Meta and Arm join forces to boost AI efficiency at scale

Partnership aims to improve AI performance from cloud servers to mobile devices

Meta and semiconductor designer Arm have entered a multi-year partnership to accelerate the development of faster, more energy-efficient AI systems across Meta’s global platforms. The collaboration is expected to enhance AI capabilities not only in large-scale data centers but also on smartphones and other low-power devices.

By combining Meta’s expertise in AI infrastructure and open-source tools with Arm’s reputation for power-efficient chip designs, the companies aim to deliver smarter and more sustainable AI experiences to billions of users worldwide.

Hardware, software, and sustainability

Under the agreement, Meta will integrate Arm’s Neoverse-based processors into its data centers. These chips offer competitive performance with significantly reduced energy consumption compared to traditional CPUs—an important factor as AI workloads continue to grow.

This shift could lead to cost savings and a reduced carbon footprint for Meta, which uses AI extensively across Facebook, Instagram, and its Meta AI assistant for functions such as recommendations, moderation, and chatbots.

Beyond hardware, the partnership also focuses on optimizing Meta’s open-source AI tools—like PyTorch and ExecuTorch—to run more efficiently on Arm architecture. These improvements will allow developers to build AI models that are faster and more battery-friendly, whether operating in the cloud or directly on a device.

AI on the edge

One key component is Arm’s KleidiAI technology, which helps improve AI performance on lightweight, battery-operated devices. This could lead to more responsive voice assistants, real-time filters, and on-device image recognition that drains less power.

Both companies said they plan to share the advancements with the broader developer community, opening the door for more energy-efficient AI applications.

For Meta, this partnership supports its goal to scale AI to more than 3 billion users. For Arm, it’s a move deeper into AI-focused computing beyond its stronghold in smartphones