Arm’s Innovations Propel AI Performance Forward

AI models are rapidly evolving, outpacing hardware capabilities, creating an opportunity for Arm to drive innovation across the compute stack.

Recently, Arm introduced new chip blueprints and software tools designed to enhance smartphones’ ability to efficiently handle AI tasks. Arm didn’t stop there—they also revamped their delivery approach, potentially speeding up adoption.

Arm is advancing its solutions to leverage the benefits of leading-edge process nodes. They unveiled the Arm Compute Subsystems (CSS) for Client, their latest cutting-edge compute solution optimized for AI applications in smartphones and PCs.

This CSS for Client promises a significant performance boost—over 30% increased compute and graphics performance, and an impressive 59% faster AI inference for AI, machine learning, and computer vision workloads.

While Arm’s technology has powered the smartphone revolution, it is also gaining traction in PCs and data centers, where energy efficiency is highly valued. Although smartphones remain Arm’s largest market, supplying IP to giants like Apple, Qualcomm, and MediaTek, the company is broadening its scope.

Arm has launched new CPU designs optimized for AI workloads and new GPUs, along with software tools to simplify the development of chatbots and other AI applications on Arm chips.

The real game changer is how these products are delivered. Traditionally, Arm provided specifications or abstract designs that chipmakers had to translate into physical blueprints—a daunting task involving billions of transistors.

For this latest offering, Arm collaborated with Samsung and TSMC to deliver physical chip blueprints ready for manufacturing, saving significant time.

Samsung’s Jongwook Kye praised the partnership, stating their 3nm process combined with Arm’s CPU solutions meets the soaring demand for generative AI in mobiles through “early and tight collaboration” in the areas of DTCO and PPA maximization, ensuring on-time silicon delivery that meets performance and efficiency demands.

TSMC’s Dan Kochpatcharin, head of the ecosystem and alliance management division, echoed this sentiment, calling the AI-optimized CSS “a prime example” of Arm-TSMC collaboration helping designers push the boundaries of semiconductor innovation for unmatched AI performance and efficiency.

“Together with Arm and our Open Innovation Platform® (OIP) ecosystem partners, we empower our customers to accelerate their AI innovation using the most advanced process technologies and design solutions,” Kochpatcharin emphasized.

Arm’s approach isn’t about competing with customers but enabling faster time-to-market by providing optimized designs for neural processors that deliver cutting-edge AI performance.

As Arm’s Chris Bergey explained, “We’re combining a platform where these accelerators can be very tightly coupled” to customer NPUs.

In essence, Arm offers more refined, “baked” designs that customers can integrate with their own accelerators to swiftly develop powerful AI-driven chips and devices.

Leave a comment