AI Chips news and analysis on Web Pulse: 6898 curated articles tracking the latest AI Chips developments, tools, and research, updated continuously from vetted sources.
The US government is expected to announce an AI compute credit donation on Thursday, October 8, 2026, building on the Department of Energy's Genesis Mission, which has drawn more than $800 million in private partnerships…
Average GPU utilization across Kubernetes clusters measured in Cast AI's 2026 State of Kubernetes Optimization Report was just 5%, and only 2% on Azure Kubernetes Service (AKS), meaning 98 cents of every dollar spent on …
Microsoft and NVIDIA launched the Surface Laptop Ultra on October 7, powered by NVIDIA's RTX Spark SoC with up to 128GB of unified memory and 1 petaflop of AI compute, capable of running 120-billion-parameter models loca…
Dell Technologies Inc. announced extensions to its AI Data Platform, adding a Unified Semantic Layer, an Enterprise Knowledge Graph and Knowledge Agents to its Data Orchestration Engine, plus a Data Processing Engine cap…
NVIDIA's DOCA GPUNetIO has evolved from a standalone GPU-centric packet-processing framework into a unified GDA-KI foundation now integrated across NVSHMEM, NCCL, Aerial 5G SDK, UCX/NIXL, NVQLink with the Holoscan Sensor…
Kilawatt Cloud completed a second live verification of its x402 pay-per-job GPU service on October 2, 2026, routing each request to a different supplier — Lambda, Hyperstack, and RunPod — with an autonomous agent paying …
Cedana published a tokens-to-GPUs calculator showing that serving 1 trillion tokens in one month (30 days) on Llama 3.3 70B requires approximately 367 H100 GPUs at a base case, with a range of 211 to 853. The calculator …
Axelera AI launched Europa, a purpose-built inference accelerator delivering 629 TOPS at INT8 across eight second-generation AI cores at 35 watts typical during LLM inference, announced at the AI Infra Summit 2026 in San…
Vast co-founder and CTO Alon Horev said a long-running AI agent session of half a million tokens can consume one-tenth to one-twentieth of a GPU's memory, and that Vast's tiered storage approach offloads KV cache from GP…
NVIDIA released AICR v1.0, an open-source AI Cluster Runtime that provides version-locked, validated recipes for GPU-accelerated Kubernetes cluster configuration, establishing a stable compatibility contract across its C…
An Epoch AI analysis modeling a semiconductor supply disruption found China would suffer higher costs and deeper economic losses than the US, given its heavy dependence on imports of advanced chip tools and materials tha…
Dell Technologies expanded its Dell AI Data Platform with a Unified Semantic Layer, NVIDIA cuDF GPU acceleration, and PowerScale clusters supporting up to 500 tenants, the company said. Dell reported that its Data Proces…
A developer benchmarked six AWS LLM inference backends — Amazon Bedrock, a SageMaker real-time endpoint, vLLM on two EC2 GPU families, and a hand-ported Gemma 4 build on Inferentia2 and Trainium — all driven by a single …
Mistral launched a public preview of Mistral Large 4, a 1-trillion-parameter natively multimodal model with 49 billion active parameters, on Mistral Studio, with weights to be released by the end of the month. Mistral sa…
A developer argues that quantizing small language models for on-device deployment functions as an uncontrolled, unsupervised fine-tune rather than a lossless compression step, silently degrading narrow behaviors like too…
Dell Technologies Inc. expanded the Dell AI Data Platform with a semantic layer and a knowledge graph designed to give AI agents trusted context from a company's own data, and added faster data processing on Nvidia Corp.…
Alibaba released Qwen3.8-Max, a 2.4-trillion-parameter mixture-of-experts model with 95 billion active parameters, claiming an 86.6 score on Terminal-Bench 2.1 and 86.1 on OSWorld-Verified, with the API live August 2 and…
Everspin Technologies demonstrated what it calls the industry's first CXL-attached DDR4 MRAM at the SNIA Developer Conference in Santa Clara, California, using four 1-GB Everspin DDR4 MRAM modules to create a contiguous …
A developer compared five GPU cloud providers for LLM inference workloads, finding Vast.ai cheapest at roughly $1.20/hr for an A100 80GB, ahead of Lambda Labs (~$1.50), RunPod (~$1.89), GCP (~$2.90) and AWS (~$3.20). For…
Reflection AI Inc. introduced Beam, an open-source large language model with 501 billion parameters, which the company says matches or beats GLM-5.2 on some tasks while using between one third and one fourth the hardware…