, driven by expanding platform partnerships with Oracle and ByteDance. Moreover, Mizuho now believes Arm can generate as much as $15 billion in agentic AI infrastructure CPU revenue by fiscal year 2031. As a result, the firm updated its earnings estimates higher for the company.
Moreover, the company posted strong results for fiscal Q4 2026 on May 7. During the quarter, the company posted revenue of $1.49 billion and adjusted EPS of $0.60, surpassing estimates of $1.47 billion and estimated EPS of $0.58. For the fiscal first quarter of 2027, management expects revenue of around $1.26 billion and adjusted EPS between $0.36 and $0.44.
Arm Holdings plc (NASDAQ:ARM) is involved in the licensing, research, marketing, and development of system IP, microprocessors, graphics processing units, physical IP, and associated systems IP, software, and tools. The company’s operations are divided into the following geographical segments: the United Kingdom, the United States, and Other Countries.
While we acknowledge the potential of ARM as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.
READ NEXT: 9 Most Undervalued Foreign Stocks to Buy Now and 10 Most Undervalued US Stocks According to Hedge Funds.** **
Disclosure: None. Follow Insider Monkey on Google News.
Four leading AI models discuss this article
"The $500 PT assumes flawless AI execution that current valuation and modest Q1 guidance make far from assured."
Mizuho's upgrade to a $500 price target on ARM leans on speculative $15B agentic AI CPU revenue by FY2031 via Oracle and ByteDance ties, yet the $366B market cap already embeds rapid adoption. Fiscal Q4 revenue of $1.49B beat estimates narrowly, but Q1 2027 guidance of $1.26B implies a sharp sequential drop that could pressure sentiment. Licensing exposure leaves ARM vulnerable to customer in-sourcing of chips and slower royalty ramps if AI infrastructure spending cools. The article downplays execution risk on unproven agentic workloads and competition from Nvidia and custom ASICs.
ARM's royalty model could scale faster than modeled if hyperscalers standardize on its IP for efficiency, turning the $15B forecast into a conservative base case rather than blue-sky.
"Ambitious AI-driven upside is contingent on sustained AI capex and execution; current valuation prices in aggressive growth, leaving little room for error."
Arm's licensing moat and AI demand could lift platform revenue, especially as Oracle and ByteDance deepen ties. Mizuho's $500 target implies ~25–30% upside from a $366B market cap, anchored by the ambitious 2031 goal of $15B in agentic AI infrastructure CPU revenue. But the article glosses over risks: Arm's business is licensing/royalties; achieving that revenue scale hinges on sustained AI capex, favorable customer mix, and broad deployment, all uncertain. Also watch for concentration risk, potential margin compression, competition from x86 incumbents and emerging RISC-V/IP, and regulatory/export-control headwinds. The missing context: margins, backlog, capex cycles, and how much of the revenue is recurring vs one-off licensing fees.
The $500 target hinges on an unlikely linear ramp to $15B of CPU revenue by 2031; a stall in AI demand or stricter export controls could trigger sharp multiple compression. Licensing revenue is inherently lumpy and customer-driven, so execution risk is non-trivial.
"ARM is currently priced for perfection, with a valuation that ignores the inherent risks of long-term revenue forecasting in the volatile AI infrastructure cycle."
Mizuho’s $500 target on ARM reflects a aggressive valuation multiple expansion based on speculative 'agentic AI' revenue projections for 2031. While ARM’s transition to the v9 architecture and higher royalty rates is impressive, the stock is currently trading at a massive premium—roughly 60x-70x forward earnings—which prices in near-perfect execution. The $15 billion revenue projection for 2031 assumes ARM successfully captures significant value in the data center CPU market against x86 incumbents and custom silicon efforts from hyperscalers. Investors are paying for a best-case scenario where ARM becomes the foundational compute layer for all agentic AI, ignoring the cyclical nature of semiconductor licensing and potential saturation in the smartphone market.
If ARM successfully captures the edge AI market, their royalty per chip could increase exponentially, justifying a permanent re-rating of their multiple as a high-margin software-like utility.
"The bull case hinges on a single $15B revenue line item in 2031 that has no near-term validation, while current valuation already prices in substantial AI upside."
Mizuho's $425→$500 PT (+17.6%) rests almost entirely on a $15B agentic AI CPU revenue projection by FY2031—seven years out, unaudited, and contingent on Oracle/ByteDance partnerships scaling dramatically. The Q4 beat was modest (1.4% revenue upside, 3.4% EPS upside) and Q1 guidance implies deceleration. ARM trades at ~68x forward P/E on 2027E earnings; even at $500, that's still 58x. The article itself hedges by recommending 'other AI stocks' with better risk/reward, which is telling. Licensing models are capital-light but also have limited upside surprise potential—ARM's moat is durable but not explosive.
If agentic AI infrastructure truly inflects as Mizuho claims, ARM's architectural dominance in low-power compute could drive outsized licensing revenue for a decade, justifying premium multiples. The ByteDance/Oracle partnerships are real, not speculative.
"ARM's China exposure via ByteDance creates unpriced binary regulatory risks beyond AI demand uncertainty."
Claude correctly flags the 68x forward multiple and modest Q4 beat, but underplays how ARM's shift to v9 architecture could accelerate royalty rates even without new agentic revenue. The ByteDance tie-in introduces China exposure risk that export controls could disrupt faster than AI capex stalls. This concentration in two customers for the $15B bet creates binary outcomes not priced into current valuations.
"Regulatory/geopolitical policy risk could derail ARM's ambitious agentic AI ramp, pressuring multiples even in a solid quarter."
Responding to Grok: The ByteDance/Oracle tie-up is indeed a tail-risk, but the bigger, underappreciated risk is regulatory and geopolitical headwinds that could abruptly choke ARM's $15B agentic AI CPU ramp. Export controls, sanctions, or supply-chain restrictions could turn 'partners scaling' into 'partners pausing' and trigger meaningful multiple compression even if Q4 beats hide the deceleration. One-off licensing deals are not immune to policy risk.
"Hyperscaler adoption of custom RISC-V silicon poses a greater long-term threat to ARM's revenue than geopolitical export controls."
ChatGPT and Grok focus heavily on export controls, but they miss the real structural threat: internal silicon development by hyperscalers. If Oracle or ByteDance successfully pivot to proprietary RISC-V designs, ARM’s 'moat' evaporates regardless of AI demand. The $15B target assumes ARM remains the standard, but the industry is actively incentivized to replace them. We are overestimating the stickiness of ARM’s architecture in a world where custom silicon is becoming a primary competitive differentiator.
"Custom silicon is a margin play, not an architectural replacement—ARM's moat survives unless royalty rates become prohibitively expensive relative to in-house R&D."
Gemini's RISC-V threat is real, but underspecified. ByteDance and Oracle have *incentive* to build custom silicon; they lack ARM's 20-year architectural moat and ecosystem lock-in across billions of devices. Custom ASICs win on margin, not on replacing an entire instruction set. The stickiness question isn't whether they *can* defect—it's whether the capex and time-to-market justify it versus licensing ARM's v9. That calculus shifts only if ARM's royalty rates spike sharply.
Panelists are divided on ARM's future, with concerns about execution, competition, and regulatory risks outweighing optimism about AI demand and licensing moat.
Potential upside from AI demand and licensing moat
Internal silicon development by hyperscalers and regulatory/export control headwinds