⚡️ Z.ai, Microsoft, Alibaba — This Week’s AI Power Moves

July 29, 2025

Hello,

Welcome to another issue of The Digital Sovereign, where we decode the AI race, uncover tools worth your time, and help you make smarter moves with tech.

This week, China's AI labs are making serious waves.

While much of the West is still speculating about OpenAI's next big release, Chinese innovators are moving fast — releasing powerful open-source models that rival or even surpass current leaders.

Here’s a look at the breakthroughs that are rapidly reshaping what "cutting-edge" and "open-source" truly mean in the global AI race.

🚀 Z.ai's Open-Source Powerhouse is Here

Z.ai (formerly Zhipu) just dropped GLM-4.5 — and it's a beast:

  • 355B parameters combining reasoning, coding, and autonomous task handling

  • Hits 90% in tool-use success — outperforming o3, Gemini 2.5 Pro, and Grok 4 in agentic tasks

  • Comes with their new "slime" training framework (also open-sourced)

Z.AI claims it's now the world’s top open-source model.

Why it matters: This isn’t just another model. It’s another reminder that China’s AI labs are iterating faster than many realize — and that pressure is now on OpenAI and others to keep up.

🧠 Microsoft’s Copilot Gets a Browser Brain

Edge just launched Copilot Mode — turning your browser into an intelligent assistant:

  • Search across tabs

  • Summarize pages

  • Suggest actions — and soon, even take them for you (like booking travel)

It’s free for now.

Why it matters: Agentic browsing is no longer theoretical. Microsoft, Perplexity, and startups like Dia are racing to embed AI directly into our daily flow. If done right, this could be the future of web navigation.

🎥 Alibaba Levels Up Video AI with Wan2.2

Forget just generating images.

Alibaba’s new Wan2.2 video model is pushing cinematic boundaries:

  • Dual-expert system: one crafts scenes, the other polishes fine details

  • Outperforms Seedance, Kling, and even rivals Sora in video realism

  • Lets you fine-tune color, lighting, and camera movement like a pro

Why it matters: Open-source innovation isn’t just about language. China’s building a full-stack AI suite, with video, voice, and vision all moving fast.

⚙️ 4 AI Tools You Should Know

  1. Runway Aleph – The new standard in video editing with AI

  2. Qwen3-Thinking – Alibaba’s reasoning-focused model

  3. Hunyuan3D World Model – Tencent’s open-world environment builder

  4. Aeneas – Google’s AI for restoring historical texts

🔥 Bonus Drop: Microsoft just released a feature to swap any character voice in your video using AI — one-click audio transformations.

🧵 Quick Hits in AI + Tech

  • 🕶️ Alibaba unveils smart AI-powered Quark Glasses

  • 💻 Anthropic caps usage due to Claude Code demand surge

  • 🤝 Tesla & Samsung’s $16.5B chip deal is locked

  • 🎬 Runway signs IMAX deal for August AI short film screenings

  • 📊 Google processed 980 trillion tokens last month alone (!)

⚖️ ICYMI – AI Policy + Lawsuits

  • 🇬🇧 UK’s Online Safety Act could force age verification across the web

  • 🇪🇺 EU planning controversial chat scanning laws by October

  • 😬 Meta accused of seeding pirated content to train its AI

  • 🔗 Google’s old URL shortener links are officially dead

  • ✈️ Aeroflot loses $50M due to airline bans — ripple effects for AI deals?

💡 The Big Takeaway

Chinese labs are not waiting.

From language to video to hardware, they're building an entire open-source AI ecosystem — fast.

If you’re betting on AI for your business, creativity, or workflow… watching what happens outside Silicon Valley might be your unfair advantage.

Talk soon,

The Digital Sovereign