Chrome is about to feel different. Google is wiring Gemini straight into the browser, letting it read your tabs, answer layered queries, and soon even run errands for you. Luma unveiled Ray3, a video model that critiques and re-renders its own footage. Apple’s local models are slipping into everyday iOS apps. And Beijing just cut Nvidia off entirely, telling its tech giants to stop buying U.S. AI chips.

📌 In today’s Generative AI Newsletter:

Google is folding Gemini directly into its Chrome browser for U.S. desktop users, marking the first time its flagship AI model becomes a native part of daily browsing. The update adds new tools for handling context across tabs, a Gemini button built into the interface, and an experimental search mode arriving soon.

Key additions:

The browser has long been a portal to the web. With Gemini embedded inside, it edges closer to becoming a portal to an AI that reads, reasons, and acts in real time. The real question is whether users will treat this as a convenience or as a new layer of control stitched into the fabric of everyday browsing.

Luma has introduced Ray3, a generative video model that combines high-end visuals with built-in reasoning. The system generates HDR footage, critiques its own outputs, and re-renders until the results meet the requested standards, giving creators more control over quality and style.

Features of Ray3:

Ray3 represents a shift from simple video generation to systems that judge and refine their own work. By combining reasoning with visual fidelity, Luma is testing whether AI can begin to act less like a one-off tool and more like a creative collaborator.

With iOS 26 now rolling out, developers are starting to plug Apple’s Foundation Models framework into their apps. The system gives them access to Apple’s small, on-device models for free, with built-in features like guided generation and tool calling. These models are lighter than those from OpenAI or Google, so the early use cases focus on convenience and small workflow boosts.

Some of the first integrations:

Apple’s approach shows a different path in the AI race. Instead of massive models in the cloud, it is embedding lightweight intelligence directly into everyday apps. The result is less about breakthroughs and more about sanding down the friction in daily tasks, one local model at a time.

China’s internet regulator has barred domestic tech companies from purchasing NVIDIA AI chips, cutting the U.S. chipmaker off from one of its largest markets. The Cyberspace Administration of China instructed firms including ByteDance and Alibaba to stop testing and ordering NVIDIA’s RTX Pro 6000D server, which had been designed specifically for Chinese buyers.

The context:

The ban marks another tightening of the technology divide between the U.S. and China. For NVIDIA, already projecting billions in lost revenue from restricted sales, it cements China as a market effectively out of reach. For Beijing, it signals a harder pivot toward building a self-sufficient semiconductor industry, even if the road to parity is long.

We’ve extended our bonus-share deadline—giving you until Oct 17, 2025 to join GenAI.works and secure up to 25% bonus shares.

GenAI.works is building the stack powering the next big platform shift:

✅ Infrastructure for agentic AI systems

✅ Trusted by 200+ companies and a 13M+ global community

✅ Bootstrapped to $3M ARR and scaling fast

👉 Don’t just believe in AI—be part of it.

^{Reg CF offering via DealMaker Securities. Investing is risky.}

🚀 Boost your business with us. Advertise where 13M+ AI leaders engage!

🌟 Sign up for the first AI Hub in the world.

📲 Our Socials