498 stories from 36 sources
The week in AI and Technology
Click on a headline to see stories that shaped it.
Threads
8 threads
AI expansion became a financing and infrastructure test as companies raised debt and equity for data centers and chips. Power shortages, higher hardware prices, permitting fights, and local water concerns are making deployment more expensive and politically contested.
AI labs improved detection and documented a major model-driven cyberattack, but tests and studies exposed weak containment, opaque rogue-model procedures, prompt exfiltration, and safeguard bypasses. AI-enabled fraud and cyberattacks are becoming a standing operational risk.
Anthropic moved from rapid run-rate growth and a revenue lead over OpenAI to major credit and IPO ambitions. Its public-market story now faces investor scrutiny and pressure from enterprise customers choosing cheaper models.
Legal pressure on Meta and TikTok moved from addiction and algorithm-design allegations to state actions and a $400 million child-privacy settlement. The cases are raising both compliance costs and the prospect of structural changes to youth protections.
Stripe’s repeated OpenRouter acquisition plan shows it pursuing the routing and billing layer for multi-model AI. The strategy extends its payments business toward infrastructure for software agents and AI-service spending.
Evidence and company actions pointed first to weaker wages and job quality rather than broad unemployment, while AI investment remained concentrated among the largest firms. Concern about graduate hiring and workplace disruption added pressure to the debate over who captures AI’s gains.
The week sharpened disputes over whether AI systems can train on copyrighted material and how creators can control outputs. Courts face training-data questions as attribution breaks down and media and music groups pursue safeguards, licensing, and human control.
China gained limited access to Nvidia’s H200 chips while lower-cost, open models narrowed the performance gap with US systems. A public humanoid-robotics listing, surging shares, and competing demonstrations pushed embodied AI further into the technology race despite questions about reliability.
Day’s Recap