Introducing TokenProcessor™
Building The Future Of AI Compute
Backed by Leading Investors
The LENZO Approach
Adaptability
Deployability
AI does not live only in data centers. Real-world systems, from robotics and industrial vision to autonomous infrastructure, require hardware that can operate efficiently across diverse environments and product lifecycles.
Long-Term Value
When AI changes faster than hardware refresh cycles, flexibility becomes a business advantage. By reducing the need for redesigns and hardware replacement, TokenProcessor™ helps organizations maximize the return on their AI investments.
LENZO A-Series

Built For Demanding AI Workloads
The A-Series combines the adaptability of TokenProcessor™ with the performance, efficiency, and developer ecosystem required for real-world deployment.
65 TOPS INT8
High-performance AI acceleration for edge and Physical AI workloads.
Up To 64GB LPDDR4
Support larger models and memory-intensive applications.
Linux Environment
Built for familiar development workflows and deployment pipelines.
Framework Support
PyTorch, TensorFlow, ONNX, and other popular AI frameworks.
Use Cases
Built For The Real World
From robotics and industrial automation to autonomous systems and intelligent infrastructure, the A-Series is designed for AI workloads that interact with the physical world.
The Technology
Built On A Different Computing Model
TokenProcessor™ is built on LENZO’s proprietary CPU-Grounded Linear Array (CGLA) architecture. Developed through years of research at NAIST, CGLA combines the flexibility of programmable computing with the efficiency of specialized acceleration, enabling a broader range of computational patterns to be executed on a single platform.
List of Publications
View the full list of publications of the NAIST Computing Architecture Lab, including peer-reviewed, invited-talk & awards, and over 60 patents.
News
Latest From LENZO
Follow product updates, technical insights, research milestones, industry events, and company announcements.






















