Athos Silicon mSoC™ | A New Compute Architecture for Physical AI
Athos Silicon's mSoC™ introduces a new compute architecture for Physical AI and agentic AI.
Instead of adapting legacy PC and mobile designs, mSoC is built around fault containment from first principles: distributed chiptiles, hardware voting, redundant power, local high-bandwidth flash for model storage, and enforced communication boundaries. The result is a system that continues operating under failure instead of relying on centralized control.
This video uses autonomous driving as a reference point, but the architecture is designed for the entire Physical AI landscape, including robotics, drones, and real-world systems that interact with humans and require deterministic behavior.
Related Chiplet
- Optical Chiplet
- Reverse Gearbox Chiplet
- FlexGen Multi-Die Smart Network-on-Chip (NoC) IP
- Ncore Multi-Die Interconnect IP
- Integrated voltage regulator (IVR) chiplet
Related Videos
- Design and Assembly of an Automotive-Grade Chiplet-Based Multiple Systems-on-Chip
- Athos Silicon’s Safety-First Chiplets for Autonomy
- Getting Moore with Less: How Chiplets and Open Interconnect Accelerate Cloud-Optimized AI Silicon
- Alphawave and Keysight: UCIe Compliance Validation Journey from System Simulation to Silicon
Latest Videos
- Scaling up AI Compute, Memory and Connectivity
- CURIE: A Cycle-Accurate and Synthesizable UCIe Interface for the Agile Design and Co-Simulation of Chiplets
- Federated Chiplet Design and Simulation
- Custom HBM as a System: Co-designing the Base Die, Die to Die Power, and Package Thermals for AI Accelerators
- Bunch of Wires Flexi 1.0: a low-cost, low power D2D interface