AI Infrastructure news and analysis on Web Pulse: 35123 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
A melted power connector on an MSI GeForce RTX 5090 Gaming Trio OC has been reported during DLSS 5 Neural Rendering testing in NBA 2K27, with power draw exceeding 600 W, according to Videocardz. The incident, documented …
Five enterprise vendors—Broadcom, Citrix, CrowdStrike, ServiceNow, and Genesys—independently shipped nearly identical three-layer AI agent infrastructure stacks between late August and early September 2026, converging on…
A new guide from Vetted Consumer identifies six bottlenecks that slow local large language models, ranked by impact, with the top cause being the model not fitting in fast memory, which can drop speed from 11+ tokens per…
CVE-2026-82526 (CVSS 9.8), CVE-2026-85695 (CVSS 9.4), and CVE-2026-85620 (CVSS 9.2) shipped this week without default authentication, exposing structural failures in AI middleware architecture, according to Ionix and oth…
OpenRouter has added GPT-6 Astra to its platform, allowing users to access the model through its existing unified API without new integrations. The listing enables simplified model switching but requires evaluation of pr…
Dograh, a self-hostable voice agent platform positioned as an open-source alternative to Vapi and Retell, is now available under a BSD-2-Clause license, enabling teams in regulated industries to keep call recordings, tra…
Developer Nirvik KC built Personal Gemini Journal, a full-stack journaling app submitted to the Accelerate AI with Cloud Run APAC Ideathon 2026, combining conversational AI, voice-to-text dictation, automated mood analyt…
Together AI, which operates one of the world's largest GPU fleets, is hiring a Senior Software Engineer for its Infra Agent Systems team in Amsterdam, offering an on-site role focused on building production AI agents tha…
OpenAI is preparing a gated incentive program that rewards ChatGPT workspace credits to users who install and successfully use selected third-party plugins, according to code in the ChatGPT Windows client version 26.901.…
A technical walkthrough details how to build a production-ready retrieval-augmented generation (RAG) pipeline using n8n for orchestration, Qdrant for vector storage, and Gemini for embeddings and answer generation, empha…
Kioxia Corporation is developing CXL memory modules using its XL-FLASH technology to replace costly DRAM in AI systems, aiming to lower data center costs and address supply chain limits. The company showcased the plan at…
Microsoft researchers reported that a spam campaign using ASCII smuggling, which inserts invisible Unicode tag characters into email text to bypass AI-powered filters, drove daily detections from 21,000 to over 1.3 milli…
UC Berkeley researchers released CUA-Lite, an open platform unifying sandboxes, data, evaluation, and reinforcement learning for computer-use agents, with Lite.OSWorld reproducing the OSWorld task suite in Docker contain…
Perplexity Engineering detailed its GPU embedding serving stack for pplx-embed, revealing that it reuses prefill and decode kernels from its LLM engine rather than building a separate embedding engine. The stack comprise…
Imbue, the AI company co-founded by Kanjun Qiu and Josh Albrecht, launched Cloud in a Bottle on September 6, an open-source personal cloud that offers a managed server starting at $5 per month for users who want to run w…
SigNoz has published a guide for monitoring Modal functions and Sandboxes with OpenTelemetry, detailing three non-overlapping telemetry paths: the Modal OpenTelemetry integration for workspace-wide logs and metrics, trac…
LoreConvo, a local-first memory backend that operates as a native MCP server, aims to standardize session persistence across AI development tools such as Claude Code, Cowork, the Codex desktop app, Cursor IDE, and Hermes…
Abacus.AI, led by Bindu Reddy, launched Astra Mode on September 5th, a configuration that builds 3D games by routing tasks to different AI models, with GPT-6 Astra handling Blender-based 3D work. The company's documentat…
South Korea's Ministry of Science and ICT has selected consortiums led by SK Telecom, Kakao, and KT to develop a national AI chatbot and public-service agents under the 'AI for All' programme, with beta testing expected …
OpenRouter's tracking of the google/gemini-3.1-pro-preview endpoint on google-ai-studio/flex recorded a B3IT total-variation shift of 0.28, with no logprob tracking possible because the endpoint does not return logprobs.…