The machine is called the Midea PortaSplit. It sold out on Amazon Germany, France, the Netherlands, and the United Kingdom through July (Nikkei). More than 200,000 units shipped to Europe in the first half of 2026, double the year before, per Midea’s own disclosure via EqualOcean. Reporting from German consumer-tech press documented widespread sell-outs and secondary-market scalping through the summer. Someone built a website purely to help buyers track stock.

Midea, in case the name is new, is a Chinese appliance manufacturer that also owns a majority stake in Toshiba’s home-appliance business and owns the German industrial-robotics firm KUKA outright.

The machine

Europe has been miserable at air conditioning for decades. About one in five European homes has it, per the International Energy Agency. The reasons are structural: heritage rules and landlord constraints that make drilling into old walls impractical or forbidden, national noise limits, EU refrigerant restrictions, and installation costs for a conventional split system that often run into four-figure euros before the unit is switched on.

The PortaSplit was built to fit within all those constraints at once. The outdoor half hangs on a bracket that clamps to the window frame, so nothing is bolted to the building. The refrigerant charge is deliberately kept low, well within every EU refrigerant threshold. In silent mode, it runs at 39 decibels (Midea product spec), quiet enough not to trigger typical residential noise complaints. No drilling. No professional install. No landlord letter.

On Chinese social media, people joke that the company’s compliance lawyers earned their annual bonus.

The address

Last week I posted a short video about the PortaSplit. One of the most-liked comments was some version of: “Why are you calling this Chinese? It’s designed in Germany.”

Here is Midea’s version. Midea conceived and designed the PortaSplit in Europe, with a European team, using R&D infrastructure Midea built inside the market it wanted to sell into. Pain points were identified in Stuttgart. Industrial design in Milan. Manufacturing at Midea’s Chinese factories in Foshan. Development took roughly three years. The Milan design centre opened in 2017; the Stuttgart R&D centre was established in 2017 and operationally commenced in 2019 (Midea).

If “Chinese company” still makes you picture a factory in Foshan and an office in Guangzhou, you are describing what Western coverage described twenty years ago. Midea runs its own R&D across multiple countries, with a proper engineering hub in Germany. Samsung looks like this. Apple looks like this. Midea now looks like this too.

TIME magazine named the PortaSplit one of its Best Inventions of 2025. Not a technical certification, but the sort of Western editorial recognition that would have skipped a comparable Chinese product a few years ago.

The meter

The current PortaSplit has an expiry date. The EU F-gas regulation prohibits fluorinated greenhouse gases above the relevant GWP threshold for this product category from 1 January 2027 (EU F-gas regulation). R32, the refrigerant Midea uses in this model, is captured. Existing units can still be repaired and stock sell-through is permitted after the date.

That could kill it. It probably won’t. Midea has already confirmed a new PortaSplit generation for 2027, expected on natural refrigerant R290.

The mirror

Here is the comparison worth sitting with.

Whirlpool, the American appliance giant that also owns KitchenAid and Maytag, handed operational control of Whirlpool China to Galanz, a Chinese appliance manufacturer, in May 2021 for US$193 million and a retained 20 per cent stake (Whirlpool 8-K). The transaction had its own portfolio context, and Whirlpool still licenses its brand in China. But strip the qualifiers and the direction of travel is legible: a Western appliance giant relinquished operational control of its Chinese business to a Chinese competitor.

Four years before that transaction closed, Midea had already opened its own R&D centre in Stuttgart and its own design centre in Milan. It spent the next several years engineering a product for European regulations and European buildings, from inside Europe, and shipped it during a heatwave Europeans could not solve any other way.

That is the mirror. One direction of travel is running much faster than the other.

Most Western firms have already said the obvious version of the lesson: invest more in China. The specific move worth watching is what Midea did in 2017. It is not localisation, in the sense of taking a home-designed product and adapting its packaging and marketing for a foreign market. It is origination: build product from inside the market, then manufacture from wherever is cheapest.

A hostile reader will point out that one product does not prove a model. Fair. But the same European R&D team unveiled a second product at IFA 2025. The H-Pack indoor heat pump, developed in Germany, designed in Italy, running on R290. Different category, same operating model. Announcing a 2027 PortaSplit successor on R290 makes three. The model is doing what a model does.

