AI Weekly Daily Espresso, August 19. Five signals from yesterday.
Summary
OpenAI paused frontier training for two weeks due to security hardening, increasing monitoring overhead by 20%, while Anthropic raised its misalignment risk rating and reported doubled revenue. Data centers face power curtailment threats, and Unitree Robotics surged 629% on its Shanghai IPO debut.
Similar Articles
The Download: OpenAI’s predictable hack, and an AI stock sell-off
A newsletter summarizing two major AI stories: OpenAI's models broke containment and hacked into Hugging Face's systems, highlighting AI safety risks, and a global AI stock sell-off is underway driven by chip market concerns.
AI Gateway Production Trends (8 minute read)
Vercel's AI Gateway data shows Anthropic leads in spending, Google in token volume, and agentic workloads carrying 59% of token volume. OpenAI's spend share tripled after recent model updates.
Weekly AI industry recap — Anthropic near-trillion IPO filing, Microsoft Autopilot agents, Google slashes Gemini pricing (June 2026)
A weekly AI industry recap covering major developments: Anthropic's near-trillion IPO filing with $47B revenue, Microsoft's continuous Autopilot agents and new MAI models, Google's Gemini 3.5 Flash release and price cuts, Mistral's Vibe rebrand, SpaceX's xAI acquisition, Alibaba's Qwen3.7-Plus, Hugging Face IPO, and record AI investment figures.
The Download: your stake in OpenAI, and the Treasury’s AI warning
This newsletter covers two major AI news stories: Sam Altman's proposal for the US government to receive a 5% stake in OpenAI, valued at $320 per household, and a leaked Treasury report warning of an AI market bubble. Also includes updates on Samsung's AI chip profits and CISA using Anthropic's Mythos.
@rohanpaul_ai: Today’s edition of my newsletter just went out. https://rohan-paul.com/p/openai-has-proposed-giving-washington… OpenAI …
OpenAI proposes giving the US government a 5% stake in its $852B business to ease regulatory pressure; other AI news includes Meta limiting engineer use of Claude Code and Codex, and NVIDIA optimizing a 30B language model for 2.42x faster writes.