MIPS is launching its Atlas chip designs for bodily AI platforms similar to industrial robots and autonomous automobiles.
The purpose is to drive real-time intelligence into bodily AI, stated Sameer Wasson, CEO of MIPS, in an interview with GamesBeat.
The brand new MIPS Atlas product suite delivers compute subsystems that empower autonomous edge options to sense, assume and act with precision, driving innovation throughout the rising bodily AI alternative in industrial robotics and autonomous platform markets.
San Jose, California-based MIPS is focusing on its MIPS Atlas portfolio at automotive, industrial and embedded expertise corporations to allow them to deploy secure, safe and environment friendly bodily AI on the edge. By combining excessive efficiency real-time computing with practical security and edge deployment of submit
generative AI fashions, the Atlas product suite allows the event of subsequent technology autonomous platforms to deal with the estimated $1 trillion bodily AI market alternative.
Creativeness bought MIPS to Tallwood Enterprise Capital in 2017. Then Wave Computing bought MIPS in 2018. However Wave Computing declared chapter in 2020, and MIPS introduced that it will abandon the MIPS structure in favor of RISC-V designs. MIPS began launching its RISC-V chip designs in 2022.
At CES 2025, Nvidia CEO Jensen Huang stated that the age of bodily AI — with self-driving automobiles and robots — can be accelerated due to artificial knowledge. That’s the usage of simulations to check knowledge for issues like self-driving automobiles in digital twins earlier than they’re deployed in the actual world. That artificial knowledge enhances the bodily world knowledge, accelerating the time it takes to totally take a look at a product.
He famous MIPS shouldn’t be competing with Nvidia on constructing AI brains. Relatively, MIPS is constructing management techniques with AI constructed into it for real-time controllers. He stated to count on merchandise on this house in 2026.
“The necessity for environment friendly autonomous platforms to advance next-generation driverless automobiles, manufacturing unit automation, and lots of different purposes is immediately aligned with the MIPS Atlas portfolio,” stated Wasson. “Our core competencies of security, environment friendly knowledge processing and expertise in autonomy have enabled us to broaden our portfolio with real-time intelligence that’s the important tech stack for Bodily AI platforms. MIPS clients can take our compute subsystems with software program stacks as a turnkey resolution to construct bodily AI platforms.”
James Prior, advertising and marketing govt at MIPS stated in an interview with GamesBeat that MIPS is offering a portfolio of options for its method to bodily AI, which entails real-time processing, knowledge motion, customized compute and practical security to allow the world’sindustrial robotics and automation suppliers with what they should construct bodily AI platforms.
“The explanation for that’s we analyzed the bodily AI drawback as three completely different computing points. There’s the sense, assume and act,” he stated.
Sensing helps robots see the world round them an make selections with AI fashions. It computes options to issues based mostly on fashions which might be educated in simulation, after which it turned that call into motion for robots to take.
The MIPS Atlas portfolio is constructed to supply turn-key enablement for the three classes of computing that make up bodily AI – Sense, Suppose, and Act. Bodily AI platforms interpret their environment utilizing a various array of sensors, producing knowledge that have to be seamlessly transferred, built-in, and processed in actual time.
This knowledge is processed by the platform’s embedded AI engine to allow quick, non-public determination making for secure and exact actions. Motion, wanted for the dexterous management of motors and actuators, is enabled through real-time compute platforms able to extremely low latency control-loop processing in robotic actions.
“The MIPS Atlas portfolio aligns tightly with the compute necessities for bodily AI and robotics, “ stated Jim McGregor, principal analyst at Tirias Analysis, in a press release. “Working with its clients, MIPS supplies a novel multi-threaded structure with optimized directions to ship event-driven low-latency efficiency together with system-level workload evaluation through the Atlas Explorer. MIPS is in a novel place to fill the necessity for rising bodily AI platforms.”
The brand new MIPS Atlas portfolio of M8500, I8600, and S8200 options ship these capabilities as utility particular turnkey enablers. Enabled by the MIPS Atlas Explorer software program platform, clients can make use of shift-left methodology to develop and deploy their purposes with MIPS Atlas subsystems in a digital twin universe for sooner time to market. Choose clients will have the ability to consider the MIPS M8500 real-time compute subsystem with Atlas Explorer in mid-2025, with analysis boards in 4Q25, and reference silicon platforms obtainable in 1H26. A MIPS buyer automotive platform that includes M8500 is anticipated to begin manufacturing in 2027.
“The MIPS Atlas portfolio represents a major leap ahead in enabling Bodily AI on the edge, delivering the real-time compute energy wanted for autonomous platforms to thrive in industrial and automotive purposes,” stated Steven Dickens, CEO at HyperFRAME Analysis, in a press release. “By integrating security, effectivity and cutting-edge intelligence, MIPS is well-positioned to speed up innovation throughout the quickly increasing $1 trillion Bodily AI market.”
A 40-year-old firm, MIPS is utilizing the RISC-V structure to drive intelligence into motion for automotive and industrial corporations with the MIPS Atlas portfolio of Sense, Suppose and Act by pushing the boundaries of autonomous platforms. MIPS is an exhibitor at Embedded World 2025 from March eleventh – thirteenth, in Nuremburg, Germany.
MIPS additionally introduced quite a lot of different chip design merchandise, together with the MIPS S8200 bodily AI subsystem for bodily AI on the edge. It additionally introduced the MIPS I8600 compute subsystem for sensor knowledge motion, the M8500 real-time compute subsystems and extra.
A brand new identification for MIPS
“This announcement ought to assist construct the identification of what we’re attempting to do MIPS,” Wasson stated. “Whereas our expertise is relevant to a number of markets, the market which we’re solely specializing in with all our power is the automotive embedded facet, the bodily AI facet,” Wasson stated.
Wasson stated that in case you’ve ever seen an autonomous lawnmower, you’ll understand it form of sucks at doing a great job chopping your garden. That’s as a result of the bodily AI drawback hasn’t been solved accurately but. He stated a garden shouldn’t be straight and grass can’t be reduce with one thing like a robotic vacuum. The grass modifications in delicate methods, and it takes actual bodily AI to unravel such issues.
“I’m not changing into a lawnmower firm, nevertheless it’s an instance of the place we’re attempting to construct a system. So once you consider robotics, the identical lawnmower drawback exists throughout the board on a number of robotic assessments. And what we’re attempting to do is carry AI within the closed loop of a management. And if you are able to do that efficiently with the processing capabilities that now we have.”
He famous that MIPS is creating all the constructing blocks of subsystems which might be wanted in autonomous automobiles and bodily robots.
“MIPS has been very IP centered previously,” he stated. “We are actually beginning to broaden into past IP. We’ve constructing blocks for this, however we’re additionally popping out with reference silicon, with the board, with the ecosystem. We’re proving out the ecosystem of the software program facet in order that individuals are able to go and may begin replicating it very simply.”
