NVIDIA bets $500B on homegrown dominance.

And PixelDance might be the biggest quiet disruptor of 2025.

Hello AI Innovators.

This week felt different.

Not just faster—but deeper. More ambitious.

From Amazon and NVIDIA reshaping AI’s foundation, to China’s quiet surge—and a new tool that might replace your entire video team.

Let’s dive in! 🚀

In today’s newsletter ⬇️

  • 🧰 4 AI tools to boost your workflow—no fluff, just fire.

  • ⚙️ Amazon drops Nova Premier—its boldest AI model yet.

  • 💸 NVIDIA’s $500B chip play could shift the global AI game.

  • 🧠 Alibaba enters the ring with Qwen3—China’s next-gen model.

  • 🎬 Can PixelDance cut your video budget by 70%?

  • 📚 Curated expert picks, tools & resources to stay sharp this week.

Ready to save precious time and let AI do the heavy lifting?

Save time and simplify your unique workflow with HubSpot’s highly anticipated AI Playbook—your guide to smarter processes and effortless productivity.

🚀 Top Tools of the Week!

The News: East Meets West

News from The WEST 👈:

  • Amazon Unveils Nova Premier: Its Most Advanced AI Yet
    Amazon has launched Nova Premier, its largest and most capable AI model to date—marking a serious move to challenge Google, OpenAI, and Anthropic in foundational models.
    💡 Why it matters: Nova Premier is built to power Amazon’s next-gen Alexa and enterprise offerings, showing Amazon’s serious intent to catch up in the foundation model race.
    👉 Read more

  • Microsoft’s Phi-4 Punches Above Its Weight
    Microsoft’s new Phi-4 model delivers performance competitive with much larger systems, all while using far fewer parameters. It’s designed to be lightweight and cost-efficient without sacrificing accuracy.
    💡 Why it matters: Phi-4 could reshape the economics of AI by offering strong capabilities in a smaller, more accessible package—perfect for scaling to millions of edge devices.
    👉 Read more

  • NVIDIA Commits $500B to US-Based AI Chip Manufacturing
    In a bold move, NVIDIA is investing a whopping $500 billion into manufacturing AI chips in the U.S., signaling a major bet on domestic capacity and geopolitical supply chain security.
    💡 Why it matters: This is one of the largest manufacturing investments in tech history—aimed at strengthening U.S. dominance in the AI hardware stack.
    👉 Read more

Would U.S.-made AI chips change your buying decisions?

Login or Subscribe to participate in polls.

News from The EAST 👉:

  • Alibaba Debuts Qwen3 to Compete in China’s AI Race
    Alibaba has released Qwen3, a next-gen foundation model designed to go head-to-head with DeepSeek’s latest offerings. It’s already being integrated into Alibaba’s cloud and e-commerce platforms.
    💡 Why it matters: The rivalry between DeepSeek and Alibaba is heating up—fueling innovation in China’s domestic AI model race.
    👉 Read more

  • DeepSeek Supercharges Its Math AI Model
    DeepSeek has rolled out a major upgrade to its Prover model, which now solves complex math problems with higher accuracy and reasoning ability.
    💡 Why it matters: Specialized AI models like Prover could disrupt how students, educators, and even researchers solve technical problems.
    👉 Read more

  • Nvidia CEO: "China Is Not Behind in AI"
    Jensen Huang, CEO of Nvidia, pushed back against U.S. narratives, asserting that China is "not behind" in AI capabilities—especially in software and foundational models.
    💡 Why it matters: This statement highlights how closely contested the global AI race really is—and the growing tension between perception and reality.
    👉 Read more

What is PixelDance—and Can It Really Cut Video Production Costs by 70%? 🤯

Video is the future—but it’s expensive, slow, and skill-heavy. PixelDance, a new AI video generation tool by ByteDance (yup, the TikTok folks), flips that on its head.

With a simple text prompt like “woman in a floral dress walking through a sunny garden,” you can now generate a cinematic video—no actors, no camera, no editing tools.

💡 Inside this guide:

  • How PixelDance works (and how it compares to traditional video).

  • Real-world use cases for startups and creators.

  • Tips for writing better prompts.

  • When and how you’ll likely get access.

The AI handles camera movement, scene transitions, and lifelike animation—powered by a DiT (Diffusion Transformer) backbone that smooths out visual noise.

What’s your biggest blocker in creating videos today?

Login or Subscribe to participate in polls.

📩 What type of content do you want more of?

Login or Subscribe to participate in polls.

Whether it’s improving your SEO, launching on platforms like Product Hunt, managing newsletters, or building a strong LinkedIn presence, we’re here to help. Curious to see how we can work together?

Until next time!
Leo & Lex

 

Hello AI Innovators.

Is AI peaking—or just getting started?