Every year I speak to Western executives who tell me they cannot compete with Chinese scale. Very few tell me they cannot compete with Chinese design capability inside their own home markets. That is where Midea has quietly walked. Someone at your firm should probably notice.

P.S. If you want to sit across the table from the people running this playbook, our December trip to Shenzhen and Guangzhou is where it stops being abstract. Founders, factories, R&D leads on the ground. Only a handful of seats left this year.

Claim A Seat

The term was black-collar workers. Someone who supervises AI, judges what it produces, and holds the switch that shuts it off. Chun Xia, a co-founder of the venture firm TSVC, used it on stage. The name comes from sport.

I want to be careful about one thing before we go further, because the discussion around AI and jobs has been almost entirely miserable for two years. AI is deleting tasks faster than it is deleting jobs. The person stays. What they spend their day doing changes.

The technology is moving faster than most companies can plan for. That is exactly why the accountability question keeps getting pushed to next quarter, and why it is worth looking at a country that has stopped pushing it.

The more interesting question is different. When AI does the work, who is accountable for the answer? Who sets the machine’s rules? Who protects the person working alongside it? China has started to answer those three.

The interfaces China is drawing

Three developments in Chinese labour policy this year interlock. I should be honest about that word. Beijing has not published a single “AI workforce” framework. It has issued the pieces separately over three months, each settling a different question. That is my reading, not Beijing’s.

The first settles who decides when software allocates work. In April, the CPC Central Committee and the State Council issued guidelines giving platform workers the right to be informed about, participate in and choose algorithm-related rules covering income distribution, pricing and delivery times. The policy target for standardising employment practices around new forms of work is 2027.

The second settles what the new human work is called. In July, the Ministry of Human Resources and Social Security opened public consultation on a draft list of 12 new occupations, including one that Xinhua translates as “embodied intelligence robot technician.” Occupational codes are the plumbing under national occupational standards, vocational training and talent evaluation. The Ministry has said national standards, training programmes and talent-evaluation frameworks will follow from the codes.

The third settles what protection follows a worker when the old employment relationship no longer holds. A pilot injury scheme for gig workers went nationwide on 1 July, covering all 31 provincial-level regions. By the end of June, 29.9 million people had enrolled across 17 provinces and 11 platforms. It attaches protection to the order rather than the employer, so a rider working three apps is covered by all three.

The three together define three interfaces. Who controls the machine’s decisions. What the new human work is called. What follows the worker when the employment container no longer holds. This is a gig-worker rule, not an AI rule. It matters here because AI-mediated work breaks into the same units these schemes already protect: the task, the transaction, the moment of accountability.

That combination is what makes the Chinese case unusually early and unusually operational. The black-collar workforce itself is not yet at scale. What is beginning to take shape is the architecture around it. One to a hundred, as I put it on stage. China rarely does zero-to-one invention. Its usual contribution is building the architecture at scale once a shift is underway, and this is that pattern applied to AI-mediated work.

You can see the cost of leaving those interfaces informal. In 2024, an Air Canada chatbot told a grieving passenger he could claim a bereavement discount after his flight. No such policy existed. The Civil Resolution Tribunal of British Columbia rejected the airline’s argument that the chatbot was a separate entity and held the airline responsible for what the bot had said. The award to the passenger was small. The precedent was not.

A note for readers in Greater China: 黑领 has historically carried several senses in Chinese usage, most often the high-end technical specialist. I am using it in Chun Xia’s newer sense.

The numbers, and the tension underneath them

Research from ShanghaiTech’s Center for Enterprise Innovation and Sustainable Development, reported in Yicai on 21 July after release at the World Artificial Intelligence Conference, analysed around 65,000 tasks across 1,267 occupations in Chinese recruitment data. It found that around 21 million jobs have been generated through AI-linked task reorganisation, representing about 18.6 per cent of total recruitment positions. Read the methodology carefully. The report is state-adjacent, released at a state-organised conference, and its top-line conclusion aligns with the government’s employment-stability line. That alignment is worth naming rather than treating the number as neutral.

What the finding does support is a shift measured in task reorganisation rather than headcount deletion.

