All-In-One
AI Compute Platform
AI-Optimized RISC-V Computing
The All-In-One chip architecture integrates Atrevido CPU, Vector Unit, Tensor Unit, and Gazzillion Misses™ for scalable, programmable AI processing
A breakthrough memory management technology that eliminates bottlenecks and maximizes throughput for AI workloads
Scalable AI Processing Without Limits
The AIO AI Compute Platform is the next evolution in AI acceleration, integrating multiple functions into a unified, zero-latency processing system. Built on RISC-V , this All-In-One solution eliminates the inefficiencies of traditional AI architectures, providing seamless scalability from edge AI devices to massive datacenter AI processors
Tensor Unit
The Tensor Unit handles matrix operations essential for AI, seamlessly integrating with the Vector Unit
Vector Unit
The Vector Unit executes multiple data elements simultaneously, delivering high-throughput parallel compute for AI workloads
Atrevido CPU
A high-performance CPU core designed for demanding AI applications
Gazzillion Misses™
A breakthrough memory management technology that eliminates bottlenecks and maximizes throughput for AI workloads
Key Features
Scalable, Limitless AI Processing
Fully Integrated AI Compute
Unifies CPU, Vector, Tensor, and Memory Acceleration for optimal performance
Scales from 1 to 100s of TOPS
Configure the ideal mix of cores for your workload
Single, Unified RISC-V Software Stack
Avoid vendor lock-in with fully programmable AI acceleration
Zero-Latency Data Flow
No bus congestion, no bottlenecks—just direct communication between all processing units
Redefining AI Hardware for Unlimited Innovation
Unlike conventional AI chips that rely on fixed-function accelerators, AIO is built for full programmability, ensuring it adapts to evolving AI models through software updates.
More Efficient
Eliminates wasted compute cycles with zero-latency interconnects
More Scalable
Customizable configurations allow AI processing from small edge devices to high-performance AI clusters
More Flexible
Tailor the CPU-to-Vector-to-Tensor ratio to balance performance and power efficiency
Instead of using rigid fixed-function accelerators, AIO enables fully flexible AI compute that adapts to future AI breakthroughs. By leveraging our Configurator tool, you can fine-tune the right balance of CPUs, Vector Units, and Tensor Units to achieve optimal PPA (Performance, Power, and Area)