Arm Holdings Is Expensive, But Not Fully Priced For 2031

Arm Holdings (ARM) was rated a Buy with a $544 price target versus $332 (about 64% upside). The bull case centers on rising royalty dollars per chip tied to Armv9/CSS, improved licensing momentum, and increased CPU demand from agentic AI/Arm AGI CPU. The analyst anchors valuation on a longer-dated EPS view, projecting non-GAAP EPS of $3.01 in 2028 growing to $8.59 by 2031.

Analysis

The durable upside is less about near-term unit growth and more about who captures the economics of custom silicon. ARM is structurally exposed to a favorable mix shift in AI-adjacent CPU content, but the largest second-order beneficiary may be hyperscalers and OEMs that use ARM to lower their own compute cost base; that caps ARM’s pricing power if customers decide the architecture is strategic enough to renegotiate harder. The more successful the custom-chip trend becomes, the more the value accrues to the chip buyer rather than the IP owner.

The market’s immediate reaction can be supportive, but the fundamental verification window is 1-3 earnings cycles, not the next week. What matters is whether licensing momentum turns into higher royalty take-rate, not just more design wins; if AI workloads remain accelerator-led, CPU demand may rise slower than the valuation bridge implies. RISC-V is the underappreciated long-dated substitution risk, especially in edge and embedded where cost sensitivity matters more than software portability.

Consensus may be overcalling the duration of the AI CPU cycle and undercalling customer bargaining power. The 2031 EPS framing is useful for a model, but it is a weak anchor for public-market multiples today; one soft quarter in licensing or royalty growth could compress the stock quickly because the bull case is back-end loaded. TGT has essentially no read-through.

Key Decisions for Investors

• Do not chase the first move higher; wait for a 3%-5% pullback or the next quarterly print to confirm royalty acceleration before adding exposure.
• For a 6-12 month expression, use a long ARM call spread instead of outright common stock to reduce valuation risk; the thesis works only if licensing converts into visible royalty leverage.
• Relative-value trade: long ARM / short INTC for a 6-18 month secular architecture-shift view; thesis is invalidated if x86 share stabilizes or ARM licensing growth stalls.
• Set a watch item on next earnings: if licensing growth and data-center commentary fail to inflect, fade the multiple expansion rather than averaging down.