MIPS Launches Three Computing Platforms for Physical AI

This page was machine-translated and may differ from the original. View original

MIPS Launches Three Computing Platforms for Physical AI

Published2026.09.03 15:30

Workload Optimization by AI Inference, Real-Time Control, and Safety Systems

MIPS, a GlobalFoundries subsidiary, has unveiled three computing platforms required for implementing AI in physical environments such as robots, automobiles, and industrial machinery. Each platform is characterized by independently optimized computing capabilities according to workload characteristics including AI inference, real-time control, and high-reliability safety systems.

On the 3rd, MIPS announced the launch of three Physical AI development platforms in San Jose, California: 'MIPS Acies™' for AI inference, 'MIPS Actus™' for real-time control, and 'MIPS Aegis™' for high-reliability and safety systems. MIPS describes this approach as 'Workload-Native.'

Sameer Watson, MIPS CEO, stated, "As AI spreads across diverse physical systems, computing platforms must also evolve to match each workload's characteristics," adding "We are providing open, software-first platforms that enable customers to develop and deploy Physical AI on a common foundation."

MIPS Acies™ is a development platform for AI inference in edge environments, featuring open standards-based NPU and open software stack.

MIPS Actus™ is a real-time platform for embedded systems requiring rapid response such as motion control, sensor fusion, and predictive maintenance.

MIPS Aegis™ targets mission-critical systems requiring low latency and autonomous AI decision-making, and is suited for environments demanding high safety and reliability.

MIPS adopted a software-requirement-first approach in platform design.

Based on open standards technology and advanced workload modeling, it supports developers in analyzing actual processing tasks and configuring systems to meet their needs.

Additionally, combined with GF's process technology and manufacturing capabilities, it enables the implementation of designed computing architectures into mass-producible semiconductors.

Manuel Uhm, AMD Embedded Silicon Marketing Director, stated "We welcome efforts to strengthen the software utilization environment, enabling diverse customers to leverage the advantages of open standards architecture."

David Ku, MediaTek Co-Chief Operating Officer and Chief Financial Officer, also stated "We welcome contributing to the advancement of RISC-V solutions based on MIPS's open standards-based approach and proven processor and software expertise."

#RISC-V #NPU #AI 추론 #피지컬 AI #엣지 컴퓨팅 #임베디드 시스템 #실시간 제어 #Physical AI #MIPS

To request a correction, reply or follow-up report on this article, see how to file a request. Previously published statements are collected in corrections & replies.