Set that against the counterweight. This is the same country where 12.2 million university graduates enter the workforce this year, according to the Ministry of Education. Youth employment has been a source of open policy anxiety for two of them. The Human Resources Minister, Wang Xiaoping, has said the intention is to use AI to create jobs and upgrade traditional ones while keeping employment stable.

China is not evidence that AI creates jobs smoothly. It is evidence of a state building institutional infrastructure while the underlying labour market is still under pressure.

The awkward parts

Take the honest counterexample. Chinese delivery riders told National Business Daily in July that their per-order pay had roughly halved after the platform delivery war. The architecture settles who is accountable and who is protected. Not what the work pays.

Coverage in the underlying scheme also remains partial. Against official estimates of around 84 million platform workers in China, current enrolment covers under half. And the ambition to use AI to create jobs runs alongside anxiety about being replaced by it. In a Tencent Research Institute survey published in September 2025, around 40 per cent of Chinese respondents said they were concerned about their roles being replaced by AI.

China is not the only country moving on this. Singapore’s Platform Workers Act, in force since January 2025, mandates work-injury compensation for platform workers. The EU’s Platform Work Directive addresses algorithmic-management transparency and the employment presumption. Spain’s 2021 Rider Law required algorithm disclosure to works councils. None of them has the combination. Occupational codes, algorithm rights, portable injury protection, all built around the same unit and tested at a workforce approaching 200 million. That combination, at that scale, is what makes the Chinese case worth watching in its own right.

If someone tells you China has solved the AI-and-work question, they are selling you a report. What China is doing is closer to institutional carpentry. It is early, uneven, contested, and worth watching for what it is actually doing rather than as a morality tale.

What to steal on Monday

Four things a Western operator can actually use.

Steal the clock. Name the person whose signature would go on the answer if the AI got it wrong. Not the vendor, not the platform team, not the model. The person. If you cannot name them, that is the interface China is formalising and you have not. Ask the harder version too. How many committees in your company would kill the line “the model said so” the first time it appeared in a decision memo? If none would, you already know what to change.

Steal the seat. The job postings I have been sent this year mostly describe the work an AI can already do. If you are hiring for a role whose actual daily work is judging AI output, put someone doing that job on the interview panel before you post it. Otherwise, you will keep hiring the previous version of the role.

Steal the design principle. In China, the interesting move is to attach the protection to the task rather than the employment relationship, and to build the classifications and algorithm rights around the same unit. The specific insurance scheme matters less than the design choice. If your workforce is a mix of employees, contractors, agents and platform labour, the useful question is what the transaction should carry and who is responsible when the machine gets it wrong. China has committed to a set of mechanisms that address parts of that question. Most Western institutional debate is still stuck on the previous one, whether the person is an employee at all.

Steal the humility. A lot of corporate paperwork exists to catch human unreliability: approvals, sign-offs, compliance checks assuming someone might lie or cut a corner. Agents can log actions in ways humans cannot, which changes what some of that scaffolding is for. Rewrite it. Do not just duplicate it with AI oversight bolted on top.

What actually changes

The unit of design is changing. Employment used to be the natural container for what an economy owed its workers: pay, safety, training, and benefits. AI-mediated work is loosening that container. Some of the work is being done by an agent, some by a supervisor, some by a rider on three apps at once, some by a graduate still looking for their first job. The interfaces China is drawing are early answers to what the new container looks like.

There is a line I use on stage, and I will use it here. The next decade will not belong to the best inventors. It will belong to the fastest adopters. Your business will never outperform the mindset of its leader. Most Western leaders are still bringing a procurement mindset to AI: which model to buy, which vendor to sign. The story above rewards something different. It is about what you design around the machine: who is answerable for what it produces, and what protection follows the person doing the work.

Every company I speak to is planning for the AI it is going to buy. Very few have written down whose job it is to be answerable for what it produces. That is the interface worth naming inside your own organisation. Nobody is going to name it for you.

P.S. Chun Xia’s April 2026 essay on the black-collar worker is worth an evening. If you want to see the country writing the occupation codes in real time, the December tour is where the abstract stops being abstract.

