ARM Holdings PLC Earnings - Q4 2025 Analysis & Highlights
ARM Holdings delivered record Q3 2026 results driven by strong royalty growth across AI and data center segments, with management emphasizing the company's positioning as the compute platform for agentic AI workloads and outlining ambitious growth strategies across edge, physical, and cloud AI markets.
Key Financial Results
Total revenue grew 26% year-over-year to a record $1.24 billion, marking the fourth consecutive quarter above $1 billion.
Royalty revenue reached a record $737 million, up 27% year-over-year, driven by record units with strength across AI and general purpose data center.
Licensing revenue was $505 million, up 25% year-over-year, as more leading companies signed high value licenses for next-generation technologies.
Non-GAAP operating income was $505 million, up 14% year-over-year, resulting in a non-GAAP operating margin of approximately 41%.
Non-GAAP EPS was $0.43, close to the high end of guidance range, driven by both higher revenue and slightly lower operating expenses than expected.
Business Segment Results
Data center royalty revenue has grown more than 100% year-over-year, with management expecting data center to become the largest business within a few years, larger than mobile.
Data center royalty revenue continues to double year-over-year, with the ramp of Arm-based chips by all major hyperscaler companies.
Royalty revenue from edge AI devices such as smartphones continues to grow much faster than the market, with all major Android OEMs now ramping smartphones with chips based on both Armv9 and Compute Subsystems (CSS).
Automotive market grew double-digits year-over-year and contributed to strong royalty performance in the physical AI segment.
SoftBank Technology Licensing and Design Services contributed $200 million of the $505 million in license revenue.
Annualized Contract Value (ACV) grew 28% year-over-year, maintaining strong momentum and continuing to be above long-term expectations of mid to high single-digit growth for license revenue.
Capital Allocation
Non-GAAP operating expenses were $716 million, up 37% year-over-year due to strong R&D investment.
R&D investments reflect ongoing engineering head count expansion to support customer demand for more Arm technology, including innovation in next-generation architectures, compute subsystems, and exploration into chiplets and complete SoCs.
Industry Trends and Dynamics
AI is changing how compute is built and where it runs across cloud infrastructure, edge devices, and physical systems.
The shift towards inference is reshaping data center design, with increasingly agent-based inference workloads that are persistent, always-on, and power constrained.
Agent-based AI requires coordination across many agents running continuously, and the CPU can only do coordination, requiring CPU chips with higher core counts and better power efficiency.
Neoverse CPUs have surpassed 1 billion cores deployed, and Arm share amongst the top hyperscalers is expected to reach 50%.
Leading hyperscalers are launching new products with increased core counts, including AWS's fifth-generation Graviton processor with 192 cores (doubling from Graviton4), NVIDIA's Vera CPU with 88 Arm-based cores (up from 72), Microsoft's Cobalt 200 with 132 cores (up from 128), and Google's second Arm-based server processor with Axion.
Google has migrated over 30,000 applications to the Arm instruction set.
More integrated platform designs are improving system efficiency, with AWS integrating Graviton with Arm-based Nitro DPUs and Trainium accelerators, and NVIDIA pairing GPUs with Arm-based Grace CPUs.
AI is now moving to everyday devices, opening up new growth opportunities in edge and physical AI markets.
Competitive Landscape
Only Arm's compute platform can address the demands of supporting AI workloads ranging from milliwatts to gigawatts across cloud infrastructure, edge devices, and physical systems.
Arm's strength in power efficiency, predictable latency, and always-on operation are best suited to on-device agents that continually monitor inputs and invoke models when needed to preserve battery life.
Arm's common software foundation across devices, vehicles, and robotics lets customers scale deployments without rebuilding software stacks.
Rivian announced its third-generation Autonomy Computer based on the Arm-based Rivian Autonomy Processor, the first production vehicle based on a custom Arm chip and the first to deploy Armv9 in a production car.
Tesla's upcoming Optimus humanoid robot is powered by a custom Arm-based AI processor.
Leading silicon providers like NVIDIA's Jetson Thor and Qualcomm's Dragonwing platforms are scaling Arm-based solutions across robotics and autonomous systems.
Arm provides the foundation for AI across all power envelopes, with a developer ecosystem of over 22 million developers, more than 80% of the global total.
Macroeconomic Environment
MediaTek discussed approximately 15% reduction in unit volume for next year, consistent with what other smartphone and handset providers expect from memory supply chain constraints.
Partners are trying to protect the high end of the market, where premium and flagship portions have the highest CSS and Armv9 royalties.
Most supply chain constraints will be felt at the bottom end of the segment, where Armv8 and older generations with dramatically smaller royalties are located.
A 20% reduction in smartphone volumes would translate to approximately 2% to 4% impact on smartphone royalties, and 1% to 2% negative impact on total royalties.
Growth Opportunities and Strategies
Compute Subsystems (CSS) demand continues to exceed expectations, with 21 CSS licenses across 12 companies and five customers now shipping CSS-based chips.
Two additional CSS licenses were signed this quarter for edge AI tablets and smartphones.
Two CSS customers are shipping a second generation platform, and the top four Android smartphone vendors are shipping CSS power devices.
CSS helps customers get to market faster by lowering integration risk and complexity, and as demand scales, it increases the value that Arm delivers per chip.
CSS royalty revenue has grown from approaching double-digit to well into the teens, with expectations to reach upwards of 50% over the next couple of years.
Arm is organized around three business units: edge AI, physical AI, and cloud AI, aligning with how customers deploy AI.
Edge AI comprises the smartphone and IoT businesses, physical AI includes automotive and robotics, and cloud AI encompasses data center and networking.
Arm is investing in next-generation architectures, compute subsystems, and silicon needed to enable higher performance, greater efficiency, and more AI use cases.
Arm is exploring chiplets and complete SoCs as part of its innovation strategy.
Arm is highly involved in SRAM and alternative memory technologies to address increasing demands on AI.
Arm is hosting an event on March 24th with plans to announce new offerings, though specific details were not provided.
Financial Guidance and Outlook
Q4 2026 revenue is expected to be $1.47 billion, plus or minus $50 million, representing approximately 18% year-over-year growth at the midpoint.
Q4 royalties are expected to be up low-teens year-over-year and licensing to be up high-teens year-over-year.
Non-GAAP operating expenses are expected to be approximately $745 million for Q4.
Non-GAAP EPS for Q4 is expected to be $0.58 plus or minus $0.04.
Full year 2026 revenue growth is expected to be approximately 22% at the midpoint, exceeding the previously stated target of at least 20%.
For fiscal 2027, the 20% growth rate is considered very reasonable at a high level, though full year guidance was not provided.
Full year royalties for next year are expected to be in the north of 20% range, which is consistent with expectations from earlier in the year.
The strength of customer demand combined with a growing base of long duration contracts at structurally higher royalty rates provides increasing confidence in future revenue profile.
Cloud AI or infrastructure business has been growing ahead of expectations and is more than compensating for risks on the memory and mobile side.
