AI Infrastructure news and analysis on Web Pulse: 31530 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
Workflow SDK 5 beta now compresses run, hook, and step inputs and outputs using zstd, reducing storage and improving performance. Compression automatically applies to larger payloads, cutting storage costs by up to 85% f…
Cerenovus, a Y Combinator Summer 2026 company founded by Harvard students, is building an AI-powered 'company brain' that aggregates and reasons over fragmented enterprise data from emails, Slack, documents, and spreadsh…
PeekAI, a new open-source tool for local-first observability and debugging of Python AI agents, launched on Hacker News. The tool stores traces in SQLite, requires no cloud accounts or API keys, and supports multi-agent …
Linzumi, a Y Combinator Spring 2026 company, launched a chat interface that lets engineering teams coordinate AI coding agents running locally on their own machines, addressing security concerns of cloud-based platforms.…
Microsoft will resume automatic installation of the Microsoft 365 Copilot app on eligible Windows devices with Microsoft 365 desktop apps starting June 2026, according to updated deployment documentation. The rollout win…
Touchmark, a Y Combinator Summer 2026 startup, is building infrastructure to price AI outputs based on quality and delivered value instead of token count. The company offers a TypeScript SDK that records sessions, evalua…
A developer discovered that standard rotating proxy setups can waste over 30% of bandwidth due to failed requests, as residential proxies are metered per gigabyte. To fix this, they built an adaptive proxy routing layer …
Large language models are stateless by default, but agents require memory to retain context across steps. A new technical guide identifies seven types of agent memory—working, semantic, episodic, procedural, retrieval, p…
Mireye, a Y Combinator Summer 2026 startup, is building a geospatial data infrastructure layer that provides AI agents with accurate, cited ground truth about physical locations. The company, founded by Ansh Chokshi and …
ByteDance released DeerFlow 2.0, a ground-up rewrite of its open-source agent runtime, on February 28, 2026. The new version transforms the deep-research tool into an isolated, stateful execution harness for autonomous s…
Banco Santander, owner of Santander UK, announced plans to cut costs by over €500 million (£433 million) by accelerating AI adoption across its global operations, targeting more than €1 billion in extra revenues and cost…
A user requests adding Tenstorrent Wormhole and Blackhole AI accelerator cards to Hugging Face's hardware catalog, suggesting a pull request to update the hardware list in the repository's packages/tasks/src/hardware.ts …
Cloudflare launched 'Temporary Cloudflare Accounts for Agents' on June 19, 2026, allowing unauthenticated users to deploy Workers via wrangler deploy --temporary without signing up. The ephemeral deployments last 60 minu…
Compass, a local-first configuration layer for AI coding agents, introduces guardrails, cost routing, and a hard budget cap to prevent runaway spending and unsafe actions. The tool enforces a dollar limit that halts sess…
The Swiss AI Initiative released Apertus Mini on June 15, a set of 16 smaller language models distilled from the open-source Apertus foundation model, designed for local deployment on constrained hardware. The move aims …
Global automakers are shifting competition in South Korea from hardware to software-defined vehicles (SDVs), with Hyundai Motor, Toyota Motor Korea, BMW Korea, Zeekr, KG Mobility, and Renault Korea launching or planning …
Global automakers are shifting competition in Korea from hardware to software-defined vehicles (SDVs), with Hyundai, BMW, Toyota, and Zeekr launching models featuring advanced operating systems and AI functions. Toyota c…
A developer built a dynamic load balancer using the Least Connections algorithm to replace a round-robin proxy that caused 502 errors during a traffic spike. By tracking active connections per backend and routing traffic…
A backend engineer migrated from GPT-4o to DeepSeek after comparing Chinese and US AI models on real production workloads, finding DeepSeek V4 Flash delivers competitive performance at 40-60x lower cost. The engineer's b…
GitHub updated its Copilot usage metrics REST API on June 19 with a new field, ai_credits_used, enabling organization admins to track per-user AI credit consumption daily. This follows GitHub's June 1 switch to usage-bas…