← All Issues
Thursday, August 27, 2026

AI Daily

Your Automated Intelligence Briefing

AI Geopolitics

The Mystery Model Has a Name Now. It Explains Why the West Is Cutting Prices.

Ox Alpha topped the world's AI usage charts for a week with no maker attached. Now Beijing lab Z.AI has claimed it as a new GLM model and promised to publish the weights. The reveal crystallises why Silicon Valley is quietly slashing prices: China's cheap, good-enough models are leaking into everyone's toolchain, even as both governments move to wall the ecosystems off.

Read full story →
AI Hardware

Nvidia's Best Customers Are Now Building the Exit

Nvidia walked into another near-doubling of revenue this week just as OpenAI unveiled JalapeƱo, an in-house chip it says beats Nvidia's own on the fastest-growing part of the market: inference. Google, Meta, Amazon and Anthropic are all doing the same. The GPU king still owns training and CUDA, but the profitable middle is exactly where its biggest buyers can carve it out.

Read full story →
AI Policy

Bill Gates Wants to Slow Dangerous AI. His First Stop Is Beijing.

As the industry talks arms race, Gates is betting the opposite: that China would agree to curb dangerous model releases if the US moved first. He is seeking a November meeting with Xi to pitch bio-attack monitoring and a new nuclear-style global body. His plan hinges on Washington leading, and its instinct right now is to export American AI, not restrain it.

Read full story →