Licensable NPU IP for edge AI inference in power-constrained endpoint SoCs.
Arm licenses the Ethos-U85 neural processing unit IP to chip designers building endpoint AI silicon. The microarchitecture targets power-efficient machine-learning inference for consumer, industrial, and IoT devices.
Design teams pair the NPU with Arm Cortex-M and Cortex-A processors and the Ethos toolchain so models compile from common frameworks into deployable firmware.
| Attribute | Arm Ethos-U85 | Vivante NPU IP | |
|---|---|---|---|
| Provider | |  | |
| Founded | 1990 | 2001 | |
| Sells To | Consumers, Enterprise | — | |
| Pricing Model | Licensing, Recurring, Transactional | — | |
| Ownership | Public | — |
Provider
Founded 1990
Sells To Consumers, Enterprise
Pricing Model Licensing, Recurring, Transactional
Ownership Public
Provider
Founded 2001
Sells To —
Pricing Model —
Ownership —
Mobilint's Edge AI Accelerators deliver power-efficient NPU solutions for edge and on-device inference.
Sparse Processing Unit is femtoAI sparsity-aware AI accelerator silicon for extreme-edge inference.
HailoHailo's Edge AI Processors enable high-performance deep learning inference on edge devices.
Kneron's Edge AI Platform delivers on-device NPUs and full-stack software for edge inference.
Licensable deep-learning accelerator IP for power-efficient edge inference.
Licensable edge-AI NPU IP for vision, audio, and sensor-fusion SoCs.
Low power machine learning system on chip for physical AI devices.
Deep VisionAra-1 is the first-generation Ara discrete neural processing unit for edge AI vision inference.
Edge software that runs AI models locally for real-time decisions, including offline.
Deep VisionAra240 is the second-generation Ara discrete neural processing unit for generative AI inference at the edge.
Axera TechnologyAxera's Edge AI Inference Chips provide high-performance, energy-efficient SoCs for edge computing workloads.
Licensable AI processor IP for on-device inference across edge silicon designs.
Mobilint's Edge AI Accelerators deliver power-efficient NPU solutions for edge and on-device inference.
Sparse Processing Unit is femtoAI sparsity-aware AI accelerator silicon for extreme-edge inference.
Hailo's Edge AI Processors enable high-performance deep learning inference on edge devices.
HailoKneron's Edge AI Platform delivers on-device NPUs and full-stack software for edge inference.
Low power machine learning system on chip for physical AI devices.
Hailo-8 is an edge AI accelerator from Hailo.
HailoAra-1 is the first-generation Ara discrete neural processing unit for edge AI vision inference.
Deep VisionHailo-10H is an edge AI accelerator for generative AI inference on devices rather than in the cloud.
HailoEdge software that runs AI models locally for real-time decisions, including offline.
Ara240 is the second-generation Ara discrete neural processing unit for generative AI inference at the edge.
Deep VisionAxera's Edge AI Inference Chips provide high-performance, energy-efficient SoCs for edge computing workloads.
Axera Technology
