AI Infrastructure news and analysis on Web Pulse: 35027 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
Serverless AI services such as Amazon Bedrock, Azure OpenAI Service, and Google Cloud Vertex AI are changing how organizations deploy AI at scale, but they come with trade-offs in cost and control. The article from InfoW…
Winbond VP Chen Shao-kun said at SEMICON Taiwan 2026 that AI demand is not a bubble but a supply chain capacity bottleneck, citing continued capital expenditure by wafer fabs, advanced packaging, testing, and equipment p…
Cohere released a serving engine for its North Mini Code model built around a decode megakernel that runs 1.25x to 1.41x faster than vLLM end-to-end on a single H100 with BF16 precision. The engine, available on GitHub, …
Data center GPUs physically last 5+ years, but economic replacement cycles are 2-4 years due to rapid performance gains, according to Data Center Knowledge. Resale or GPU-as-a-Service models help maximize value as organi…
PRAMA-Dynamagh, a domain-agnostic control layer for autonomous agents operating through Telegraph Protocol, has been developed to convert mandates into auditable execution trajectories, integrating real Telegraph miner a…
New York Gov. Kathy Hochul and Pennsylvania Gov. Josh Shapiro are among 11 elected officials who have shifted their positions on data centers amid public backlash, with Hochul imposing the first state moratorium on hyper…
Arm Ltd. unveiled its CSS for Mobile 2 platform at the Arm Everywhere China event in Shanghai, featuring the new C2-Ultra CPU core and the Mali G2-Ultra NX GPU, which the company claims delivers up to 15% higher single-t…
TSMC and Samsung have committed to adopting ASML's High NA EUV lithography tools for advanced chipmaking, driven by AI-fueled demand for smaller, more complex semiconductors. The move underscores the critical role of ASM…
AMD at IFA 2026 unveiled Strix Halo and Gorgon Halo chips for Windows PCs capable of running 125-billion-parameter models locally, alongside a 96-core Threadripper Halo Station personal AI workstation, and announced Proj…
NylonME, a memory engine developed by an unnamed team, argues that enterprise AI systems require a centralized memory store akin to databases, predicting the technology will become essential infrastructure. The vendor co…
NVIDIA invested $3.5 billion in MediaTek to advance NVLink Fusion, an interconnect technology that extends NVIDIA's proprietary NVLink protocol to custom AI chips from other vendors, enabling them to integrate into NVIDI…
South Korea's power demand is set to surge by 25 to 30 gigawatts due to AI-driven needs, according to Energy Minister Kim Sung-whan, as chipmakers Samsung Electronics and SK Hynix expand production and new AI data center…
SpaceX AI's Grok 4.6 is reshaping performance-per-dollar benchmarks, while coding agent valuations surge past $40 billion and neo-cloud infrastructure providers secure multi-year backlogs, according to an investigation i…
Doc-scraper, a configurable, concurrent, and resumable web crawler written in Go, has been released as an open-source tool on GitHub to scrape technical documentation websites and convert them into clean Markdown for ing…
OpenAI released GPT-6 Astra, a 500B parameter model with 128K context length and 40% faster inference latency than GPT-5, enabling real-time agent workflows with complex memory systems at half the cost, according to a po…
AMD and the vLLM project detailed five speculative decoding methods for vLLM on AMD GPUs, achieving up to 2.59x throughput improvement for certain large language models. The techniques enable production deployments on AM…
Mistral AI has expanded its AI platform with regional inference endpoints in the EU and US, support for third-party open-weight models starting with GLM-5.2 from Z AI, and a proposed European compute coalition. The updat…
Dash0, an observability platform built on OpenTelemetry, has joined the Agentic AI Foundation (AAIF) as a Silver member to help shape open standards for agentic AI, including the Model Context Protocol (MCP). The company…
Nokia plans to cut Bell Labs research headcount from 681 to 408 by the end of next year, a reduction of about 46% since 2024, and shrink its dedicated research centers from 19 to six, according to internal sources and a …
Samsung Electronics has joined a consortium led by Netherlands-based ASML to develop next-generation semiconductor manufacturing technology, focusing on advancing 12-inch photomask technology to replace the 6-inch format…