Claim A Seat

Fifteen years ago in Shenzhen you’d queue with a little plastic stool for lunch, because the wait was long enough to sit down for. People still queue for a good hotpot. What moved off the street was the queue for everyday things: errands, groceries, painkillers, phones. It lives inside an app now.

The Western reader’s yardstick for fast is still Amazon Prime, one day. That bar was cleared in China years ago. The question that matters now is who else gets to be that fast.

What actually got built

Meituan runs the biggest of China’s quick-commerce rails, hitting 120 million instant-delivery orders a day at its July 2025 peak across food and retail (SCMP), dispatched from a network of around 30,000 lightning warehouses headed for 100,000 by 2027.

Alibaba’s answer, Taobao Instant Commerce plus Ele.me, cleared 80 million daily orders in July 2025, with more than 13 million from non-food (TechNode).

JD’s own instant-commerce arm reached 25 million daily orders only four months after launch (JD.com IR), dispatching from over 100,000 offline stores at a sub-30-minute average.

One popcorn number to keep in your head: CAITEC estimates China’s instant retail market crossing 2 trillion yuan by 2030 (People’s Daily), up from 650 billion yuan in 2023. Around US$285 billion at the 2030 mark. Not far off the entire UK grocery sector today.

The rider at my friend’s building wasn’t sent by Li-Ning. He was sent by Meituan. Which means the independent sportswear retailer that had the shoes in stock got to fulfil the order at Nike’s speed, for a per-order fee, using the rail.

The West cracked food delivery. It never made the same leap to general retail.

Food delivery works. Deliveroo, Uber Eats and DoorDash all run at national scale. What didn’t cross over was consumer goods. The general-retail attempts (Getir, Gorillas, Zapp) failed to bridge from delivering a burger to delivering a pair of trainers.

Getir bought Gorillas for €1.1 billion in December 2022 and exited the UK, Germany, the Netherlands and the US sixteen months later, in April 2024, with around 1,500 UK jobs going with it (Tech.eu). Zapp scaled back before that. What survived is food, and mostly for chain partners. General retail from an independent shop, delivered in half an hour, hasn’t become an ambient expectation in most Western cities.

The Western small merchant is having a bad decade. Independent closures across UK high streets, US main streets and European city centres come from many pressures, and delivery is only one of them. It’s the one this piece is about.

What the marketplace actually looks like

Amazon’s third-party sellers make up 69% of Amazon’s sales. That sounds decentralised until you look at where the money actually goes: the top 1.6% of Amazon sellers pull half of that (Marketplace Pulse). And those sellers, however small, are tenants on Amazon’s platform. Fulfilment by Amazon is shared logistics too, but only for sellers who put their store on Amazon. It doesn’t help the shop next door running its own website.

Pinduoduo, by contrast, holds no inventory of its own. Every yuan on the platform belongs to a third-party merchant, most of them small factories and small-town retailers previously underserved by Alibaba and JD. Taobao is the same structure. One billion users, all of the GMV coming from independent sellers, small factories, and neighbourhood vendors. Meituan Shangou’s rider network dispatches from a Watsons chain and from a family-run pharmacy through the same app, on broadly the same delivery promise.

Living here, the surprising thing isn’t the twenty-minute delivery. It’s watching a family-run pharmacy on Nanshan Road dispatch a rider through the same app the Watsons chain uses, on the same window. The rail doesn’t ask who you are before it moves the parcel.

The hard parts

None of which means Chinese platforms are benevolent.

Meituan takes up to 26% commission from some new merchants and averages around 13%. In 2022 regulators pushed it to cap fees at 5% for around a million SMBs under operational strain (Bloomberg), and the cap only exists because the strain was public. Rider labour is a serious problem. Meituan runs around 7.5 million riders on the network (Caixin), largely without social protection or health insurance. Strikes over pay cuts are recurring. A 55-year-old delivery rider in Hangzhou died on his bike in September 2024 after an 18-hour shift (RFA, SCMP); Meituan announced social security benefits for some riders from Q2 2025 in the aftermath. Pinduoduo has a counterfeit history. Platform lock-in is real, and once you’re on the rail, switching is a real undertaking.

The argument is narrower than that. The infrastructure Chinese platforms built lets small merchants use it on the same terms as the biggest brands. The Western equivalent doesn’t.

