AI Chips news and analysis on Web Pulse: 6902 curated articles tracking the latest AI Chips developments, tools, and research, updated continuously from vetted sources.
China Telecom AI released Xing4.0-29B-A4B on September 22, 2026, which the model card calls "the first model of this scale trained entirely on the Ascend NPU platform with the MindSpore framework," running on Ascend 910C…
China Telecom AI released Xing4.0-29B-A4B, an open-weight agentic mixture-of-experts model under Apache 2.0 that was trained entirely on Huawei Ascend 910C NPUs with the MindSpore framework and no Nvidia hardware. The 29…
The New Jersey Department of Environmental Protection fined DataOne $1.07 million on September 22nd after inspectors found 62 unpermitted natural-gas generators at its Vineland AI data-center construction site, the state…
A developer published a command-line guide for monitoring GPU utilization and temperature on remote servers, using nvidia-smi queries, cron-scheduled CSV logging, webhook-based temperature alerts, and the nvitop terminal…
Huawei Cloud announced the global launch of its latest AI Cluster Service (AICS) at HUAWEI CONNECT 2026 in Shanghai on September 18, 2026, with commercial availability in China on September 30 and outside China on Novemb…
TileRT powered GLM-5.3 to 469 tok/s of single-user generation throughput on 8× AMD Instinct MI355X GPUs, taking the top spot on the InferenceX AgentX leaderboard and beating the second-place NVIDIA GB300 NVL72 system by …
OpenEuler released version 24.03 LTS SP3 in December 2025, which it calls the world's first operating system built for SuperPoD, a tightly coupled scale-up architecture that lets software treat many compute nodes as one …
AMD reported that UltraQuant, its 4-bit KV-cache method for the Qwen3.8-2.4T MXFP4 model on AMD Instinct MI355X GPUs, serves 29% more tokens per second at 24% lower per-token latency than the fastest 8-bit KV backend at …
AMD detailed a local quantization workflow using AMD Quark 0.12.post1 that compressed the Qwen3.6-35B-A3B Mixture-of-Experts model from BF16 to W4A16 on a single Strix Halo system, cutting model weights from roughly 70 G…
Intel released OpenVINO 2026.4, adding EAGLE-3 tree drafting that lifts speculative decoding throughput on Qwen3-8B from 1.65x to over 1.85x at batch size 1, Multi-Token Prediction for Gemma 4, Qwen3.5 and Qwen3.6 on CPU…
Google's Project Suncatcher, a collaboration with Planet Labs, will launch a refrigerator-sized prototype satellite carrying four Trillium-generation Tensor Processing Units aboard SpaceX's Transporter-18 mission on Octo…
PrismML, an AI lab founded by Caltech researchers and advised by UC Berkeley's Ion Stoica, unveiled a 1-bit Bonsai LLM that runs locally on AI smart glasses built on Qualcomm's Snapdragon AR1 Gen 1 Platform, showcased We…
Gartner named Google a Leader in the 2026 Gartner Magic Quadrant for Container Management for the fourth consecutive year, positioning Google highest in Ability to Execute of all vendors evaluated, and ranked Google Clou…
Delos Data announced at the AI Infra Summit in Santa Clara, California, that it is developing Apollo, a data interface chiplet offered as a 30+ Tbps I/O chiplet next to an XPU, a 10+ Tbps near-packaged optical interface,…
Qualcomm released the Early Developer Preview of Linux on Snapdragon X2 on September 23, 2026, the first official path to native Linux on an X2 ARM64 laptop, exposing the Hexagon NPU via FastRPC for 80 TOPS of on-device …
A developer benchmarked a Jev-style decision read of Gemma 4 on a single Google Cloud TPU v6e chip with vLLM, finding that the E2B, E4B and 12B checkpoints fit at bf16 while the 26B-A4B fp8 build also loads, but no 31B c…
Google DeepMind, Nvidia, and EMBL-EBI released a March 2026 update to the AlphaFold Protein Structure Database adding roughly 1.7 to 1.8 million high-confidence protein complex predictions covering about 2,812 viral prot…
TSMC is introducing a new AI Design Kit (ADK) to enable AI-driven agentic workflows, announced at its Open Innovation Platform (OIP) event, according to Aveek Sarkar, director of ecosystem and alliance management at TSMC…
Qualcomm is bringing first-class Linux support to its Snapdragon X2 Elite and X2 Plus processors, integrating full SoC enablement into the Linux kernel with initial support for the Hexagon NPU, Adreno GPU, and CPU cluste…
A custom self-attention in-memory computing architecture built on charge-based gain cells reduces attention latency by up to two orders of magnitude and energy consumption by up to four orders of magnitude compared with …