SkyEdge
Edge AI SoC chip design and validation. An integrated RTL candidate exists; broader verification and implementation preparation remain open.
Chip design in progressDevelopment status →ARCEXA / Research
Research directions grounded in stated program progress.
Our work spans semiconductor architecture, processor and SoC design, edge AI, verification, and engineering software. This page distinguishes current development from documented baselines and future research.
Current Development
These three programs are the current development focus. No ARCEXA product is commercially available.
Edge AI SoC chip design and validation. An integrated RTL candidate exists; broader verification and implementation preparation remain open.
Chip design in progressDevelopment status →Vision software for edge AI applications. Development is in progress; no released product is claimed.
Software developmentEDA and engineering tool platform. Development has reached an MVP stage; no commercial release is claimed.
MVP stageTechnical Scope
Exploring compute, memory movement, interconnect, and power-aware design choices for future silicon.
Research directionSkyEdge integrates edge AI processing at SoC level. Other processor concepts remain at their stated baseline or planning stages.
SkyEdge in developmentYazhi-A1 and Skavy-E1 are documented RISC-V CPU baselines with unresolved RTL, integration, or verification tasks.
Baselines documentedArchitecture direction for specialized AI, NPU, and ML workloads. A dedicated accelerator implementation is not claimed here.
Portfolio conceptFunctional simulation and regression are reported for named baselines. Coverage closure, formal, FPGA, and synthesis work are program-dependent and remain open where stated.
Partial evidenceAtlas is at MVP stage. The broader tools portfolio is described on the Vision page without implying commercial availability.
Atlas MVPResearch Directions
Research in progress
ARCEXA describes ongoing research into event-driven computing and neuromorphic architectures. This is research activity, not a released processor.
Explore R&D programs →Planned or exploratory
AI accelerator concepts, DevX Core, network-on-chip studies, quantum research, and research silicon ideas are portfolio directions. Their implementation and milestone status must be evaluated individually.
Explore research roadmap →Evidence & Limits
Published summaries
Not claimed