AI Chips news and analysis on Web Pulse: 6906 curated articles tracking the latest AI Chips developments, tools, and research, updated continuously from vetted sources.
Huawei upgraded its Stellar AI Network solution, adding a revamped NetMaster network agent that the vendor claims can perform root-cause analysis for 95% of network faults "within minutes," alongside new high-density swi…
Durabook, the rugged mobile solutions brand owned by Twinhead International Corporation, announced the Z14I-DX3 Rugged Mobile Laptop, a workstation-class edge AI platform with three integrated 14-inch full HD displays. T…
Alibaba unveiled the Zhenwu V900 AI chip at its annual flagship conference in Hangzhou on Tuesday, which CEO Eddie Wu called "most powerful AI chip in China today" and said delivers three times the performance of the pre…
Alibaba's T-Head semiconductor unit unveiled the Zhenwu V900 AI chip, which Yicai, citing Securities Times, reports delivers three times the computing performance of the previous Zhenwu M890. At the Yunqi Conference, Ali…
Alibaba Group said on Tuesday it is developing an artificial intelligence model up to four times larger than its flagship system and unveiled a next-generation AI chip it called China's most powerful, sending Alibaba sha…
China's active AI agents could grow from 4.95 million in 2026 to 197 million by 2030, a 151.2% compound annual growth rate, according to the 2026 China Artificial Intelligence Computing Power Development Assessment Repor…
Alibaba Group Holding unveiled what it called China's most powerful artificial intelligence chip and teased plans to train an AI model with up to 10 trillion parameters at its annual Apsara Conference in Hangzhou on Tues…
LG Chairman Koo Kwang-mo and Microsoft CEO Satya Nadella agreed on a broader artificial intelligence partnership spanning data-center infrastructure, workplace AI and physical AI at their first meeting Monday at Microsof…
LG Electronics and LG Energy Solution have qualified products under Nvidia's newly launched DSX Ready program, with LG Energy Solution's JF2 AC Link battery energy storage system the only Korean battery maker in the init…
U.S. President Donald Trump is hosting Chinese leader Xi Jinping in Washington this week for their third meeting since Trump returned to the White House, with AI competition, trade and Iran among the flashpoints despite …
Firebird, a San Francisco-based startup founded by people of Armenian descent, began operating a data center in Hrazdan, Armenia, in August that is set to reach 300 megawatts and more than 70,000 Nvidia chips by the end …
Rambus director of product management Berardino Carnevale told Lets Data Science that AI teams renting cloud GPUs must verify which components were measured, who evaluated the evidence, and what happens when a check fail…
Alibaba Group CEO Eddie Wu said on Tuesday the company plans to train a new artificial intelligence model with 5 trillion to 10 trillion parameters, part of a push spanning AI models, chips and data centres. Alibaba also…
Alibaba's semiconductor arm T-Head unveiled the Zhenwu M890 AI chip at the Alibaba Cloud Summit in Hangzhou, claiming it triples the performance of the previous-generation Zhenwu 810E. The accelerator, which packs 144 GB…
Alibaba Group Holding Ltd. unveiled the Zhenwu V900 AI accelerator from its T-Head chip division on Tuesday, which CEO Eddie Wu said triples the performance of its predecessor and can be clustered in up to 500,000 units …
LG Electronics' 2.5-megawatt Coolant Distribution Unit qualified as Nvidia DSX Ready on September 22, 2026, one day after Nvidia launched the DSX Ready program on September 21, 2026, confirming the unit meets specificati…
Moonshot AI released the weights for Kimi-K3, a 2.8-trillion-parameter, 1M-context, natively-MXFP4 Mixture-of-Experts model, with AMD Instinct support available on day 0 across three serving frameworks: vLLM, SGLang, and…
A custom inference stack running Qwen 3.8 27B on NVIDIA B300 SXM6 GPUs reached 9,825 tokens per second on 8× B300 with DP4×TP2 plus Suffix4, and a maximum cluster ceiling of 16,931 tokens per second on 8× B300 via the Fi…
AMD researchers introduced Zebra-HyLo, an upcycling recipe that converts pretrained Transformer models into long-context hybrid LLMs through post-training on roughly 10B tokens, without pretraining from scratch. The meth…
Specialized storage is increasingly shaping inference efficiency for AI cloud providers by preserving AI context, reducing recomputation, and improving GPU productivity, response times, and inference throughput, accordin…