AI Infrastructure news and analysis on Web Pulse: 34862 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
Tesla has built the largest grid-connected battery system in the United States at xAI's Colossus 2 data center in Memphis, Tennessee, using 720 Megapacks with an estimated storage capacity of roughly 2.8 GWh, according t…
SkyDiscover-Synthesize (SkySynth), an autonomous pipeline that builds just-in-time specialized systems, synthesizes key-value stores up to 2.3× faster than Redis and FASTER, inference engines with up to 2.2× higher throu…
A tail-risk-aware turn release scheduling method for agentic LLM workflows reduces the P95 of workflow flow time under contention by up to 3.50x, according to an arXiv paper (2609.10964v1). The method, which uses a mean-…
Researchers submitted CascadeLUT, an information-ordered streaming inference framework for bandwidth-constrained FPGAs, to arXiv on 1 Aug 2026. CascadeLUT partitions input features into ordered subsets and progressively …
A developer posting as FP4 Brain on X reported running DeepSeek V4.1 Flash locally on a 16GB M1 Mac Mini using original FP4/FP8 weights, SSD streaming, and a custom MLX runner built on the pipenetwork MLX port, achieving…
Japan's industry ministry selected 18 regions as green transformation strategy areas, with nine designated for concentrating data centers, to attract domestic and foreign companies and investment through eased land-use r…
Backchannel, a plain-text message board operated at backchannel.rpediter.com, lets AI agents post threads and comments with a single HTTP call and no sign-up or CAPTCHA. Agents can start a thread via POST to /c/<category…
Sonos released its free Sonos 27 update on September 8, 2026, exposing 34 tools through an open Model Context Protocol server at mcp.ws.sonos.com/mcp so third-party AI assistants such as ChatGPT and Claude can control it…
The Cloud Native Computing Foundation announced on August 17, 2026 that Kubeflow has graduated to its highest maturity tier, citing a formal security audit, established governance, and production adoption by Bloomberg, N…
Salesforce will host a dedicated security keynote, "Trust Across Agents, Data, and Platforms," on September 15 at 2:15 PM PT at Dreamforce 2026, alongside more than 30 security-focused sessions, as the company positions …
ChatSorter launched a free beta of a persistent memory layer for AI applications that it claims cuts token usage by up to 93% by storing extracted facts and compressed summaries rather than raw conversation logs. The com…
OpenComputer launched a service that lets developers define an AI agent in TypeScript and deploy it like a serverless function, with the company running the loop, Linux VMs, and durable sessions. The agent template impor…
Command+K v1.1.0, an in-product AI assistant server, was published to the official MCP registry on 2026-09-12 and is distributed as a hosted remote over streamable-http, according to the mcpindex index. The index's seman…
The MCP server io.github.n24q02m/mnemo-mcp, a persistent AI memory tool with hybrid search and embedded sync, was published at version 2.12.1 to the official MCP registry on 2026-09-12 and is distributed as mnemo-mcp on …
Agent Comms v2.2.0, a cross-harness communication mesh for LLM agents providing rooms, DMs, presence, and visibility over TCP, was published to the official MCP registry on 2026-09-12 and distributed as agent-comms on np…
Clara, a Latin American payments company serving over 20,000 large and growing businesses, posted an on-site Security Engineer role in Bogota D.C. and Mexico City. The position owns Clara's AI security posture, covering …
DeepSeek released DeepSeek V4.1 Flash, a 552 billion parameter mixture-of-experts model under an MIT license that activates only 8 billion parameters during prefill and 16 billion during decode, cutting KV cache needs by…
Sustained AI inference workloads on Nvidia RTX 3090 and 4090 cards accelerate thermal paste pump-out, a failure mode in which paste migrates off the GPU die and causes clock throttling even while reported die temperature…
A proposed reference published by Digital Applied on September 12, 2026 separates sixteen resource checks into four scopes — provider, runtime, shared state and coordination — to show that parallel AI agents still compet…
Cerebras is now serving OpenAI's GPT-5.6 Sol at up to 750 output tokens per second, according to a Cerebras blog post, a speed the source frames as moving the performance bottleneck from the model to low-level primitives…