What that means for a Western operator

Western retailers keep diagnosing this as a delivery problem. It’s an infrastructure problem, and Amazon’s speed comes from a network independent merchants can’t plug into. Better delivery, better tools, better ads: none of them fix the missing rail.

The honest options for a Western brand are two, and neither is what the consultancy deck is going to sell. One: pick the sectors where 30-minute delivery does not matter and defend those. Two: build the whole stack yourself, factory to platform to rider network to warehouse. That’s a bet almost no Western company is set up to make.

The small merchant in Manchester or Sacramento is losing to Amazon for many reasons: selection, Prime, price, algorithmic reach. Shared logistics is the one that’s most consistently underestimated, because there is no independent rider rail for them to plug into, and no one has built one at meaningful scale. In the West, delivery is somebody’s moat, not somebody’s platform.

China is the toughest gym a Western retailer can walk into. This is one of the workouts they can’t skip.

If your 2027 retail strategy is about catching Amazon on delivery, it’s aimed at the wrong problem.

The rider who brought my friend’s Li-Ning shoes wasn’t sent by Nike, and wasn’t sent by Li-Ning. He was sent by Meituan.

P.S. If you want to see a rider network and a factory floor on the same trip, that’s what the Shenzhen days are for.

Claim A Seat

That was Luckin’s subsidy era. A large latte for a few yuan, sometimes free, an app of vouchers that refilled faster than I could spend them. Taxis were the same, with DiDi practically paying me to sit in the back seat. Then it stopped. The vouchers thinned, the free ride became a normal fare, and one morning the discount was just smaller.

Watch the price of AI right now, and you are watching the opening act of that same play. A task that costs a few US dollars on a top American model runs for pennies on a Chinese one. Having lived through the coffee and the taxis, I already know how the first act ends.

The playbook has a name duller than it deserves: penetration pricing. Subsidise hard, make the habit stick, build the base while rivals bleed, then turn the meter on once you are the default. Taxis, coffee, shared bikes, community group-buying: China has run this move for a decade, right up to the graveyards of the companies that mistimed it. AI is the newest and biggest version.

Two things about how it ends decide whether the number on your 2027 contract is a bargain or a trap. The crowd shouting that it will all collapse has missed both.

Act one: get everyone in the room

Two things stand out: how cheap it got, and how fast it got there.

DeepSeek kicked it off in 2024, pricing its model at around 1 yuan, roughly fourteen US cents, per million tokens and daring everyone to follow. They did. Alibaba cut its Qwen model from US$1.10 to seven cents per million tokens. ByteDance came in at the floor with Doubao, its top tier priced around ninety-nine percent below the going enterprise rate. Baidu made two of its models free the same afternoon. Within months, the price of machine intelligence in China had fallen off a table.

A market that moves this fast and this deliberately has almost always run the move before. The pace is real: when DeepSeek’s rise rattled everyone in early 2025, Alibaba fired back with Qwen 2.5-Max on the first day of the Lunar New Year, shipping on the one day the whole country stops while rivals were off with their families. That is the engine under the falling price.

The spending is aimed at one thing: getting the tool into people’s daily routine before the price ever moves. Over Chinese New Year, Alibaba reportedly put around US$400 million into getting ordinary people to use its AI assistant, a subsidy aimed at the exact moment a tool stops being a novelty and becomes a reflex. Local governments in Beijing, Shanghai and Shenzhen handed startups vouchers for computing power, the way a mall hands out coupons on opening weekend. The low price is just bait for the habit that follows.

It is working. On OpenRouter, the marketplace that routes developer traffic across models, Chinese AI models climbed from under a tenth of the flow in early 2025 to a record of roughly 58 percent this summer. Which Chinese model sits at number one keeps changing as the rankings churn, but the direction underneath is steady. And OpenRouter is one developer marketplace rather than global market share, so treat it as a strong signal, not the final word.

One honest note before anyone declares a winner: that is share of volume, not revenue. When a model costs pennies, a great deal of usage still adds up to modest income. The traffic has moved; the money has followed far more slowly.

You might object that none of this should transfer to AI. You can swap models with a line of code, so where is the lock-in that made taxis and coffee stick? It builds more quietly. A model gets written into an enterprise’s procurement and compliance sign-off, its quirks learned by a company’s developers, its outputs baked into fine-tuned workflows and tooling. By the time the free trial ends, switching is a migration project, not a toggle.

