# infrastructure

Anthropic has revealed how its Managed Agents architecture evolved: by separating sessions, harnesses, and sandboxes, it went from “pets” to “cattle.” The result was a reduction of more than 90% in p95 time to first token, while also solving problems with security and scalability.

Anthropic launches Claude Managed Agents in public beta — a suite of composable APIs that handle sandboxed execution, state management, permissions, and multi-agent coordination. Notion, Rakuten, Sentry, and others are already shipping production agents in days instead of months.

SemiAnalysis examines the 2026 GPU rental crunch: H100 one-year lease rates jumped from $1.70 to $2.35 per GPU-hour, on-demand capacity sold out, and Blackwell deployments filled through September as a self-reinforcing shortage took hold.

Box CEO Aaron Levie argues that as agents expand from coding into all knowledge work, existing software simply wasn't built for them. Every platform needs dedicated Agent APIs and CLIs, and agent interoperability will become software's core competitive edge.

Cloudflare's "Markdown for Agents" lets AI request markdown instead of HTML, cutting token usage by 80%. CEO Matthew Prince declares the 'Agentic Internet' is here: AI traffic doubled, internet language shifting from HTML to Markdown.

Anthropic pledges to cover 100% of grid upgrade costs & compensate consumers for electricity price hikes from its data centers. This comes as AI data centers may consume 12% of US electricity by 2028, facing political pressure & building a 'responsible AI' brand.

Michael Bloch asks what becomes more valuable when AI intelligence is nearly free. His 12 endgame positions include energy, atoms, capital, regulation, trust, proprietary data, attention, network effects, operations, security, physical space, and intelligence itself.

Crypto guy Jordan Lyall spent a week researching security before installing OpenClaw — this is the security guide he wished existed, written for people who don't want to become the next victim