AI Infrastructure news and analysis on Web Pulse: 35109 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
The OKF Agent Memory repository on GitHub has garnered 395 stars and 20 forks, introducing Git-native persistent memory for AI coding agents. This technology enables agents to maintain state across sessions directly with…
South Korea and the United States have agreed on a planned investment of around $22.3 billion to build a 6.3-gigawatt gas plant in Encinal, Texas, to meet rising power demand for AI data centres, Korean media outlet Edai…
Vista Equity Partners published an analysis of inference economics in enterprise AI, highlighting that inference costs—not training—now dominate AI spending as models scale, with enterprises facing up to 10x higher infer…
Noul, a Korean AI diagnostics startup, has won a public procurement contract from Benin's health ministry to supply its miLab MAL platform to major hospitals treating severe malaria, with an initial order of 20 units and…
A new paper reports that a neural network model achieves 6,616 tokens per second on a single Intel AMX core, demonstrating extreme efficiency for small models on Intel hardware. The result highlights the potential of Int…
Malaysia is advancing sovereign AI initiatives with Huawei's Ascend chips, planning to deploy roughly 3,000 chips by 2026, despite initial government distancing. Huawei inaugurated its first AI Lab outside China in Kuala…
DeepSeek plans to deploy at least 160,000 Huawei Ascend 950DT accelerators at a new data center in Inner Mongolia, according to sources familiar with the matter. The cluster, earmarked for inference workloads, would be a…
HP unveiled the ProBook 4 Flip G2i, a 13-inch convertible AI PC with Intel Core processors, an NPU, up to 32 GB DDR5 memory, and 1 TB storage, targeting executives who work in motion. The device features a 360-degree hin…
Salesforce made Anthropic's Claude the default reasoning model across Agentforce, Slack, and its engineering org on 26 August, but kept the deal non-exclusive and Claude selectable in Agent Builder, preserving model choi…
Kubernetes 1.37 “Garhwal,” released August 26 with 67 enhancements, now natively supports scaling workloads to zero replicas via the Horizontal Pod Autoscaler (HPA) in beta, eliminating the need for KEDA in basic queue-c…
OpenAI published two essays on September 7, 2026, detailing its $8,000-per-day-per-researcher compute spending on an automated AI researcher and urging voluntary slowdowns, while Nvidia CEO Jensen Huang declared AGI has …
Stacklok has released ToolHive, an open-source platform under Apache 2.0 that runs Model Context Protocol (MCP) servers inside containers, isolating them from host credentials and network access to enhance security for A…
DeepSeek plans to deploy at least 160,000 Huawei Ascend 950DT accelerators at a data center in Inner Mongolia, which would be among the largest known clusters of Huawei AI chips, according to people familiar with the mat…
A new guide breaks down the real costs of AI development for businesses, separating build costs from ongoing operational expenses. It highlights that data engineering is the most underestimated line item and that compute…
A developer's guide details the emergence of Agent-to-Agent (A2A) marketplaces in 2026, where autonomous AI agents discover, negotiate with, and pay each other for specialized services. The article provides a technical a…
Aivres, a renamed subsidiary of blacklisted Chinese server maker Inspur, exported at least $5.6 billion of advanced U.S. technology, including over $3 billion of computers with Nvidia Blackwell chips, to Southeast Asia f…
The Institute of Foundation Models (IFM), the frontier lab launched by MBZUAI in May 2025, released K2 Horizon, a fleet of six Apache 2.0-licensed open-source models ranging from 0.9B to 375B parameters, alongside the pr…
A developer detailed how they built an autonomous AI agent that accepts USDC micropayments on the Base Layer-2 network, using an HTTP 402 workflow and on-chain verification via Cloudflare Workers and viem. The agent oper…
Engrim, a universal local-first SQLite memory engine for AI CLIs, has been released on Hacker News, claiming to preserve project decisions and state across model switches with a 99%+ reduction in reloaded context cost. T…
An engineer has developed PII Guardrail Studio, a local, air-gapped reverse privacy proxy and encrypted token vault designed to run inside a private VPC, enabling enterprises to use external LLM APIs while complying with…