Act two: the meter is already being installed

In coffee and in taxis, the discount was only ever the setup for the day after it ended. The first signs of that day are already showing in AI.

DeepSeek began charging more during peak Beijing hours from July, roughly double the rate in the busy windows, the behaviour of a company that has decided its compute is worth pricing rather than giving away. MiniMax now ships its top model under a licence that is free until you get big, then needs the company’s sign-off once your product passes US$20 million in revenue. Goldman Sachs expects this drift across the board, from pure free-and-open toward open-but-invoiced, and projects China’s model revenue climbing from about US$4.8 billion this year to US$121 billion by 2030.

It shows up in capacity, too. When demand for Moonshot’s Kimi spiked this summer, the company paused new subscriptions rather than keep serving them free. A product you cannot sign up for is a price signal of its own.

Thin margins forever were never the objective. The prize is to become the layer everyone else builds on. Whoever supplies the cheap intelligence that a million products are wired into does not need to win on price for long. It needs to be the default when the price finally moves.

Free is a strategy with an expiry date, so a 2027 budget that assumes four cents forever is built on a price that is already starting to move.

The people expecting a collapse are right that the rock-bottom price won’t last, but they are wrong about what happens after it.

Why the edge outlasts the discount

The West ran a version of this play too, to be fair about it. Cheap tokens got businesses hooked. Companies rebuilt their workflows around AI, let people go, and wired models into everything on the assumption the price would only keep falling. Call it tokenmaxxing: spend freely now, because the next query is basically free.

Then the bill landed. Klarna replaced around 700 customer-service staff with an OpenAI-built assistant in 2023, and spent 2025 quietly hiring humans back, its chief executive admitting the company had focused too much on efficiency and cost and that the drop in quality wasn’t sustainable. This year Uber’s technology chief reportedly exhausted his entire 2026 AI budget on tokens by the spring; Microsoft engineers are said to have burned theirs on Claude within months, and an Nvidia executive has put it bluntly that compute now costs his team far more than the people do. For a growing list of jobs, a human turns out to be the cheaper option, a punchline the 2023 press releases did not see coming. That is the tell of a subsidised cheap; the reckoning always finds whoever leaned on it.

When the meter turns on in China, the price will rise. It will still, in all likelihood, sit below the West’s for years, and that comes down to plain arithmetic.

Look at where the American cheap comes from. The US frontier labs are, for now, pricing their most-used services below what it costs to serve them, burning investor money to hold market share. By one account, OpenAI spends close to two dollars for every dollar it earns on inference. When it shut down its video tool Sora earlier this year, the thing was reportedly burning around US$15 million a day against a couple of million dollars in lifetime revenue. That is a price held down by a chequebook, and chequebooks close.

Now look at where the Chinese cheap comes from, and this is the part that speaks to how long it lasts. The data centre is roughly half as expensive to build in the first place: about US$5.5 to 6.5 million per megawatt in China against US$10 to 12 million in the United States, before the AI chips even go in. Then it runs cheaper. China puts its compute where the power is, and a hub in Inner Mongolia pays about four US cents a kilowatt-hour, less than half the coastal rate and well below typical US industrial power. Add state-allocated land and cheaper construction and operating labour, and the low price is simply what a lower cost base produces. Water is the one input where neither side has an edge; both are running into the same drought and cooling limits.

There are two kinds of cheap in this story. One is a discount somebody is funding and will eventually stop funding. The other is simply a lower cost of production.

None of this is as clean as “American cheap is fake and Chinese cheap is real.” China’s prices lean on venture money and provincial incentives too, and the US labs have genuine scale and hardware efficiencies of their own. The difference is one of degree. More of the American price is a subsidy that has to end; more of the Chinese price rests on inputs that stay low.

The honest forecast is duller than either camp wants: the free trial ends, prices normalise, and even then the floor stays lower in China, because the country pairs real technical capability with the cheapest inputs for making intelligence anywhere. It also means China can hold out longer: when your costs are structurally lower, you can keep prices low without bleeding, long after a subsidised rival has had to stop. If I were advising a Western team, this is the line I would make them repeat back to me: competing on price alone is a losing game, and the sooner a strategy accepts it, the better.

