← All Issues
Monday, June 22, 2026

AI Daily

Your Automated Intelligence Briefing

Business

A National Security Risk at Home, a Flagship Partner Abroad

Ten days after the US gave Anthropic 90 minutes to pull its top models, the freeze is thawing: Trump now says the company is no longer a threat. The same week, Samsung, LG and NAVER all deployed Claude across their workforces. Both are true at once, and the gap between them shows how AI is really being governed.

Read full story →
Models

The Selling Point Is Now: Nobody Can Switch It Off

China's GLM-5.2 trails the best closed model by a single point on coding, and Vercel wired it in within three days. But the sharper pitch after the Fable 5 ban is structural: a model you have downloaded cannot be recalled by any government letter. An export control became the open-weight world's best marketing.

Read full story →
Infrastructure

The Quiet Bet That AI Compute Won't All Live in One Place

Tesla filed a trademark to turn idle Supercharger power into modular data centers; a startup is selling 'owned intelligence' that runs on factory floors. Both wager against the megacampus, and both note the hyperscalers won't lead the charge because distributed compute quietly undercuts their core business.

Read full story →
Hardware

The AI Boom Reaches the Gamer's Wallet

AMD may raise Radeon prices 10 to 15 percent in July. The cause isn't AI chips eating graphics cards directly: it's memory makers shifting wafers to the pricier HBM that AI accelerators use, starving the cheaper GDDR6 in consumer cards. When capital floods one corner of a supply chain, the bill travels.

Read full story →