Arm Holdings posted a strong second quarter for fiscal 2026, with revenue rising 34% year-over-year to $1.14 billion — its third consecutive billion-dollar quarter. Royalty revenue grew 21% to a record $620 million, while licensing surged 56% to $515 million, driven by new high-value agreements for next-generation compute technologies. Non-GAAP operating income reached $467 million, and non-GAAP EPS was $0.39, both exceeding guidance. Arm also posted $411 million in free cash flow and ended the quarter with $3.3 billion in cash and short-term investments .
The results highlight Arm’s central role in powering AI—from hyperscale data centers to smartphones and vehicles. Over 1 billion Arm Neoverse CPUs have now been deployed across hyperscalers including AWS, Google, Microsoft, and Meta, which are scaling Arm-based AI infrastructure. AWS Graviton 4 and Google Axion chips demonstrate strong efficiency gains, while Microsoft’s Cobalt CPUs now operate in 29 regions. Arm expects its CPU share among top hyperscalers to approach 50% by year-end. At the edge, new Arm-based devices such as Google’s Tensor G5 in the Pixel 10 and NVIDIA’s DGX Spark developer system extend AI capabilities from cloud to desktop .
Arm’s Compute Subsystems (CSS) also gained traction, with three new licenses signed in Q2, spanning smartphones, tablets, and data centers. The company introduced its Lumex CSS platform for smartphones, delivering up to 5x faster AI performance on CPU and 3x energy efficiency over the prior generation. Ecosystem partners such as MediaTek, OPPO, and vivo have adopted Lumex for flagship devices. Meanwhile, Tesla and leading automakers expanded use of Arm for vehicle intelligence and AI-driven safety systems, with the new Tesla AI5 chip delivering 40x faster AI performance than its predecessor. Across smartphones, vehicles, and IoT, Arm’s AI architecture is becoming the foundation for what it calls “AI everywhere.”
• Q2 revenue $1.14 billion, up 34% Y/Y
• Royalty revenue $620 million (+21% Y/Y); licensing $515 million (+56% Y/Y)
• Non-GAAP operating margin 41.1%; EPS $0.39 vs. $0.30 last year
• Compute Subsystem licenses total 19 across 11 companies
• Armv9 adoption continues to lift average royalty per chip
• Over 1 billion Arm Neoverse cores now deployed in AI and hyperscaler infrastructure
“Our business continues to demonstrate the strength of Arm’s platform in the AI era,” said Rene Haas, CEO of Arm. “As workload complexity accelerates with every new model and every new agent, Arm is delivering high-performance, power-efficient AI everywhere.”
🌐 Analysis: Arm’s results confirm its evolution from a mobile IP licensor into a key infrastructure provider for AI compute. Its growing footprint across hyperscalers—AWS, Google, Microsoft, Meta, and NVIDIA—positions it as the architectural backbone for cloud-to-edge AI. The rapid uptake of Armv9 and CSS platforms signals a deeper move into system-level design, while the Meta and Tesla partnerships extend its reach into AI hardware co-design. Competitors such as Intel and AMD are now racing to match Arm’s energy efficiency as the industry shifts to AI-optimized heterogeneous compute.
🌐 We’re tracking the latest developments in semiconductors. Follow our ongoing coverage at: https://convergedigest.com/category/semiconductors/
