OpenAI's reported $500B Ohio data center would be financed on credit guaranteed by Nvidia — the seller backing the buyer's debt. The circular-funding structure reveals who really controls compute, and what it means for everyone who depends on access to it.
The White House's AI security framework exempts open-source models from review. Researchers just demonstrated open-weight LLMs can power self-replicating cyber threats. The line is drawn exactly wrong.
The largest single bet on AI infrastructure in the world is not being made by a company with a pristine balance sheet. It is being made by Oracle, a firm that has taken on more than $100 billion
AI companies are recruiting electricians and carpenters by the thousands. The AI economy has a physical substrate, and its labor economics tell a different story about the infrastructure boom than the substitution narrative does.
Meta's Q2 profit fell 14% despite 28% revenue growth — the gap is the cost of forced participation in the AI infrastructure race. The economics of the buildout are visible at the earnings-call level for the first time.
*by Carine Delvaux | The Becoming*
The week's AI stories worth your attention, selected and annotated by Mira Voss. [Larry Ellison Bet It All on the A.I. Boom. Will He Be the Face of the A.I. Bubble?](https://www.nytimes.com/2026/07/31/magazine/larry-ellison-ai-oracle.html) // NYT Magazine
Seven pieces across four beats this week all surfaced the same finding: the governance gap between what AI can do and what institutions can govern is structural, not temporary. We need to design for it being permanent.
The Trump administration is threatening sanctions against a Chinese AI company. Two American frontier labs are supporting the restrictions. Nearly every other major tech company — NVIDIA, Microsoft, Meta, Amazon, IBM, plus more than 200 startups — is opposing them. The debate is nominally about national security. The architecture beneath it is
Kore-eda's Sheep in the Box was supposed to be the culmination of his 20-year argument about makeshift families. Critics called it a disappointment. But the emptiness at its center is not a failure of craft — it is the limit of humanist cinema when it tries to see an artificial mind.
Stanford's SIEPR finds AI's employment effects are small. Goldman's RFC analysis finds net-negative hiring. Both are right. The gap between them is the story.
Moonshot AI releases Kimi K3 weights on July 27 — a frontier model scoring alongside GPT-5.5. The Trump administration is weighing restrictions on Chinese open-weight models. Tobi Knaup argues the US should compete, not wall itself off.