AI Infrastructure news and analysis on Web Pulse: 34976 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
Physical AI is moving AI beyond screens into robots, but the field is split on where inference should run: on-device versus in datacenters. Figure runs its Helix model entirely onboard, while Physical Intelligence's π0.7…
Red Hat AI 3.5, announced today, adds governance, safety, and observability features, including the general availability of EvalHub, a framework for benchmarking and certifying AI application safety. The update also intr…
Apple Inc. is set to debut its 2nm M6 chip and a quad-die variant in updated Mac mini and Mac Studio desktops this September, a move that resets the AI race by delivering significant performance gains for on-device AI wo…
Google committed €13 billion to AI infrastructure in Finland, its largest single capital deployment in Europe, and signed a 22-year power purchase agreement with Fortum for output from the Loviisa nuclear plant, marking …
NVIDIA Dynamo, an open-source inference framework, supports encode-prefill-decode (EPD) disaggregation to accelerate multimodal model serving, achieving up to 5x faster time to first token (TTFT) and 7x faster end-to-end…
Perplexity launched Q2D-Web on September 9th, a benchmark pairing about 190 million web documents with 69,721 machine-written queries derived from roughly 23,000 production searches to test retrieval systems for AI agent…
OpenAI added GPT-6 Astra and GPT-5.6 to ChatGPT Voice on September 9th, replacing the separate Instant, Medium, and High intelligence settings with explicit model and reasoning effort choices, according to OpenAI's updat…
NVIDIA released CUDA Toolkit 13.4, adding Windows on Arm support, early developer preview for the NVIDIA Rubin GPU architecture, and enhanced GPU management capabilities including Multi-Process Service V3 and CUDA Comput…
System76 announced on Sept. 9 the Thelio Mira AI, a lower-cost workstation optimized for AI development and local AI workloads, featuring dual NVIDIA GPUs and a lower-wattage AMD CPU to prioritize GPU compute.
Axis Communications, a provider of network cameras and physical security solutions, is urging organizations to prepare their infrastructure for the rise of agentic AI, which will automate administrative and optimization …
Kubernetes 1.37 “Garhwal,” released August 26, 2026, includes gang scheduling in Beta, which prevents partial-job deadlocks that can leave idle GPUs costing around $467 per event. The feature, detailed in a September 8 K…
Apple's technical specifications page reveals that the iPhone 18 Pro and iPhone 18 Pro Max may use different cellular modems, with the Pro listing Apple's C2 modem and the Pro Max listing no modem at all. The two models …
AWS announced August 2026 updates to Amazon Bedrock, AgentCore, and the open-source Strands Agent Harness SDK, including million-token context windows for GPT-5.6 Sol, Terra, and Luna models, web search integration, and …
OpenAI announced on September 8, 2026, that an internal AI system, using a swarm of roughly 10,000 agents, produced a proof addressing the Navier–Stokes existence and smoothness problem, one of the Millennium Prize Probl…
Google announced a record €13bn ($15bn) investment in Finland's AI infrastructure, its largest single investment in Europe, which will fund three new data centres, expand an existing site, and support energy projects. Th…
Apple CEO John Ternus unveiled the iPhone 18 series, Duo foldable phone, AirPods 5, and Apple Watch Series 12 and Ultra 4 at a hardware event Wednesday, positioning the devices as 'purpose built to serve as your own inte…
HP's ZGX Fury AI station is now orderable, featuring NVIDIA's GB300 Grace Blackwell Ultra Desktop Superchip with 748GB of unified memory and up to 20 petaFLOPS of FP4 compute, and HP has partnered with Red Hat and NVIDIA…
BMC Helix leaders Erhan Giral and Ryan Manning revealed in an Eye on AI interview that their agentic IT operating system uses multi-agent topologies, synthetic chaos gyms, and open-weight mixture-of-experts architectures…
A forum user asked for recommended specs for self-hosted AI workloads with Odysseus, and a respondent said a 3090 Ti should suffice, noting that a Qwen3.8-27b quant fits in its VRAM and is a good small model, with contex…
Google announced the general availability of the AlloyDB Omni Red Hat RPM orchestrator, which brings production-ready security, resiliency, and low-downtime operations to PostgreSQL workloads in enterprise environments. …