AI Infrastructure news and analysis on Web Pulse: 31281 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
Krea AI released Krea 2, a series of foundation models for image generation focused on creative exploration, featuring a diffusion transformer architecture and systems for prompt expansion and style referencing. The mode…
Mozilla AI has expanded its encoderfile project to support image classification, allowing vision models to run as standalone executables without Python or serving infrastructure. The update includes preprocessing via Lua…
Cloud storage provider Backblaze secured a 5-year, multi-exabyte storage deal with AI cloud provider CoreWeave worth $335 million. Backblaze will provide HDD-based storage tiers for CoreWeave AI Object Storage, serving c…
A developer released the Cascade Graph, an interactive knowledge map of 393 nodes tracing how physical and economic constraints—from energy to materials—cascade through the real economy, identifying chokepoints, feedback…
Neo4j launched the Agent Memory Service (NAMS), a managed cloud service providing persistent, structured memory for LLM agents using a graph database with native vector indexes. The service offers short-term, long-term, …
Markdown files are a legitimate pattern for agent memory in single-agent, single-user scenarios, but the approach breaks at scale, under concurrent agents, and as facts change and errors compound. Companies like Manus, C…
Three open-weight coding models launched in the week after the US Commerce Department pulled Fable 5 and Mythos 5 offline, including Zhipu's GLM 5.2, Moonshot's Kimi K2.7-Code, and Cohere's North Mini Code. The incident …
OpenAI reported $13.07 billion in revenue against $34 billion in costs and expenses in 2025, resulting in a net loss of $38.53 billion, according to financials seen by Ed Zitron. The losses, nearly eight times higher tha…
Developer NotASithLord released peerd, an open-source AI agent harness that runs entirely inside the browser as a Chrome/Firefox extension, enabling agents to read and drive web pages, spin up sandboxed compute, and comm…
Union Square Ventures (USV) argues that a 'rebel alliance' of open-source tools and increased competition at the model layer is fueling an agentic consumer explosion, creating opportunities for new apps and platforms bey…
The S&P 500 is up 9.8% year-to-date, but many other asset classes are outperforming it, including small caps (+21.7%), value stocks (+15.1%), and emerging markets (+30.8%). This broad market strength marks a shift from p…
NVIDIA announced that DFlash speculative decoding boosts inference performance on Blackwell GPUs by up to 15x for gpt-oss-120b and nearly doubles interactivity for Llama 3.1 8B compared to EAGLE-3. The open-source block-…
China's CXMT, the country's leading DRAM manufacturer, is set to go public on the STAR Market in what could be the largest semiconductor IPO in China. Founded by Zhu Yiming, the company has leveraged technology from bank…
Hugging Face's inference provider information for Meta's Llama 3.1 405B model may be outdated, as the listed provider Featherless AI returns errors and does not show the model as available. A similar issue may affect Bai…
Several Magnificent Seven stocks have declined recently due to uncertainty over AI growth sustainability and increased competition. Fund manager Inge Heydorn of Chelverton Asset Management sees a buying opportunity in Mi…
Microsoft announced at Build 2026 a suite of upgrades to Azure Kubernetes Service (AKS) to optimize it for AI workloads, including AKS on Bare Metal in public preview for direct hardware access, general availability of A…
Memcached's refusal to add features like persistence or replication makes it the most reliable choice for stateless, high-performance caching, according to a developer analysis. The tool's strict architectural constraint…
Mistral AI released OCR 4, a document extraction model with bounding boxes, block classification, and confidence scores, supporting 170 languages and self-hosted deployment. Independent annotators preferred it over leadi…
Google and Microsoft co-developed WebMCP, an open web standard that lets websites expose structured, callable tools directly to browser-based agents, replacing unreliable vision-based or DOM-scraping methods. Chrome 149 …
SEO teams face complex build-versus-buy decisions for AI tools, as AI lowers barriers to building custom solutions but introduces hidden costs in token usage, API calls, and maintenance. Companies like Uber have blown th…