Meta released Muse Glimmer 30B, its first open-weights model in the Muse family, under the Apache 2.0 license, featuring a 1.8B vision encoder and a 128K+ context window for coding, agentic workflows,…
Meta introduced Muse Glimmer, an open-weight 30-billion-parameter model distilled from Muse Spark for on-device agentic workflows, with ExecuTorch adding end-to-end support for running it on NVIDIA GP…
Meta released Muse Glimmer, a 30B open-weight dense model with a 120K+ context window, optimized for local AI agentic workflows on NVIDIA platforms, delivering over 20K tokens/sec on a single GPU. The…
Siemens highlights that hardware-assisted verification (HAV) methodologies, including emulation and FPGA-based prototyping, are essential for developing advanced AI chip designs, citing its Veloce CS …
Indonesia's Ministry of Communication and Digital Affairs welcomed the launch of Zankore by Indosat, a collaboration among Indosat Ooredoo Hutchison, Ooredoo Group, Nokia, and NVIDIA to build Southeas…
South Korean President Lee Jae-myung is pushing to relocate the Gwangju Air Base, a joint US military facility, by mid-2028 to build a semiconductor megacluster, with Samsung Electronics and SK Hynix …
TSMC reported a 45% year-over-year increase in July 2026 monthly sales, following a June revenue of NT$442.68 billion (up 67.9% YoY), and raised its full-year 2026 revenue growth outlook to slightly a…
TileRT's persistent engine on NVIDIA GPUs achieves up to 500 tokens/s/user on the InferenceX GLM5 FP8 744B benchmark on a single B200 decode server, approximately 3× faster than GB300 NVL72 running tr…
The Autonomous Driving Mobility Expo 2026 (AME 2026) is scheduled for August 25-27 at Coex Hall B in Seoul, featuring more than 60 exhibitors and 150 booths across autonomous-driving systems, AI softw…
TIME magazine is reportedly serving ads that only AI bots can see, a practice that creates two versions of web pages—one for humans and one for AI crawlers. The trend, highlighted alongside Microsoft'…
NVIDIA's H100, H200, and B200 GPUs offer different trade-offs for AI workloads, and choosing among them depends on factors like model size, memory requirements, and whether the use case is training or…
SK hynix and Sandisk unveiled the first HBF (High Bandwidth Flash) standard specifications at FMS 2026, presenting AI memory solutions. The announcement, made on August 4, 2026, aims to address the gr…
NVIDIA released NemotronLabs VoiceChat 11B, an open 11B end-to-end speech-to-speech model for real-time full-duplex conversation, achieving 448 ms smooth turn-taking latency on Full-Duplex-Bench 1.0 a…
SK hynix showcased a full-stack AI memory portfolio at Future of Memory and Storage 2026, held August 4-6 at the Santa Clara Convention Center, including 16-layer 48GB HBM4, wafer-bonded 375-layer 4D …
NVIDIA released its Alpamayo self-driving models as open-source code for commercial use, claiming 'exceptional performance' for the new Alpamayo 2 Super model on the LingoQA test. The move targets Jap…
Google has open-sourced TPU Raiden, an inference optimization library for KV-cache transfer in large language model serving, under the Apache-2.0 license on GitHub, positioning it as a direct counterp…
Amazon is offering the NIMO AI-Creator-Gaming-Desktop, featuring an Intel i5-14400F CPU, RTX 5070 Ti GPU, 32GB DDR5 RAM, and 1TB NVMe SSD, at a 41% discount, reducing the price from $3,899.99 to $2,29…
Firebird launched a 6,144-GPU NVIDIA B200 Blackwell AI factory near Hrazdan, Armenia, which NVIDIA calls the largest AI factory in the Commonwealth of Independent States. The first phase is backed by …
Red Hat published a post arguing that 'the CPU is back' for LLM inference, citing an Intel and Georgia Tech paper that found CPU-side tool processing accounts for 50–90% of total latency in agentic wo…
Delta Electronics unveiled its GoCool-150 liquid-to-air coolant distribution unit at ASRock Rack's Computex 2026 booth, designed to discharge up to 150kW of heat from a single liquid-cooled server rac…