Arm reaches out to Automotive with new chips, subsystem plans

Arm is set to roll up its sleeves and immerse itself up to its elbows in the automotive market. The company this week announced a new line of automotive chips, along with plans for a new automotive-focused compute subsystem as part of a broader automotive roadmap.

The automotive market, despite some ups and downs related to inventory and other dynamics, continues to be a burgeoning growth segment for semiconductor firms, and it is only expected to get bigger as more electronics, sensors, AI capabilities, and autonomous functionality are added to more vehicles in the coming years. Arm is targeting this opportunity with a chip family for automotive that is based on the Arm v9 CPU architecture enhancement set typically seen in data centers in the form of Neoverse server chips, a move that dovetails with the strategic foresight of sector leaders like Jensen Huang, CEO and co-founder of Nvidia and a major Arm customer, who has envisioned cars becoming more like edge data centers in the future.

Specifically, the new line-up of products includes the Arm Neoverse V3AE, which Dipti Vachani, senior vice president and general manager of Arm’s Automotive Line of Business, said brings the server-class performance of the company’s existing Neoverse technology for the first time to automotive sector to tackle the sector’s growing AI-accelerated autonomous and ADAS workloads.

Arm also unveiled the Arm v9-based Cortex-A processors purpose-built for automotive, including the Cortex-A720AE for a broad range of software-defined vehicle (SDV) applications, and the Cortex-A520AE, which aims to bring power efficiency with functional safety features to scale across automotive use cases.

Also included are the Cortex-R82AE, which delivers 64-bit computing to real-time processing to support increasingly important functional safety features, and the Mali-C720AE configurable image signal processor optimized for computer vision and human vision use cases. Along with this batch of chip designs, Arm also said it is delivering a range of configurable system IP to its  ecosystem to help drive development of scalable, high-performance automotive SoCs.

“As we look at the auto sector, we see the systems are only going to get more complex and new needs for safety will become even greater,” Vachani said. “We need to look at bringing all of this together with these compute components for all our partners and standardize for consistency where it matters.”

Nvidia is already adopting the new technology in its DRIVE Thor automotive system-on-a-chip platform, Arm said. Others adopting it include Marvell, MediaTek, NXP, Renesas, Telechips, Texas Instruments, and more.

“The DRIVE Thor SoC integrates Arm’s next-generation CPU, the Neoverse V3AE, to provide the industry-leading, single-thread CPU performance necessary for intelligent cabin experiences and self-driving capabilities built for safety and security,” said Ashish Karandikar, vice president, hardware engineering at Nvidia.

In addition to the new automotive chip line-up, Vachani outlined plans for multiple Arm Compute Subsystems (CSS) for Automotive, with the first automotive-specific CSS expected to be delivered in 2025. Arm CSS for Automotive will provide pre-integrated and validated configurations of Arm AE IP optimized for performance, power, and area in the firm of virtual prototypes that Vachani said will allow software development to begin sooner than the traditional very linear automotive development cycle has allowed for.

“These solutions will have a huge impact on the developer experience, and will shorten time to market [for OEMs] up to two years,” she added. “The industry can immediately start innovating without having to wait for silicon to be in production.”

Arm is working with cloud partners and others, including Amazon Web Services (AWS), Cadence, Corellium, Siemens, and more, to deliver these virtual platforms and cloud solutions, also in collaboration with Autoware Foundation, BlackBerry QNX, Elektrobit, Kernkonzept, LeddarTech, Mapbox, Sensory, Tata Technologies, TIER IV, Vector, and others.