OK. I’ll rewrite the provided content as a new,evergreen news article for a target site,following all the instructions regarding paraphrasing,quotations,ad placements,adn enhancement modules.
Arm Unveils Lumex Platform for On-Device AI Acceleration
CAMBRIDGE,UK – Arm has announced its new Lumex compute subsystem (CSS) platform,designed to accelerate AI experiences on flagship smartphones and next-generation PCs. The platform integrates high-performance CPUs with Scalable Matrix Extension version 2 (SME2), gpus, and system IP.
Lumex aims to enable faster growth of AI devices and deliver enhanced experiences,from mobile gaming to real-time translation and personalized applications. arm plans to integrate SME2 across its CPU platforms,projecting over 10 billion TOPS of compute across more than 3 billion devices by 2030.Partners can integrate Lumex into their SoCs, using either the complete platform or configuring the RTL for specific tiers. The platform includes:
Next-generation SME2-enabled Armv9.3 CPU cluster, featuring C1-Ultra and C1-Pro CPUs.
New C1-Premium CPU, designed for the sub-flagship market.
New Mali G1-Ultra GPU with ray tracing.
C1-DSU DynamIQ Shared Unit.
Optimized physical implementations for 3nm nodes.
KleidiAI libraries for AI acceleration.
The SME2-enabled Arm C1 CPU cluster offers AI performance gains for AI-driven tasks, including up to 5x uplift in AI performance, 4.7x lower latency for speech-based workloads, and 2.8x faster audio generation.
This performance enables real-time,on-device AI inference capabilities for smoother experiences across audio generation,computer vision,and contextual assistants.A Smart Yoga Tutor demo app saw a 2.4x boost in text-to-speech, and collaboration with Alipay and vivo resulted in a 40% reduction in LLM response time. Neural camera denoising can now run at over 120fps in 1080p or 30fps in 4K on a single core.
Lumex brings intelligence directly to the device, offering faster, safer, and always-available AI, unlike cloud-first AI.
“The validation of LLM inference using SME2 has been completed on vivo’s next generation flagship smartphone… We observe that prefill and decode performance can be improved by over 40% and 25% respectively.”
Lumex offers architectural freedom for various product tiers:
| CPU | Key benefit | Performance and efficiency gains | Ideal use cases |
| ———- | ——————————– | ————————————————————- | ——————————————————————- |
| C1-Ultra | Flagship peak performance | +25% single-thread performance, Double-digit IPC gain year-on-year | Large-model inference, computational photography, content creation, generative AI |
| C1-Premium | C1-Ultra performance with greater area efficiency | 35% smaller area than C1-Ultra | sub-flagship mobile segments, voice assistants, multitasking |
| C1-Pro | Sustained efficiency | +16% sustained performance | video playback, streaming inference |
| C1-nano | Extremely power-efficient | +26% efficiency, using less area | wearables, smallest form factors |
The new Arm Mali G1-Ultra GPU enhances mobile gaming with console-class graphics, featuring a Ray Tracing Unit v2 (RTUv2) that delivers a 2x uplift in ray tracing performance. The G1-Ultra also enables up to 20% faster inference performance for AI workloads.
The Mali G1-Ultra delivers 20% better performance across graphics benchmarks, with improvements for titles like Arena Breakout, Fortnite, Genshin Impact, and Honkai Star Rail.
For developers, the KleidiAI integration across frameworks like PyTorch ExecuTorch, Google LiteRT, Alibaba MNN, and Microsoft ONNX Runtime allows apps to benefit from SME2 acceleration without code changes. Google apps like Gmail, YouTube, and google Photos are already SME2-ready. Optimizations built for Android can extend to Windows on Arm and other platforms.
“Through deep integration with SME2, MNN enables low-latency, quantized inference for billion-parameter models like Qwen on smartphones – showcasing Arm and Alibaba’s joint innovation in scalable, next-gen mobile AI,” said Xiaotang Jiang, Head of MNN, Taobao and tmall Group, Alibaba.
Nak Hee Seong, Vice President and Head of SOC IP Development Team at Samsung Electronics stated, “At Samsung, we’re excited to continue our collaboration with Arm by leveraging Arm’s compute subsystem platform to develop the next generation of flagship mobile products. This partnership enables us to push the boundaries of on-device AI, delivering smarter, faster, and more efficient experiences for our users.”
Iliyan Malchev, Distinguished Software Engineer, Android at Google added, “SME2-enhanced hardware enables more advanced AI models, like Gemma 3, to run directly on a wide range of devices. As SME2 continues to scale, it will enable mobile developers to seamlessly deploy the next generation of AI features across ecosystems.This will ultimately benefit end-users with low-latency experiences that are widely available on their smartphones.”
Arm Lumex aims to provide the tools for delivering personal, private, and high-performance AI at the edge.
About Arm
Arm technology is at the heart of a computing and data revolution. Arm’s advanced technology is enabling new markets and transforming societies and culture.
