#edge AI

79 stories · 13 in 7 days · 23 in 30 days
Stories, products and papers tagged "edge AI", newest first.
Nov 16
Nov 14
Chinese content华为麒麟 9050 系列马良 GPU 与 AI ISP 细节公开:二代硬件光追、2 亿像素融合拍照

麒麟 9050 的 GPU 升级数字很实在:性能涨 40%、光追涨 78%,还有 AI ISP 带来的 2 亿像素融合拍照。做端侧图形和影像开发的可以关注马良 API 与 HarmonyOS 的协同方案,对比高通骁龙新平台也有参考价值。

IT之家Source
ModelsBuildersSelected01:02
Liquid AI open-sources LiquidCodec for low-latency on-device voice AI

Codec throughput, not model size, is often the hidden bottleneck in streaming voice agents, and parallel feature generation attacks exactly that constraint. Builders of on-device speech assistants and edge inference tools should benchmark it against codecs like EnCodec or SoundStream for latency and quality. The M4 Max figures also hint it fits Apple Silicon laptops and phones well, useful for teams shipping local-first voice features.

Gorden Sun@Gorden_SunSource
Nov 13
Nov 12
Nov 11
Chinese content联发科赋能今日宜休 ISHO iFlow 智能温控系统,全系搭载 Genio 420

联发科把 3 月发布的 Genio 420(6nm、7.2 TOPS)塞进了今日宜休的智能床垫温控系统,心率呼吸数据全在端侧处理,不上云。标准版 9999 元、Pro 版 12999 元起,做睡眠硬件或端侧 AI 产品的团队可以看看这套芯片加模组的方案。

IT之家Source
Chinese contentArm 发布 Cortex-M42 处理器:向量算力最高达前代 3 倍,面向小型设备本地 AI

做嵌入式和端侧开发的朋友可以关注下:Cortex-M42 把 Helium 向量扩展下放到低成本 MCU,向量算力比 Cortex-M33 最高快 3 倍,跑关键词识别、传感器分析这类本地推理不用再上大芯片。芯片厂升级现有 M33 方案也省事。

IT之家Source
Nov 10
Nov 8
Nov 6
Oct 30
Oct 27
Oct 26
Oct 21
Oct 19
Oct 18
Oct 14
IndustrySelected10:25
Counterpoint: Global prescription hearing aid sales to exceed 31 million by 2030

The SNR gains are the practical headline: a jump from 6.4 dB to over 10 dB is what makes restaurant and street conversations viable for users with mild-to-moderate hearing loss, not just louder amplification. Hearing-aid buyers and audiologists comparing Phonak's AI models against traditional DSP devices should weigh the on-device AI share, which Counterpoint expects to reach half of all units by 2030. The 4% CAGR signals a steady, premiumisation-driven market rather than explosive growth.

IT之家Source
Oct 9
Oct 7
ModelsBuildersSelected76°08:57
Google EmbeddingGemma 2 Model Released

EmbeddingGemma 2 enables local deployment for high-compliance sectors like healthcare, legal, finance, and government, addressing privacy concerns. For developers in Southeast Asia, this provides a cost-effective alternative to cloud-based multimodal RAG systems when paired with Gemma 4, reducing latency and data transfer costs for applications like local video search in regional languages.

AIGCLINK@aigclinkSource
Sep 25
Sep 24
Sep 23
ProductsBuildersSelected73°04:00
Qualcomm brings HBC architecture to Snapdragon chips

The shift means the bottleneck for on-device AI moves beyond NPU compute to memory bandwidth, reshaping how flagship phones will handle large local models. Mobile developers and hardware buyers should watch Snapdragon roadmaps, since HBC-based chips could enable longer context windows and bigger parameter models running locally without cloud calls. The technology was first shown at Qualcomm's investor day in June 2026 for data centre products.

IT之家Source
Sep 21
Sep 14
Sep 10