Domestically, this is how it always ended: whoever could fund the discount longest was the last one standing, and the rivals quietly disappeared. You can see the international version starting. Chinese models have gone from under a tenth of the traffic on OpenRouter to a record 58 percent in eighteen months, with DeepSeek among the biggest providers on it. But a global market has firewalls a domestic one does not. Much of that 58 percent is cheap, high-volume work, and the money still sits at the frontier; Western governments and banks are restricted from touching Chinese models at all; and the US labs are funded by backers who can absorb losses for years. So the likelier ending is a divided market. China takes the volume and the price-sensitive world; the West keeps the frontier and the regulated core.

The catch worth taking seriously

None of this means route everything through a Chinese model tomorrow.

The cheapest option is also the one whose data handling drew warnings. Through 2025, bodies from NASA to the US Navy to Italy’s privacy regulator restricted DeepSeek over where user data was stored. There are real questions about whether quality holds up as prices fall. And the neat story about endless cheap power in the western provinces has credible doubters, who argue that much of that capacity is overbuilt and underused. Name those things plainly, or you are selling, not analysing.

The practical line is not dramatic. Take the price on the work that doesn’t need protecting: the internal drafts, the routine code, the summaries. Keep regulated data, the customer records and the health and financial information, off anything routing through servers you can’t see. You can do both at once. Most of the upside is in the boring, high-volume work anyway.

What this leaves you with

The uncomfortable part has little to do with the falling price. What changed is that the thing you were quietly counting on as an advantage, access to a powerful model, just stopped being one. When the same task costs a few dollars for one company and a few cents for its competitor, and the gap is structural rather than temporary, the model you rent cannot be your moat. Your product has to be. Your data has to be. Your judgement about what to build has to be.

The coffee got more expensive again. People kept buying it, because by then the habit was set and the price was beside the point. That is the whole game. The four cents was never the offer. Becoming the thing you build on, before anyone thought to check the bill, always was.

P.S. Four days on the ground in Shenzhen is where the pace behind all this stops being abstract. If you want to see the engine up close, that is where to do it.

Claim A Seat

This is not how Chinese brands used to go West. For 30 years, the move was to hide where you came from: a Western-sounding name and a buried origin, on the quiet hope nobody checked. Anker, the electronics giant out of Shenzhen, sounds vaguely German. Shein spent years letting shoppers assume it was anything but Chinese. Urban Revivo is still routinely taken for French. CHAGEE is doing the opposite on purpose, and I think this represents a change in the cultural perception of China.

The founder’s story is almost too neat to be true. It is. Zhang Junjie could not read until he was 18. Orphaned at 10, a milk-tea apprentice at 17, he rang the opening bell at the NASDAQ eight years after starting his own company and became, on paper, a billionaire in his early thirties. He is himself a piece of the new China, and he built a brand that wears where it is from like a flag rather than a secret.

I sat down recently with Emily Chang, CHAGEE’s chief commercial officer. She splits her life between Los Angeles and Shanghai, and explains America to China and China to America with equal ease. She left a CEO role and took a step down in title to join Zhang. When I asked her how a Chinese brand cracks the West, she started with where the tea comes from. “We are proudly from Yunnan,” she said, the province where tea was born.

That word, Yunnan, is the whole strategy. CHAGEE is selling one specific province, right down to the leaf and the 2,700-year-old legend on the cup. Here’s what I actually think. When a brand leads with its province instead of the whole country, that’s a sign it actually understands China. The West loves to flatten the place into a single word. It’s really dozens of cultures, each with its own food and its own traditions. So when I watch CHAGEE go out as proudly Yunnan instead of generically Chinese, that’s the shift I can’t look away from: Western curiosity has finally grown specific enough to reward it. The easy move when going global is to sand all that down to something smooth and placeless. CHAGEE refuses to do this.