Some days it feels like we’re sprinting toward AGI. Other days, we’re watching AI play Quake II... and crash into a wall.

This week, Meta dropped Llama 4. Microsoft showed off an AI-generated game demo (with a few bugs). DeepSeek flexed new reasoning upgrades. Meanwhile, Samsung’s AI chip sales stumbled, and AI copyright lawsuits are heating up fast.

But beyond the headlines? AI apps are making real money, changing how we work, and proving that smart, useful products always win.

Let’s dive in! 🚀

In today’s newsletter ⬇️

  • 🧰 4 AI tools that’ll 10x your workflow this week.

  • 🐑 Meta’s Llama 4 is here—and it’s not playing

  • 🎮 Microsoft drops an AI-made Quake II demo (but it’s… rough)

  • 📉 Samsung’s profits take a 21% AI chip hit

  • 🧠 DeepSeek’s new method makes AI way smarter

  • 📚 Quick links + must-read resources to level up 👀

🚀 Top Tools of the Week!

The News: East Meets West

News from The WEST 👈:

  • 🚨 Meta Drops Llama 4 — And It’s a Beast
    Meta just launched Llama 4, its newest open-weight AI model lineup—and it’s gunning straight for GPT-4’s throne. The models are already available to researchers, with Llama 4b (smaller, faster) landing first, and the full-scale Llama 4 model coming soon.

    💡 Why it matters: Meta’s Llama family is no longer playing catch-up—this release could seriously shake up the open-source AI scene.

    🔗 Read More

  • 🚨 Microsoft’s AI-Generated Quake II Demo Drops—With a Catch
    Microsoft just dropped a Quake II level entirely generated by AI, showing off its latest Copilot Runtime tools. The twist? They admit it’s clunky and needs major refinement. It's cool, but not quite ready for primetime.

    💡 Why it matters: AI in game dev is real—but still rough around the edges. This demo shows what’s possible (and what’s not... yet).

    🔗 Watch the demo here

  • 🚨 OpenAI & Microsoft Hit with Merged Copyright Lawsuits
    A U.S. court just combined several lawsuits from authors and news publishers accusing OpenAI and Microsoft of copyright infringement. The consolidated case will play out in New York and could set major legal precedent.

    💡 Why it matters: This could reshape how AI companies use training data—and who gets paid when AI learns from your content.

    🔗 Catch the full story

News from The EAST 👉:

  • 🚨 Samsung Forecasts 21% Profit Drop—AI Chips to Blame
    Samsung expects Q1 profits to slump 21%, citing weak AI chip demand and ongoing losses in its foundry business. Not exactly the AI gold rush they hoped for.

    💡 Why it matters: Even the giants are feeling the pressure—AI chip wars aren’t always a win, especially when demand cools.

    🔗 More on this here

  • 🚨 Japan’s Rapidus in Talks to Mass-Produce Chips with U.S. Tech Giants
    Japan's homegrown chipmaker Rapidus is deep in talks with American tech companies to start mass production of next-gen semiconductors. Their goal? Be the global alternative to TSMC by 2027.

    💡 Why it matters: The U.S.-Japan chip alliance could be key to countering China's dominance—and reshaping the global AI supply chain.

    🔗 Explore the partnership

  • 🚨 DeepSeek Reveals New AI Reasoning Method
    China’s DeepSeek just dropped a new technique aimed at improving LLM reasoning performance. Early tests show it boosts accuracy significantly without bloating compute costs.

    💡 Why it matters: DeepSeek’s momentum isn’t slowing. This breakthrough could make it a serious rival to GPT and Claude in reasoning-heavy tasks.

    🔗 See how it works

The Best AI Apps of 2025—and What’s Fueling Their Success 🚀👀

AI-powered subscription apps aren’t just trending—they’re outperforming nearly every other category in the store.

From ChatGPT and Jasper to Perplexity and Remini, the best AI apps of 2025 are:

  • Launching faster

  • Monetizing smarter

  • Earning 400x more than the average app.

And here’s the kicker—most don’t even offer free trials anymore. They’re that confident.

📈 Key Intel:

  • AI apps now make $0.63 per install (2x the industry average).

  • General AI assistants pulled in $726M+ revenue in 2024 alone.

  • Only 14% offer trials—yet trial start rates are 96%.

  • The best apps use smarter pricing, better onboarding & vertical focus.

💡 Why It Matters:
If you're building or investing in apps and not thinking about AI—you're already behind. The winners in 2025 are the ones blending fast dev cycles with smart monetization, personalization, and real value.

Which AI app feature matters most to you?

Login or Subscribe to participate in polls.

Whether it’s improving your SEO, launching on platforms like Product Hunt, managing newsletters, or building a strong LinkedIn presence, we’re here to help. Curious to see how we can work together?

Until next time!
Leo & Lex

 

Reply

or to participate.