AI Infrastructure news and analysis on Web Pulse: 35233 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
A developer has compiled a list of eight open-source tools for building AI agents with memory, including Statewave, Mem0, Graphiti, Cognee, LangGraph, and Letta. The tools are grouped by their function, such as storage, …
Microsoft AI released MAI-Transcribe-2 on Thursday at 10 cents per hour of audio, a 72% price cut from the $0.36 per hour charged for MAI-Transcribe-1 five months ago, undercutting rivals OpenAI, Google, and ElevenLabs. …
NVIDIA is investing $3.5 billion in convertible bonds issued by MediaTek, which will adopt NVIDIA's NVLink Fusion platform to help customers build custom AI accelerators that connect to NVIDIA's rack-scale AI systems. Th…
The loomcycle memory architecture, as of v1.75.0, spans five storage planes—key/value, vector, SQL Memory, chunked-graph Documents, and an entity graph—with four scopes (agent, user, tenant, global) and server-side owner…
Alibaba's Qwen team released Qwen3.8-Flash-Next on August 26, 2026, an open-weight experimental model with 125B total parameters but only 6B active per token, previewing the architecture intended for Qwen4. The model int…
Perplexity has launched a hybrid AI feature for its Pro, Max, and Enterprise subscribers that splits tasks between cloud models and a small on-device model running on Apple silicon, enabling sensitive files such as tax r…
Fireworks AI announced the general availability of its Training API and Fireworks Lab, enabling companies to train specialized AI models beyond frontier labs' offerings without building their own training stack. The Trai…
DumbDetector, a new community-driven platform, launched to aggregate anonymous reports of quality drops, errors, or instability in AI tools, monitoring products from Anthropic, OpenAI, GitHub, Google, Cognition, Perplexi…
At IFA 2026, Acer showcased its SFF RTX Spark, a small-form-factor desktop PC based on NVIDIA's RTX Spark platform, featuring up to a 6,144-core GeForce RTX "Blackwell" GPU and a 20-core NVIDIA "Grace" CPU, delivering 1 …
Anthropic's Fable 5.1 rollout keeps headline prices flat but cuts cache reads by 75%, reducing agentic AI compute costs and potentially driving more compute demand, according to Forbes contributor Jon Markman.
A developer mapped Microsoft's AI stack across six layers, from models and infrastructure to data, agentic platform, Copilot distribution, and a vertical governance layer, and compared it against Microsoft's new AI-500 e…
Pangolin, an open-source zero-trust network access suite, launched a new AI gateway that replaces API keys with WireGuard tunnels authenticated via existing identity providers such as Okta, Azure, and Google. The gateway…
Forbes Councils member Viktoria Semaan argues that enterprise AI adoption requires a control plane to manage usage, model selection, and spending, as companies shift from per-seat licensing to consumption-based AI costs.…
Stack Overflow released Stack Internal 2026.6, adding centralized admin security with header-enforced API keys, session inactivity controls, and daily rate limits up to 10,000 requests per app, plus automated article exp…
Figure AI has signed a $3.5 billion compute deal with UK-based Nscale to deploy up to 100,000 NVIDIA Vera Rubin GPUs for training its Helix AI models, with the deal expected to grow beyond $6 billion. Deployments begin i…
A new analysis from Uptime Institute finds that one in ten data center outages still cause serious or severe disruption, underscoring the critical role of 'gray space' infrastructure—power, cooling, building management, …
Coder has launched Coder Agent Relay, a self-hosted execution environment for cloud coding agents, with SpaceXAI as its launch partner, enabling Cursor Cloud Agents to run inside Coder workspaces on customer-controlled i…
Mastercard's Start Path program has admitted its inaugural cohort of 22 to 23 startups focused on agentic commerce, including SolvaPay and Crossmint, to build infrastructure for AI agents to initiate payments and hold id…
Anker introduced MindBase, a local AI hub for the home, at IFA 2026 in Berlin, featuring an on-device LLM called the TrueSmart Agent powered by a 26 TOPS chip, 64GB of internal flash, and support for up to 48TB of expand…
Cisco disclosed CVE-2026-20212, a critical vulnerability with a CVSS score of 9.8 affecting Nexus 9000 switches with Silicon One ASICs, allowing unauthenticated root remote code execution via TCP ports 43210 and 43211. T…