Arm Holdings: In-Line Q4 Expected, but FY27 Debate Intensifies; PT Raised to $191

Core Conclusion

Arm’s FY4Q26 print on May 6 will likely confirm guidance of $1.47bn revenue and $0.58 EPS, with cloud AI driving royalty upside to $708mn (+17% y/y) and licensing buoyed by another ~$200mn from SoftBank. The quarter is a non-event. Investor debate centers on FY27: handset royalty risk from a projected 13% smartphone shipment decline, SoftBank licensing contribution remaining elevated instead of plateauing, and the chip business’s revenue not materializing until 4Q27. We maintain Equal-weight with a $191 price target, reflecting a re-rating (chip business multiple raised to 20x, IP to 45x) that already prices in the CPU story but leaves minimal upside at the current $202 share price.

Evidence Chain

1. Cloud AI drives sequential royalty growth. Arm’s v9/CSS architecture benefits from data center expansion. We model 4Q26 cloud AI royalties at $113mn (+14% q/q, +62% y/y). This lifts total royalties to $708mn, in line with consensus and at the upper end of mid-teens guidance. For FY26, total royalties reach $2.65bn. The FY31 TAM for cloud AI exceeds $1trn, per Arm Everywhere.

2. Handset risk surfaces for FY27. Our AlphaWise survey forecasts CY26 smartphone builds at -13% y/y (Android -15%, Apple -2%), pressured by higher memory costs. Qualcomm’s weak handset guide reinforces this. While Arm’s premium exposure (Apple + premium Android) provides relative resilience, lower volume will create quarterly headwinds from 1Q27. We expect management to flag this on the call.

3. SoftBank licensing shows no sign of plateauing. SoftBank contributed $200mn to licensing in 3Q26, up from $178mn in 2Q26. We model another $200mn in 4Q26. Management previously guided for a plateau, but risk exists that contributions continue to increase in FY27. Total licensing revenue of $762mn in 4Q26 (+51% q/q) is in line. The FY27 licensing trajectory is H2-weighted, typical seasonality.

4. Opex remains elevated but understood. Opex of $1bn in 4Q26 (+25% y/y, -3% q/q) reflects ongoing R&D for the chip transition. Management’s FY26–31 opex CAGR guidance of mid-teens (down from 26% in FY24–26) is already priced. Consensus opex is only slightly lower at $988mn.

Key Divergences and Risks

Risk #1: Handset royalty headwinds materialize more severely. If DRAM shortages persist into FY28, Arm’s royalty growth from smartphones could decelerate beyond our base case. Premium exposure only partially mitigates volume risk.

Risk #2: SoftBank licensing does not plateau. Continued high SoftBank contributions in FY27 would distort licensing revenue quality and create comparison challenges in FY28. Investors may penalize Arm for lack of organic licensing breadth.

Risk #3: Chip business execution delays. Arm’s chip revenue is not expected before 4Q27. Any delay in design wins or production ramp would push the large TAM opportunity further out, raising near-term valuation risk given the stock’s 62% rally in one month.

Risk #4: Rising opex pressure on margins. Adjusted EBIT margin is forecast to fall from 42.9% in FY26 to 39.9% in FY27 as gross margin contracts to 95.8% (from 97.1%). Consensus EBIT margin of 42.1% for FY27 is 223bp above our estimate, suggesting potential for disappointment.

Valuation or Trading Implications

We raise our PT to $191 (from $150) on multiple expansion: 20x for the chip business (previously 15x, now a discount to global chipmakers) and 45x for IP (in line with historical IP valuations). Discounting FY29E at 11% WACC yields the target. At the current share price of ~$202, the stock trades 5.5% above our PT, implying limited upside. The risk-reward is balanced: bull case $335 (bullish AI/cloud adoption and higher royalty rates), bear case $65 (litigation loss, loss of share, no chip contribution). Options-implied probabilities suggest a ~36% chance of reaching above $191 in 12 months. We remain Equal-weight, as the CPU narrative is already priced while near-term earnings headwinds and execution risk provide no catalyst for outperformance.

Appendix

Table: Key Estimates vs Consensus