🧋 Hello human and artificial friends,

AI actually makes the world a better place.

A Chinese teenager mixed human imagination, AI, and a bit of hardware to build a self-moving trash can that auto-aims and catches anything you throw at it.

This might be the most satisfying thing you’ll see in 2026.

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.

A team in Beijing raised $20 million to solve a problem we’ve all just accepted as part of being human: We forget almost everything.

Meet Looki. They closed a Series A led by Ant Group and Meituan for their L1 wearable. It’s a 32-gram, cat-shaped camera that hangs around your neck like a pendant.

✅ It doesn't wait for you to ask: Most AI is lazy—it waits for a prompt. Looki is "proactive." At CES, they showed off "Scene-Adaptive Intelligence" that watches you in real-time.

✅ A Google Search for your physical life: Because the AI sees what you see, your day becomes a searchable database. You can ask: "What was the name of that shop I passed on Tuesday?" or "Where did I leave my keys?"

✅ The "Comic Book" recap: Nobody wants to watch 8 hours of raw footage. Looki solves this by auto-cutting your day into highlight reels or—more uniquely—automated comic strips and vlogs. It turns your boring Tuesday into shareable content without you touching an editor.

We know it is hard to build a product from scratch, but if you do, building a wearable like Looki could be an inspiration for you. For example, you could build:

💡 An "industry-specific" life graph: Pick a niche where physical details matter: medical students, flight mechanics, or luxury retail staff. Build a wearable "coach" that logs their training and gives them a "Mastery Score" based on their actual movements and actions.

💡 An AI "memory" for the elderly: Build a simplified version of this for people with early-stage memory loss. The AI proactively whispers the names of people they are talking to or reminds them of tasks they started but didn't finish.

AI has changed how consumers shop by speeding up research. But one thing hasn’t changed: shoppers still trust people more than AI.

Levanta’s new Affiliate 3.0 Consumer Report reveals a major shift in how shoppers blend AI tools with human influence. Consumers use AI to explore options, but when it comes time to buy, they still turn to creators, communities, and real experiences to validate their decisions.

The data shows:

The most effective brands are combining AI discovery with authentic human influence to drive measurable conversions.

Affiliate marketing isn’t being replaced by AI, it’s being amplified by it.

Download the full report to see what this means for your brand.

💽 Baidu moves AI chip unit Kunlunxin toward Hong Kong IPO: The company has confidentially filed for a Hong Kong listing that could value the unit at at least around USD 3 billion. Baidu aims to unlock Kunlunxin’s standalone value, raise fresh capital for chips, software, and server systems, and accelerate its role as a domestic alternative to Nvidia. The timing fits a broader wave of Chinese AI chipmakers heading to market, backed by Beijing’s push for tech self-reliance.

🧠 DeepSeek pitches a cheaper way to train AI: The Chinese startup has published a new training framework designed to cut compute and energy costs as China pushes efficiency under chip constraints. The method, dubbed Manifold-Constrained Hyper-Connections, aims to scale advanced models without top-tier Nvidia access. The paper fuels expectations for DeepSeek’s next flagship R2 model. Analysts say R2 could again disrupt global AI rankings, where Chinese budget models are increasingly punching above their weight.

🤖 Fujitsu and Nvidia team up on multi-agent AI: The Japanese tech group has developed a platform with Nvidia that lets multiple AI agents work together to automate complex workflows, starting with tasks like procurement checks. Built on Nvidia’s agent tools and Fujitsu’s Japanese-language LLM Takane, the system adds security layers so agents can handle sensitive corporate data. Next step: physical AI. The partners plan to extend the platform to industrial robots, including potential use with Yaskawa, as Japan pushes AI from screens into factories.

A new kind of creator just showed up in your feed. She looks real, posts gym selfies, lands big-brand deals, and never misses a deadline. She is also software.

Across Asia, studios are industrializing this idea. From Shenzhen to Singapore, teams brief AI talent like real models, localize looks for each market, and ship campaigns overnight.

✅ Agency-made, client-ready talent

Barcelona agency The Clueless created and runs Aitana López, a hyper-real AI influencer who reportedly generates about US$11,700 per month for the firm, with recent campaigns for Amazon and Razer.

✅ Hyper-real looks, global distribution

AI personas like Mia Zelu and her “sister” Ana post picture-perfect lifestyle content and build six-figure followings. The appeal is polish at scale and zero travel. You can localize looks, settings, and scripts in hours.

✅ A market forming around them

Analysts valued the virtual influencer market at ~US$6.06B in 2024, with projections reaching ~US$45.9B by 2030.

🇸🇬 Asian production shops are about to master the playbook: Singapore’s agencies already run hybrid shoots that blend real products and actors with AI scenes to keep realism high and costs low, facing consumer needs and concerns.

That’s how many AI-enabled devices Samsung plans to have in the market by 2026.

🧠 AI everywhere, not just phones: Samsung will double the number of Galaxy AI-enabled devices across smartphones and a growing list of gadgets, including Gemini-powered Bespoke refrigerators. The pitch: smarter food recognition, auto grocery lists, and recipe or shopping suggestions based on what’s actually inside the fridge.

🔎 A distribution win for Google: Much of the AI stack runs on Gemini. Samsung is scaling Google’s model across hundreds of millions of consumer devices, giving Alphabet what OpenAI lacks most: default, on-device distribution rather than browser-based usage.

Watch: This reshapes the consumer AI race. Google gains reach through Samsung’s hardware footprint, Samsung sharpens its edge against Apple, and pressure mounts on Apple to deliver system-wide AI.

Mark Zuckerberg just dropped over $2 billion to acquire the Singapore-based startup behind Manus, effectively pivoting from "talking AI" to "acting AI" to outpace Google and OpenAI.

Practical tips for your workflow:

🌙 Moonshot AI bags cash to scale Kimi: The Beijing startup behind China’s ChatGPT-style assistant Kimi raised $500M in a Series C, valuing it at $4.3B—and says it now sits on ~$1.4B cash to train models and grab users without IPO pressure.

💬 Shanda founder launches “Tanka,” a chat app with long-term AI memory: Chen Tianqiao is building an open-source messaging app that uses AI to remember past conversations and proactively surface context, betting that human choice and accountability will define the next AI platforms.

🧪 Singapore’s Galatek to automate labs and chip packaging: The automation+AI startup closed a ~$30M Series A to build “smart lab” systems for life sciences workflows and precision tools for advanced semiconductor packaging, expanding teams across Europe, North America, and Southeast Asia.

✅ Try on hundreds of virtual hairstyles and see what suits you best.
✅ Guess which one is AI and which one is real.
✅ Find out why you should have invested 10K USD in Chinese AI stocks last year.
✅ Watch and learn 7 AI skills you need for 2026.
✅ Get ready for your next China adventure with the Asiabits China Guide.*

_{*Buy now and receive 30% discount with the code: Asiabits88. Thanks :)}

Country-side in China be like… 🐓