Compute Subsystems Enhance AI Platforms

MIPS has introduced the MIPS Atlas portfolio of compute subsystems, specifically tailored to empower leading automotive, industrial, and embedded technology companies in deploying safe, secure, and efficient Physical AI at the edge. The Atlas portfolio is meticulously designed to offer turn-key enablement for the three fundamental categories of computing that constitute physical AI – Sense, Think, and Act. Physical AI platforms rely on a diverse array of sensors to interpret their environment, generating crucial data that needs to be seamlessly transferred, integrated, and processed in real-time.

This data is then processed by the platform’s embedded AI engine to facilitate swift, private decision-making for safe and precise actions. The adept control of motors and actuators is achieved through real-time compute platforms capable of remarkably low-latency control-loop processing in robotic movements. “The need for efficient autonomous platforms to drive next-generation driverless vehicles, factory automation, and various other applications perfectly aligns with the MIPS Atlas portfolio,” stated Sameer Wasson, MIPS CEO.

Wasson further emphasized, “Our core competencies in safety, efficient data processing, and experience in autonomy have enabled us to expand our portfolio with real-time intelligence, which forms the essential tech stack for Physical AI platforms. MIPS customers can leverage our compute subsystems with software stacks as a turnkey solution to construct physical AI platforms.” By merging high-performance real-time computing with functional safety and edge deployment of post-generative AI models, the Atlas product suite facilitates the development of next-generation autonomous platforms to tap into the vast AI market opportunity.

The new MIPS Atlas portfolio includes the M8500, I8600, and S8200 compute subsystems, serving as application-specific turnkey enablers. With the support of the MIPS Atlas Explorer software platform, customers can adopt a shift-left methodology to expedite the development and deployment of their applications using MIPS Atlas subsystems in a digital twin universe, ultimately leading to faster time to market.

Selected customers will have the opportunity to evaluate the MIPS M8500 real-time compute subsystem with Atlas Explorer in mid-2025, followed by evaluation boards in 4Q25, and reference silicon platforms becoming available in 1H26. A MIPS automotive platform featuring M8500 is anticipated to enter production in 2027. “The MIPS Atlas portfolio marks a significant advancement in enabling Physical AI at the edge, providing the essential real-time compute power required for autonomous platforms to excel in industrial and automotive applications,” remarked Steven Dickens, CEO and Principal Analyst at HyperFRAME Research.