2885 articles tagged with AI Hardware
70,000 Samsung workers will strike for 18 days beginning Thursday. The wage dispute over AI memory chip profits threatens a projected $66 billion in economic damage in South Korea....
Lam Research has launched a new research lab in Salzburg, Austria. This facility will focus on advanced chip packaging technology. The goal is to boost chip density and lower produ...
Victor M (@victormustar)NVIDIA의 SANA-WM 세계 모델이 오픈소스로 공개되었습니다. 카메라 조건부 world model이며 단일 GPU에서 동작하고, 2.6B 파라미터로 720p 영상 60초를 34초에 처리한다고 소개합니다. 세계 모델·시뮬레이션·에이전트 연구에 실용적인 진전입니다.https:/...
T-Head, Alibaba chip design arm, launches the Zhengwu M890 AI chip with up to 3x performance over the previous generation. The new chip covers the full spectrum of AI tasks from hi...
Tech giant Alibaba released on Wednesday a new artificial intelligence chip it said performed three times as well as its predecessor, showcasing growing domestic chipmaker prowess ...
Rising AI infrastructure demand and delayed semiconductor capacity expansion could keep electronics prices high through 2027, industry leaders warn.
RT @mweinbach: Wer sagt, dass TPUs nicht schnell sein können! Das ist in etwa Groq-Geschwindigkeit, aber aus einer TPU 8i, und bisher mit dem Gemini Flash-Modell deutlich intellige...
Gemini 3.5 Flashの総パラメータ数は2500億～3000億か、Hacker NewsユーザーがTPU性能から逆算 https://fed.brid.gy/r/https://gigazine.net/news/20260520-gemini-3-5-flash-parameter/
Как развернуть Mistral 7B на GPU-сервере через vLLMЕсли бюджет и ресурсы ограничены, а развернуть self-hosted LLM нужно, присмотритесь к такой связке: Mistral-7B-Instruct-v0.3 + ви...
Alibaba Group's cloud unit launched a new artificial intelligence chip called the Zhenwu M890 at its 2026 cloud summit, the company said Wednesday. Alibaba Cloud said the M890 deli...
Alibaba launched a faster AI chip, strengthening China’s domestic semiconductor ambitions amid continuing U.S. restrictions on Nvidia exports.
Cloud GPU pricing has the exact emotional texture of booking a flight at 1:17am. You open one...
Alibaba Group has unveiled its 3X powerful Zhenwu M890 AI chip and teased the upcoming launch of Qwen3.7-Max. Importance Rank: 1 read more
Alibaba has unveiled its latest artificial intelligence (AI) chip, the Zhenwu M890, which is three times more powerful than its predecessor, the Zhenwu 810E.
Unlock the Power of AI with GPU as a Service (GPUaaS)!Training AI models and handling big data workloads no longer requires massive upfront investment in hardware. With GPU as a Se...
Alibaba Group Holding Ltd. added a new processor to its expanding AI technology stack, enhancing a push to cover every aspect of artificial intelligence development.
A major strike at Samsung Electronics is set to begin after pay talks collapsed, raising concerns over possible disruptions to global semiconductor supply chains amid strong AI-dri...
Blackstone is deepening its bet on the artificial intelligence infrastructure boom with a new $5 billion partnership with Google, marking another episode of how the race for AI dom...
Google and Amazon are both moving hard on custom chips to cut NVIDIA reliance. At the same time folks are dropping code to ...
Pretraining large language models with MXFP4 on Native FP4 Hardware#ROCm #LLM #MXFP4 #FP4https://hgpu.org/?p=30811
KernelBenchX: A Comprehensive Benchmark for Evaluating LLM-Generated GPU Kernels#LLM #Benchmarking #Packagehttps://hgpu.org/?p=30812
CUDAHercules: Benchmarking Hardware-Aware Expert-level CUDA Optimization for LLMs#CUDA #LLM #Benchmarkinghttps://hgpu.org/?p=30807
TL;DR — 6 machines, 12 GPUs, 1,000+ concurrent agents, P95 18 ms, voice
Pope's AI Encyclical, Musk Loses, $500M Chip War & $190B Merger Top 6 AI stories from the last 24 hours. From the Vatican to ...