AI Processor IP + software stack built to keep pace with your model roadmap.
One IP foundation
Inference. 1 – 864 TOPS.
One architecture
Low latency. Power efficient.
One toolchain
Any chip. Your models.
