The agentic era needs a computing platform everywhere – Arm is building it 

We’re moving from an era dominated by generative AI into an increasingly agentic world. This next era of computing won’t be built in one place – it will form increasingly connected parts of one computing continuum – at the edge, where intelligence is personal; in the physical world, where it takes action, and in the cloud, where it is created. This creates a system-level challenge – workloads are becoming more distributed; models and compute requirements are increasingly diverse; systems are becoming more heterogeneous and autonomous; and software environments vary.

The industry needs a platform that can deliver the consistency developers need with the freedom to differentiate. Arm provides that platform, combining industry-leading performance per watt with the flexibility to choose between IP, compute subsystems or silicon, all underpinned by an unmatched software ecosystem of more than 22 million developers. Today, we unveiled a series of updates that illustrate our continued momentum in delivering intelligence everywhere.

Edge: Where intelligence becomes personal 

At the edge, AI is becoming more continuous, contextual and immersive, placing new demands on the entire mobile compute platform. These experiences aren’t defined by a single AI workload or processor – AI needs to run on the right compute engine for the task. Our new Arm CSS for Mobile 2 platform is designed for that shift, delivering an AI-native compute foundation powered by two new best-in-class compute engines built for the AI era. The Arm C2 Ultra CPU with SME2 brings a new level of responsive, efficient on-device AI to the CPU, while the Arm Mali G2-Ultra NX GPU introduces dedicated neural accelerators that enable neural graphics to run natively on the GPU.

The ecosystem is already building on these new capabilities. SME2 adoption across partners including Alipay, OPPO and vivo, alongside neural graphics collaboration with Infold Games, NetEase, Tencent Games, Unity China and Unreal Engine. Combined with Arm’s software ecosystem, CSS for Mobile 2 gives silicon partners and developers a flexible compute foundation to bring the next generation of AI experiences to mobile at scale.

Read more from Chris Bergey, EVP, Edge AI, on CSS for Mobile 2. 

Physical AI: Where intelligence takes physical action

Agentic AI is moving beyond digital tasks to real-world action in robots, vehicles and autonomous systems, representing trillions of dollars in global economic activity. Enabling intelligent machines to sense, reason and act requires us to navigate complexities spanning AI models, software, compute, sensors and actuators. 

No single company can solve these challenges alone – we must start by aligning the ecosystem around a common foundation. Arm is expanding its proven Arm Total Design program to physical AI, bringing together more than 80 industry leaders including AWS, Hugging Face, Liquid AI, QNX and Unitree Robotics, among others. An initial focus is a Robotics Capability Framework, establishing a common language to describe, compare and communicate the capabilities of increasingly intelligent autonomous machines.

Read more from Drew Henry, EVP, Physical AI, on Arm Total Design for Physical AI and a Robotics Capability Framework.

Cloud: Where intelligence is created

AI infrastructure is increasingly moving to Arm, with adoption across every major hyperscaler and Arm-based rack-scale servers overtaking x86 as the dominant accelerated computing platform, according to IDC. As agentic AI drives more diverse requirements, Neoverse CSS N4 gives silicon partners Arm’s most configurable CSS yet for throughput-efficient scale-out workloads, while Arm AGI CPU provides production-ready, high-performance compute.

Companies including OpenAI, Meta and Oracle are developing solutions around AGI CPU, with ByteDance’s Volcano Engine bringing the first agentic sandboxes powered by the platform to market. Together, CSS N4 and AGI CPU give customers greater choice in building next-generation AI infrastructure on a common Arm Neoverse platform and software ecosystem. 

Read more from Mohamed Awad, EVP, Cloud AI, on Arm AGI CPU and Neoverse CSS N4.

Enabling developers and agents to build everywhere

As development becomes increasingly agentic across cloud, edge and physical devices, agents need consistent access to that ecosystem too. The new Arm AI Portal makes that possible, helping developers and agents discover optimized models, understand performance and access the software they need.

With intelligence everywhere, the industry is converging on Arm

The Arm platform and ecosystem remain central to delivering the consistency, choice and breadth that the agentic world demands. Today’s announcements are further proof of our progress towards that vision, bringing intelligence everywhere on one global computing platform.