Q3 2025 Earnings Summary
- Accelerated demand for Arm's Compute Subsystem Solutions (CSS) is driving significant revenue opportunities, with partner demand being stronger than initially anticipated. This growth is fueled by AI workloads requiring more complex chip designs, leading to increased CSS adoption across all end markets, including both data center and client markets. ,
- Arm's central role in major AI initiatives like Stargate and Cristal intelligence positions the company at the heart of the expanding AI ecosystem. These projects highlight Arm's significance in AI infrastructure deployment and the development of AI agents, driving increased demand for compute from the cloud to the edge and leveraging Arm's efficiency and pervasive platform. ,
- Leadership in the transition to chiplet-based architectures (CSA) offers significant monetization opportunities for Arm. The industry's move towards more complex, heterogeneous system-on-chips is supported by Arm's ecosystem around CSA, enabling the company to capitalize on the shift towards chiplet-based strategies and strengthen its position in the semiconductor value chain. ,
- Stalled adoption of Arm's v9 architecture at 25% over several quarters, which may indicate a slower migration to newer technologies than initially expected. Analysts noted that the percentage contribution of v9 has not increased as anticipated, potentially impacting future growth.
- Uncertainty stemming from an ongoing lawsuit with a major customer, where there is disagreement over contract expiration dates. This could potentially impact future royalty revenues, as Arm assumes it will continue receiving royalties at the same rates but acknowledges the unpredictability of legal outcomes.
- Dependence on Arm China for a significant portion of revenue, which is expected to decrease over time, potentially affecting overall growth. Arm China's contribution was around 25% this quarter but is anticipated to fall to the mid-teens, raising concerns about the company's revenue streams amid geopolitical tensions.
| Metric | Period | Previous Guidance | Current Guidance | Change |
|---|---|---|---|---|
Revenue (Quarterly) | Q4 2025 | no prior guidance | $1.175 billion to $1.275 billion | no prior guidance |
Non-GAAP Operating Expense (Quarterly) | Q4 2025 | no prior guidance | ~$590 million | no prior guidance |
Non-GAAP EPS (Quarterly) | Q4 2025 | no prior guidance | $0.48 to $0.56 | no prior guidance |
Revenue (Annual) | FY 2025 | $3.8 billion to $4.1 billion | Midpoint ~$4 billion | raised |
Non-GAAP Operating Expense (Annual) | FY 2025 | ~$2.05 billion | ~$2.1 billion | raised |
Non-GAAP EPS (Annual) | FY 2025 | $1.45 to $1.65 | $1.56 to $1.64 | raised |
Royalty Revenue Growth (Annual) | FY 2025 | High teens | High teens | no change |
License Revenue Growth (Annual) | FY 2025 | no prior guidance | around 30% year-on-year | no prior guidance |
| Topic | Previous Mentions | Current Period | Trend |
|---|---|---|---|
Stalling or varied adoption of Armv9 architecture across quarters | • Q4 2024: Accelerating v9 adoption, no stalling. <br> • Q1 2025: ~5 percentage point increases, steady progress. <br> • Q2 2025: Reached 25% of royalties but adoption not linear. | Stalled at 25% share but absolute Armv9 dollars grew triple digits. | Bearish near-term (stall) but bullish on eventual 60–70% target |
Expanding AI workloads and Arm’s role in major AI initiatives (Stargate, Cristal) | • Q4 2024, Q1 2025, Q2 2025: Discussed AI broadly, but no mention of Stargate or Cristal. | Debuted in Q3 2025 with Stargate ($100B initial) and Cristal (agentic AI) featuring Arm’s CPUs. | New major AI initiatives in Q3 2025, indicating bullish long-term AI opportunity |
Introduction of chiplet-based architectures (CSA) and Compute Subsystem Solutions (CSS) | • Q4 2024, Q1 2025, Q2 2025: CSS viewed as key growth driver (e.g., Dimensity 9400); CSA not mentioned. | CSA highlighted for heterogeneous designs; CSS demand strong in smartphones, data centers, IoT. | CSA is new in Q3 2025; CSS momentum remains bullish |
Large potential for market share gains in PCs (targeting 50% share) | • Q4 2024: Optimism but no numeric goal. <br> • Q1 2025: 50% share target in ~5 years. <br> • Q2 2025: Reaffirmed 50% ambition. | No mention in Q3 2025. | Still a bullish long-term ambition, but no current update |
Ongoing litigation and legal disputes affecting royalty revenue | • Q2 2025: Qualcomm suit has no impact on revenue assumptions. | Litigation remains no revenue impact, assumptions unchanged. | Neutral; lawsuit continues but no financial downside expected |
Shifts in reliance on Arm China and associated geopolitical considerations | • Q4 2024: Arm China influenced 50% sequential royalty growth but no geopolitical detail. <br> • Q1/Q2 2025: No new shifts. | Arm China ~25% of revenue, expected to decline to mid-teens eventually. | Mild bearish on continued reliance, though share projected to decrease over time |
Broad-based licensing demand but reduced royalty growth guidance over multiple quarters | • Q4 2024: Strong licensing (+60% YoY), but royalty guidance lowered near-term. <br> • Q1 2025: Another cut from mid-20% to low-20%. <br> • Q2 2025: Licensing strong, royalty guide shifts to high teens. | Licensing +14% YoY, royalty +23% YoY, but v9 stall dampens near-term upside. | Bullish on licensing, tempered royalty growth outlook |
