In partnership with


Today’s edition is written by:
Anna, Michael, & Thomas

🧋 Hello human and artificial friends,

Musicians are sounding the alarm as AI-generated tracks are uploaded under real artists’ names on Spotify and Apple Music.

As one artist put it: “You just say yes, that’s me. It’s the easiest scam in the world.”

Today, we will bring you ideas on how you can make money without scamming people. 👇🏻

STEAL THIS IDEA

Asia is where AI moves fastest. New products, new habits, new ways of using tech show up every week. Most disappear, but some quietly change how people shop, learn, travel, or create. We pick those ideas so you can study them and build your own version.

Today’s idea

💰 STOP selling the shovel. START selling the gold.

SaaS got cold feet.

Let's be honest. The AI Gold Rush is full of fake prospectors.

  • Most of the new apps you see are just SaaS 2.0: You still pay a monthly subscription to access a tool that sometimes works and sometimes fails.

VCs, especially in China, have noticed thus. They are now converging on the next trillion-dollar thesis: Outcome-as-a-Service (OaaS). It’s also called AI RaaS (Result-as-a-Service) or, put simply: Pay for the Result.

Meet Shihang Intelligence. They build robots that clean ship hulls underwater. They could sell a $500,000 robot plus a $5,000/month software license.

But they don't. They just sell measurable fuel savings.

What’s new about it

They charge for the saving, not the service

A clean hull saves a ship owner $10,000 in fuel per day. Shihang charges a fraction of that proven saving. This simple twist got them to profitability in a deep-tech market where everyone else is burning VC cash.

The AI owns the whole P&L

Shihang owns the hardware, the operation, and the data. The AI runs the whole job. Companies that sell the robot will never get the operational data Shihang collects by cleaning thousands of ships in different oceans. That data is the moat.

The only way to guarantee ROI

This model forces the provider to deliver true business value, not just "enhanced workflows."

  • Cost Savings: Companies like Ramp (FinTech) guarantee you'll save X dollars on spending, or you don't pay.

  • Performance: The "Pay Per Mastery" AI Tutor guarantees skill acquisition, charging a fee only after a student passes a verified, high-stakes assessment.

Why this works

This model was almost impossible before AI.

It works today because:

  • AI can replace real labor

  • Sensors and data make outcomes measurable

  • Automation makes delivery repeatable at scale

What you can build today

💡 Dispute Win-Rate-as-a-Service: Build an AI Agent for massive Amazon/Shopify sellers that handles the entire chargeback defense loop. You charge a success fee (e.g., 20%) only on the money you successfully recover for them.

💡 Guaranteed Qualified Lead Agent Focus on painful recruiting niches (e.g., rare biotech engineers). Your AI Agent handles sourcing, verification, and outreach. You charge a flat fee only when the client is in a meeting with a pre-vetted, engaged candidate.

🥡 Takeaway

If you want to build a valuable AI company, you need to transition from charging for time or access to charging a fraction of the tangible, verifiable value you create.

OUR PARTNER

Effortless Tutorial Video Creation with Guidde

Transform your team’s static training materials into dynamic, engaging video guides with Guidde.

Here’s what you’ll love about Guidde:

1️⃣ Easy to Create: Turn PDFs or manuals into stunning video tutorials with a single click.
2️⃣ Easy to Update: Update video content in seconds to keep your training materials relevant.
3️⃣ Easy to Localize: Generate multilingual guides to ensure accessibility for global teams.

Empower your teammates with interactive learning.

And the best part? The browser extension is 100% free.

RECENT AI

🇰🇷 Korea shifts AI from planning to build mode in 2026: The government plans to ramp compute to 52,000 GPUs by 2028 and 260,000 by 2030 via public-private funding. A new “co-scientist” research model will launch in 2026, positioning AI as an active partner in discovery, while major government-industry spending on 6G will begin as 5G is deemed insufficient for an AI-driven economy. The push is backed by a 2026 AI budget of ~6.7 billion USD and the rollout of the AI Basic Act on Jan. 22, set to make Korea the first country with a fully implemented AI regulatory framework.

🇻🇳 Vietnam passes its own AI law: Parliament approved a standalone AI Law, set to take effect on March 1, 2026, aiming to promote innovation. The framework introduces risk-based rules with stricter requirements for high-risk applications, mandates human oversight in critical decisions, and rolls out a National AI Fund, AI vouchers, and regulatory sandboxes. In parallel, the government plans to invest in a national AI computing center, controlled open-data systems, and large-scale talent development to make Vietnams AI globally competitive.

🧠 China nudges state AI data centers toward domestic chips: Beijing is urging state-owned AI data centers to prioritize Chinese-made semiconductors, reinforcing its self-reliance push even as the US eases access to Nvidia’s H200 chips. A government-backed body is preparing a recommended chip list, favoring Huawei and rising local players like Moore Threads, with major cities targeting over 70% domestic chip use in AI data centers by 2027. The move is expected to curb Nvidia’s dominance in China and accelerate the rollout of homegrown alternatives.

TREND TO WATCH

🧸 AI toys become “emotional hardware”

China’s toy makers are fusing plush and plastic with AI to sell comfort, coaching, and play. Companionship is the hook. Data and software are the moat.

