Arm’s vital role in the age of AI from cloud to edge: Five takeaways from the Moor Insights and Strategy report 

Artificial intelligence (AI) is no longer a discrete technology category; it’s now a foundational capability embedded across devices, enterprise workflows, and cloud infrastructure. In this context, the real challenge is no longer just about building more powerful AI models, but deploying and scaling them efficiently in production environments.  

The new From Devices to the Cloud: Arm’s Relevance in the Age of AI report from Moor Insights and Strategy puts Arm at the center of this AI shift, with the Arm compute platform and ecosystem defining how AI is being deployed from cloud to edge. Below are five key takeaways from the market report. 

AI’s true impact is happening at inference – and Arm is built for it 

While much of the focus today is on training large models, the report makes it clear that AI’s economic value will be realized with inference where models are deployed and used at scale across a vast range of real-world use cases. With Arm-based platforms optimized for efficient, scalable inference across different environments, Arm plays a critical role. By prioritizing performance per watt and broad deployability, Arm enables organizations to bring AI inference into production at scale. 

AI is an infrastructure challenge – and Arm addresses it at the system level 

The report notes that the deployment of AI is no longer constrained by algorithms or software innovation, but by infrastructure limitations, such as power, efficiency, latency, and scalability, and system design. Through a relentless focus on efficiency, flexibility, and system-level balance, the Arm compute platform is designed with these real-world constraints in mind. This enables organizations to build infrastructure that can scale AI workloads sustainably. 

The CPU is more important than ever, sitting at the center of AI orchestration 

The CPU remains the central orchestrator layer of modern AI systems, coordinating workloads, memory, and data movement. The report notes that “the more specialized the silicon landscape becomes, the more critical the CPU’s role.” The new Arm AGI CPU – as well as other Arm-based CPUs – are designed to fulfill this role across AI infrastructure. As AI systems become more complex – especially with the rise of agentic workloads – the need for efficient coordination grows, reinforcing Arm’s vital role.  

Arm’s architectural DNA – efficiency, scale, and flexibility – aligns directly with AI’s needs 

The report highlights that Arm’s relevance in the AI era is rooted in our design principles: performance per watt, modularity, and ecosystem enablement. These are directly aligned with the needs of modern AI infrastructure where the core challenge is being able to deliver more compute within fixed power and cooling limits. Hyperscalers, including AWS, Google Cloud, and Microsoft Azure, are already deploying Arm-based CPUs in their platforms.  

The future of AI is distributed – and Arm spans the entire continuum

AI is no longer confined to the cloud, with it now spanning devices, edge systems, enterprise environments, and hyperscale infrastructure – each with different requirements. The report notes that “hybrid AI – distributing intelligence across the cloud and edge – has emerged as a preferred approach.” Against the backdrop of this shift, Arm’s ability to operate across this entire continuum is a key advantage. From consumer devices to autonomous machines to cloud data centers, Arm provides a consistent compute foundation that allows AI workloads to move seamlessly between environments.  

Looking ahead: Arm’s role in the next phase of AI 

As AI evolves to more complex, continuous, and agentic systems, the demands on infrastructure will only increase. Arm’s combination of technology innovation – including new developments like the Arm AGI CPU – and ecosystem enablement positions it as a key foundational platform for this next phase. In an AI era defined by distribution, complexity, and scale, Arm is helping make AI not just more powerful, but deployable everywhere, efficiently.