AI Software for a New Age of Development

The pervasive Arm CPU has become a highly effective platform for running common AI inference workloads from cloud to edge and endpoint. It provides optimum power efficiency for nearly every task and code size, and puts AI in the hands of billions of users.


Our software enablement work empowers over 20 million developers to achieve their full AI potential and is focused on simplifying AI deployment, accelerating AI performance with minimal or no effort for developers, and safeguarding AI-based solutions with built-in security and support.

Simplifying AI Deployment Through Software Partnerships

 


Whether leveraging generative AI or sensory data for autonomous functions, developers need to deploy AI-based workloads efficiently. Arm streamlines the entire developer experience, ensuring AI development and deployment across diverse Arm platforms is fast, easy, and frictionless.

We collaborate across critical AI frameworks, ML software providers, cloud service providers, and open-source projects to ensure that leading platforms for AI deployment in the edge or in the cloud just work on Arm.

Hear about one of our AI software partnerships in this video from Meta.

 

Simplified Workflows and Open Standards

A huge component of enabling deployment at pace is our ongoing upstream contributions across the open-source software community, our involvement in leading open standards, and simplified developer workflows via a range of CI/CD, ML Ops, and virtual platforms. 
Learn more about Arm’s support for efficient AI workflows and standards below.

GitHub Actions: Simplifying Arm-based Application Development

Discover how GitHub's partnership with Arm is revolutionizing developer workflows. By integrating Arm Virtual Hardware and native hardware, GitHub's CI/CD pipelines streamline build and test procedures, bringing cloud-native workflows to automotive, IoT, and embedded developers.

Supporting the FP8 Standard to Accelerate Neural Network Model Deployments

Learn how Arm, NVIDIA and Intel are collaborating in support of the Eight-bit floating-point specification, saving energy, performance overheads and accelerating the pace of model deployment.

Deploying Across a Range of Processors with TOSA

Discover how TOSA, the standard enabling applications or frameworks to be more easily deployed across a range of processors and custom hardware, plays a key role in enabling edge deployments on Meta’s ExecuTorch.

Accelerate AI Performance Seamlessly in Every Application

Over the past two decades, we have introduced a stream of new architecture features for efficiently processing and accelerating AI-based applications, including Arm Neon, Arm Helium, Arm Scalable Vector Extensions, and Arm Scalable Metric Extensions.

 

But how does Arm ensure that millions of developers can easily access the latest advancements for AI in every AI framework and in every application? The answer is Arm Kleidi, developer enablement resources for unlocking AI performance – out of the box and everywhere.

 

Arm Kleidi helps accelerate billions of workloads on Arm CPU without developer overhead and through seamless integration into the most popular ML frameworks, such as Mediapipe and PyTorch.

Accelerating AI Performance with Little to No Effort for Developers

Alongside our technologies and collaborations to accelerate AI deployments, we invite AI innovators everywhere to join the Arm AI Developer Community and take advantage of rich developer resources, including detailed tutorials and comprehensive documentation to streamline AI development.

AI at the Edge

Accelerate AI inference deployment and workloads in ExecuTorch and across the broader edge AI software ecosystem.

AI in the Cloud

Leverage AI performance with the Arm power-efficient Neoverse CPUs andcoding best practices, optimized AI libraries, and tools.

Arm Developer Program

The Arm Developer Program connects you to a global network of peers and experts, providing the support and tools you need to succeed.

Safeguarding AI-Based Solutions​

There is mounting pressure from consumers and governments to ensure that AI-based applications and models are sustainable, secure, and ethical. Arm is building a power-efficient architecture for running AI and generative AI everywhere on a platform where developers can simply run their code, confident in consistent standards for hardware-backed security.

Learn more about the inherent features for safeguarding AI below.

Securing the AI-Enabled Software-Defined Vehicle

Scalable software attacks in other markets have exposed significant vulnerabilities. With a billion lines of code expected in AI-enabled software-defined vehicles (SDVs), see how Armv9 can help secure the vehicles of the future.

Arm CCA will Put Confidential Compute in the Hands of Every Developer

The introduction of the Arm Confidential Compute Architecture (Arm CCA), a key feature of Armv9-A, is transforming the trust model of compute environments in every application.

Targeting Security Vulnerabilities in Software with PAC/BTI Technologies

Mitigate up to 70% of security vulnerabilities in software in one simple step, with Arm Pointer Authentication (PAC) and Branch Technology Identification (BTI).

Your AI Journey Starts Here

Whether you're optimizing existing applications or embarking on new AI projects, the Arm platform is designed to support your ambitions. Embrace the future of AI with Arm – where your creativity meets our technology.

Visit Heterogeneous AI

Subscribe to the Latest AI News from Arm

Newsletter Signup