What’s new

  • Scale: The Ministry of Industry and Information Technology puts AI toys on a national consumption priority list. Market is expected to grow from ~3.5B USD in 2024 to ~4.12B USD in 2025.

  • Products that bond:

    • Haivivi’s CocoMate chats with kids and lets parents see transcripts.

    • AI robo-pet Loona maps your home and recognizes up to five people.

    • Chongker’s AI cat adds a simulated heartbeat and cloud “memories” to deepen bonding.

  • Go-to-market proof: Robopoet’s Fuzozo (¥399) sold 1,000 units in ten minutes during the 618 presale and ~100,000 this year. Aims for 1 million next year and ~$56,8–$71,0 Mio GMV.

Why it matters

  • New value prop: “Emotional LLMs” turn toys into companions that learn habits and feel personal.

  • China’s edge: Full supply chain, price discipline, and export muscle. Domestic toy retail hit ~$13.9B in 2024 (+25.5% vs. 2020); exports reached US$39.87B (+19.1% vs. 2020).

Watch the friction

  • Child safety and privacy: U.S. PIRG found some toys sharing risky info. Parents want controls, logs, and clear data policies.

Founder angles

  • Build “toy + app” loops that reward repeated play.

  • Offer parent dashboards with controls, transcripts, and off-cloud modes.

  • Start with one emotion job to be done: soothe, coach, or co-play. Then expand.

BIG DATA

This is how much shares of Chinese GPU maker Moore Threads have surged since its late-November IPO.

🔥 A rocket without news: Moore Threads is being framed as a domestic GPU champion. For many investors, the “China’s Nvidia” narrative alone has been enough to ignite speculation, even without new catalysts or meaningful changes in fundamentals.

📉 Warning, then whiplash: After the company cautioned about potential overheating and “irrational speculation,” the stock dropped as much as 19% on Friday and closed about 13% lower. Momentum can reverse fast when confidence cracks.

Watch: This is China tech euphoria in pure form. Self-reliance as a mega-theme, GPUs as a scarce asset, and capital chasing the “Chinese Nvidia” story. But when a company itself urges caution, that’s usually not a growth signal, but it’s a volatility warning.

TOOLS OF THE WEEK

🖥️ Baidu Wenku made a comeback by turning AI into a practical PPT helper that actually saves office workers time instead of adding more work.

🫂 Ant A-Fu is Ant Group’s health-focused AI, built to be a trusted daily companion by combining medical expertise, real services, and long-term user trust.

Special:

🎓 XuetangX is Tsinghua University’s online learning platform, offering thousands of free university courses worldwide and turning top academic content into open, global education.

  • They launched an AI-Tutor platform that offers full K-12 curriculum support, proving the Outcome-as-a-Service (OaaS) model we talked about above by only charging students for verified skill mastery, not for monthly access.


STARTUP LAB

🧠 China’s Nvidia challenger heads for the stock market: Shanghai-based AI chip startup Biren is preparing a Hong Kong IPO that could raise around 300 million USD. Valued at roughly 2 billion USD, the company sits at the heart of China’s push to build its own AI compute stack as US chip restrictions tighten.

💳 Zed brings AI credit cards to young Filipinos: The fintech startup closed a 16.5 million USD Series A to scale its data-first approach to lending. Instead of relying on classic credit scores, Zed uses AI to read income flows and real spending behavior. The bet seems to be working. The app has nearly 200,000 sign-ups and monthly spend is up 500 percent since early 2025.

🛡️ Saladin raises Series A to take Vietnam’s insurance online: The Hanoi-based insurtech secured a Series A led by SBI Ven Capital to push deeper into life and health insurance, categories long dominated by offline sales. Saladin uses AI to streamline underwriting, claims, and personalization, and already works with 15 insurers, nearly one million users, and a network of 15,000 agents.


BITS TO DO


Read what someone who worked in AI Security would never tell a chatbot.
Watch the internet go wild over McDonalds’s Christmas ad, that’s made entirely with AI!
Get smarter and listen to The AI Training crisis.
Watch and laugh as scammers get scammed right back.
Upgrade your knowledge about the world’s hottest data centers.

*Yes, it’s an affiliate link. No, you won’t pay more. But it helps support asiabits. Thanks :)

FUTURE COOKIE

🥟 Dumplings never tasted better.

CAREER BITS



🌏 Amazon: AI Content Expert, Hyderabad
🌏 OpenAI: AI Adoption Manager, Tokyo
🌏 Salesforce: Solution Engineer Trainee, Tokyo
🌏 Taiwan AILabs: Machine Learning Engineer, Taipei
🌏 Spacevision AI: Data Engineer, Seoul
🌏 AMD: Software Product Engineer, Beijing
🌏 Pattern: Automations Engineer, Hong Kong
🌏 Softbrain: SaaS Engineer, Tokyo
🌏 Qualcomm: Voice AI Test Engineer, Shenzhen
🌏 PeroptyxAI Content Analyst, Japan

How would you rate today’s issue?

It helps us improve the content for you!

Login or Subscribe to participate

Keep reading

No posts found