Termsslug: npu

NPU

神经处理单元

Processor unit specialized for accelerating neural network matrix operations, common in on-device inference.

Curated Intelligence

8 items · Page 1/1
IT之家 · 智能时代T2·Infra
58 pts
Nordic Unveils nRF71 Wi-Fi SoCs with On-Chip Edge AI NPU

Nordic Semiconductor has launched the nRF71 series Wi-Fi SoCs, which integrate an Axon NPU to run edge AI inference such as keyword spotting, anomaly detection and image classification directly on-chip without a separate AI processor. The lineup includes the dual-band Wi-Fi 6 nRF7120 and tri-band Wi-Fi 6E nRF7120E, both with Bluetooth LE, Thread and Matter support, and cuts standby power by up to 20x versus the nRF70.

量子位官网T2·Infra
42 pts
Qualcomm Charts On-Device AI Shift Toward MoE-Powered Agents

Qualcomm argues that on-device AI is moving from single-purpose features such as imaging and speech transcription toward agentic experiences capable of intent understanding and task orchestration. Its approach pairs an 'on-device first' cloud-edge continuum with Mixture-of-Experts sparsity to cut per-inference compute, backed by NPU hardware and a full-stack software ecosystem.

InfoQ 中文T2·Infra
68 pts
Qualcomm Releases Snapdragon X2 Linux Preview: Upstream Kernel Integration and Open Hexagon NPU Drivers

Qualcomm released an early developer preview for Snapdragon X2 Linux, aiming to push support into the mainline kernel to resolve long-standing software fragmentation on ARM laptops. Built on Debian 13 with standard UEFI and systemd-boot, it integrates graphics via Freedreno, Turnip, and Rusticl in Mesa. A key breakthrough is exposing the Hexagon NPU to user-space frameworks through upstream FastRPC drivers, eliminating proprietary daemon dependencies and enabling standardized local AI inference.

IT之家 · IT业界T2·Infra
68 pts
Huawei Unveils Kirin 9050: First Logic-Folding Chip Enables On-Device 30B MoE LLM Deployment

Huawei officially launched the Kirin 9050 series, featuring its first application of 'Tao's Law' and logic-folding technology to boost performance and efficiency. As the first native AI Phone platform, the Kirin 9050 Pro integrates a new Da Vinci NPU architecture, enabling the industry's first on-device deployment of a 30-billion-parameter Mixture-of-Experts (MoE) large model.

Hugging Face BlogT1·Tools
45 pts
Phi-2 Chatbot Runs on Intel Meteor Lake Laptops

Microsoft's Phi-2 model successfully runs locally on laptops equipped with Intel Meteor Lake chips, demonstrating the capability of lightweight LLMs for edge deployment. This case validates the feasibility of using new-architecture NPUs to accelerate small language model inference, offering a reference for offline AI applications.