MIPS Unveils Atlas Chip Designs for Advanced Physical AI Applications
MIPS is set to introduce its innovative Atlas chip designs tailored for physical AI applications such as industrial automation and autonomous vehicles.
Enhancing Real-Time Intelligence
The primary goal is to integrate real-time intelligence into physical AI systems, according to Sameer Wasson, the CEO of MIPS, in a conversation with GamesBeat.
The new MIPS Atlas suite offers computational subsystems that enable autonomous edge devices to perceive their environment, process information, and execute actions with remarkable accuracy. This breakthrough aims to spur advancements across the burgeoning sectors of physical AI relevant to industrial robotics and self-driving platforms.
Targeting Key Industries
Based in San Jose, California, MIPS has strategically directed its Atlas portfolio at automotive manufacturers along with industrial and embedded tech companies. This initiative enables these firms to implement reliable, secure, and efficient physical AI solutions at the edge. By merging high-performance real-time computing capabilities with safety features as well as deploying post-generative AI models at the edge, the Atlas suite supports the creation of next-generation autonomous systems targeting an estimated $1 trillion market potential in physical AI.
A Brief History of MIPS
MIPS’ journey has seen significant changes; it was sold by Imagination Technologies to Tallwood Venture Capital back in 2017. A year later, Wave Computing acquired MIPS but faced bankruptcy by 2020. Following this setback, MIPS announced a transition from its traditional architecture toward RISC-V designs and began rolling out its RISC-V chip offerings starting in 2022.
Synthetic Data’s Role in Accelerating Physical AI
At CES 2025, Jensen Huang from Nvidia emphasized that synthetic data would significantly hasten developments within physical AI realms such as self-driving vehicles and robotics. Utilizing simulations allows developers to test parameters closely mimicking reality before deploying products across various scenarios—this hybrid approach accelerates product validation timelines dramatically.
Differentiated Goals: Control Systems vs. General-Purpose AI
Huang made it clear that while Nvidia may focus on creating comprehensive general-purpose artificial intelligence systems—termed ‘AI brains’—MIPS concentrates on developing integral control systems embedded with specialized real-time controllers aimed at specific applications expected for release by 2026.
A Focused Vision for Autonomous Platforms
“The pressing demand for cutting-edge autonomous solutions designed for revolutionary driverless cars or factory automation aligns perfectly with our objectives surrounding the MIPS Atlas portfolio,” commented Wasson. “Our expertise encompasses functionality around efficiency alongside proficiency within safety technology—which collectively bolster our capability offering essential components necessary for robust Physical AI platforms,” he concluded.
Simplifying Complex Computing Needs through Three Pillars: Sense, Think & Act
James Prior from MIPS elaborated on how they approached solving challenges inherent in physical AI by segregating them into three distinct computational domains: sensing information about environments; processing this data intelligently; enabling effective actions based upon processed insights—a trifecta essential when programming any robotic system today.
The Mechanics Behind Sensing & Decision-Making Affordability
The substructure supporting this framework involves smart sensors which empower robots not only observe surroundings but analyze conditions using trained models yield results computed via simulative learning processes ultimately translating decisions into corresponding motorized movements performed accurately through integrated action mechanisms supported through immersive low-latency compute infrastructures capable of outstanding speed related adjustments occurring during operational sequences!
Tuning Into Requirements Specific To Robotics And Beyond
“Merging robust computing qualifications seamlessly aligns our technology towards meeting unique demands emerging around contemporary robotic innovations,” asserted Jim McGregor from Tirias Research while commenting positively regarding ongoing collaborations occurring between both enterprise leaders exploring intuitive multithreaded architectures alongside optimized functionalities meant curtailing latencies frequently witnessed during deployment experiences,” said McGregor while acknowledging dynamic workloads assessed utilizing frameworks like ~Atlas Explorer~!
This announcement should strengthen awareness about what direction we intend steering both us overall functionality being driven respective innovative efforts dedicated paving thorough foundations rooted competence encapsulating present atmosphere catalyzing futuristic opportunities ahead!” stated Wasson respectively active promoting heightened engagement surrounding ambitious endeavors consolidating purposes directing full spectrum earnest aspirations emphasizing automotive Embedded dimension relative interests focusing requisite inventive dynamism averting generalized sidesteps hindering noteworthy effectiveness synchronizing emphasis optimism recognizes generation pervasive revolution now approaching intersection impending deployment horizons centered significantly arising potentials!