Weakness in automotive, IoT, networking, and industrial segments noted in earlier periods | • Q4 2024: Networking, industrial softness, some IoT weakness. <br> • Q1 2025: IoT, networking, industrial remain weak; auto stronger. <br> • Q2 2025: IoT still weak, auto pipeline strong. | IoT shows recovery, automotive & networking meet expectations. | Bearish in prior quarters; now improving in IoT with stable auto & networking |
New product and extension releases leveraging v9 features (security, confidential compute) for data center and AI | • Q4 2024: Google Axion, Neoverse V3 CSS, Grace Blackwell highlight v9. <br> • Q1 2025: Emphasis on security & confidential compute. <br> • Q2 2025: Ongoing expansions but no new specific security mention. | No major new v9 product releases announced; references existing HPC solutions. | Bullish momentum in prior quarters; flat for new releases in Q3 |
Increasing compute complexity driving higher demand for Arm’s IP across diverse end markets | • Q4 2024: AI fueling widespread demand. <br> • Q1 2025: AI data centers driving licensing. <br> • Q2 2025: Broad-based complexity pushes IP uptake. | Emphasized as a key driver (CSS in smartphones, autos, data centers). | Continues bullish as a core growth catalyst across multiple segments |
-
FY '26 Royalty Growth Outlook
Q: What royalty growth and v9 mix is expected for FY '26?
A: Arm expects royalties to grow at a mid-20% rate in FY '26, driven by v9 adoption and CSS. CSS has a royalty rate roughly double that of v9, which is roughly double that of v8. CSS rates vary and increase over time with new solutions. -
Opportunities in AI Projects
Q: What are Arm's opportunities with Stargate and Cristal intelligence?
A: Arm is excited to be the CPU of choice for Stargate, a significant infrastructure project with $100 billion immediate investment and $500 billion over time. This partnership includes OpenAI, Oracle, SoftBank, Microsoft, NVIDIA, and Arm. Additionally, Cristal intelligence represents a significant opportunity as AI workloads will run on all endpoints, from earbuds to data centers, leveraging Arm's pervasive compute platform. Licensing revenue is expected to grow around 60% year-on-year, driven by AI and CSS deals. -
CSS Royalty Contribution
Q: How will CSS impact royalty revenue entering fiscal '26?
A: While not ready to discuss fiscal '26, Arm feels very good about the momentum, with the last quarter royalty revenue at $500 million, about 13% higher than the previous record. This quarter marked the first with material CSS revenue, led by the Dimensity 9400 chip and Cobalt CSS shipments, contributing to royalty revenue. They expect continued momentum over the next couple of quarters. -
v9 Adoption and Royalty Growth
Q: Why is v9 royalty percentage stalling at 25%?
A: Although the v9 percentage remained at 25%, the absolute dollars grew by triple-digit rates. This suggests a longer runway for future growth, and Arm expects v9 to grow to 67%-70% of total royalties. The transition rate is consistent with expectations, driven by OEM product transitions. -
Qualcomm Lawsuit Impact
Q: What's the financial impact of the Qualcomm lawsuit?
A: There's no impact on revenue from the Qualcomm lawsuit. Arm has forecasted, since IPO, assuming they would not prevail in the lawsuit, and from a financial perspective, they assume they will continue receiving royalties at the same rates as in the past. -
Arm China Revenue Outlook
Q: Will related party (Arm China) revenue ramp long term?
A: Related party revenue, mainly from Arm China, is expected to remain consistent in the near term. Currently around 25% of revenue, it's expected to decrease to mid-teens over time, but not change much in the next few quarters. Arm China contributes about 20-25% of ACV. -
CSS Demand Across Markets
Q: Is CSS license activity mostly data center or balanced?
A: CSS license activity is roughly 50-50 between infrastructure (data center) and client business. The chiplet approach is becoming pervasive across SoCs, representing a significant opportunity for Arm. This drives strong CSS demand across all end markets. -
M&A Speculation
Q: Is Arm moving beyond IP model due to M&A?
A: While unable to comment on M&A speculation, Arm is focused on growth and addressing the rapid evolution of AI, which puts pressure on developing products faster and more efficiently. Arm is central to solving these customer problems. -
AI Training Optimization Impact
Q: How does DeepSeek's training optimization affect Arm's AI opportunity?
A: DeepSeek's advancements in training optimization are beneficial, driving efficiency and lowering costs, which expands demand for compute. Lower inference costs make it easier to run AI applications in power-constrained areas, which is terrific for Arm's focus on efficiency. Increased investments by major players indicate growing compute demand, benefiting Arm. -
Licensing Complexity and Timing
Q: Are client agreements taking longer due to complexity?
A: There's no structural change in the complexity or timing of client agreements. New contracts with new partners may take longer due to approvals, especially larger deals, but this is consistent with past experience.