And the timing is deliberate. Something is shifting in the West that its own consumers are driving. China has thrown its doors open, visa-free access now reaching 77 countries and visa-free arrivals up nearly 50 percent in a single year. Online, young Westerners have built a whole “chinamaxxing” trend on top of it, drinking warm water, praising high-speed rail, taking up tai chi, and posting billions of views under tags like #becomingchinese. The interest that once stopped at the takeaway menu now books a flight and comes home wanting the real thing. Chang’s rule for catching it is plain: meet the customer where they are, hand them something they already half-recognise, and let them walk toward the rest. That is the opening, and CHAGEE means to be waiting on the other side of it.

None of this is won yet, and I would be lying if I made it sound like it is. CHAGEE has more than 7,000 tea houses, and all but a few hundred of them are still in China. The stock has had a rough ride since that opening-day pop. This is a brand at the very start of a hard road, and plenty of Chinese names with better numbers have tried to go global and stalled.

But the direction of travel is what I’d watch. Here’s my bet. The next wave of global brands shows up proudly, unmistakably foreign, and asks you to meet it halfway. That old reflex, where a foreign brand has to assimilate before you’ll take it seriously? That’s the one that loses now. Curiosity is the edge.

And they will keep coming, because a country moving as fast as China still produces founders like Zhang at a rate the West has mostly forgotten. Not because China is fairer, because it is not. Get born in the wrong village there and the ladder is just as short. But the velocity is real, and so is the confidence that now travels with it. A brand that would once have hidden its origin to be taken seriously abroad now leads with it. That shift is worth understanding, whether or not you ever buy the tea.

Watching which Chinese brands crack the West, and how, is a lot of what I do. This September I am taking a small group into Beijing to sit with the founders, researchers and policy people building China’s innovation engine up close, the same engine that keeps turning out companies like CHAGEE. If you would rather see where the next wave comes from than read about it later, come with me.

6–9 September 2026 · Beijing

Claim A Seat

A few weeks ago, a Chinese company called UBTech put a humanoid robot on sale with one stated purpose: to keep you company. It speaks in a soft voice. If it senses you are tired or stressed, it softens further. It reminds your mother to take her pills and remembers what she tells it. It is built to be a presence in the room.

By launch day, 13,361 people had pre-ordered one. Not a factory arm or a household maid. Company. The entry model runs about US$16,600 and the top tier model closer to US$140,000, and UBTech is selling it to adults for a single purpose: emotional companionship.

The loneliness that smartphones built

Modern life got lonelier over two decades, much of it through the smartphones in everyone’s pocket. Now the fix is arriving as a robot.

The numbers are not subtle. About half of American adults report feeling lonely. The US Surgeon General puts the health toll at around fifteen cigarettes a day, which is a strange thing to know and to see printed on no packet anywhere. People who spend more than two hours a day on social media are about twice as likely to feel isolated. The tools sold to connect us did some of the connecting, and a good deal of the isolating.

Why China feels it first

China feels this first, and harder. It now has more than 100 million single-person households, a figure heading for 150 to 200 million by 2030. But look at who is actually buying: not the over-sixties the headlines reach for, but people 35 to 55, urban, high income, and living alone. This is the first generation raised on the internet and the mobile phone, the very devices that quietly thinned out everyone’s social life, so it is no surprise they are first in the queue for the technology now sold to patch it. Companionship got productised here first for a reason. China's emotional-companionship market is on track to grow from about US$530 million in 2025 to US$8.2 billion by 2028. That is roughly fifteen times bigger in three years.

Loneliness, now in a shopping cart

Emotional value has become something you can put in a shopping cart, with a price and a delivery date in September. Loneliness is now a market, and China has just built the most literal version of it.

Don’t bet against it

I know how this sounds. A humanoid robot that offers companionship sounds dystopian and a long way from your average consumer. But I would be careful betting against this technology. Plenty of people swore they’d never carry a device that tracked their every move, and now feel naked without it. Camera glasses were creepy in 2014 and took the better part of a decade to stop being a punchline. When a technology is that useful, “I would never” has a way of ageing into “how did I live without this.” A robot that keeps your father company while you are three time zones away is exactly that kind of useful. Ten years from now, one in the corner of an office or a living room will not look strange at all.

Where I land

So here is where I land, at least for now. The loneliness is real, and the demand is honest, and I will not sneer at anyone who buys one of these for a parent who lights up when it walks in. But I am not putting a machine that copies my mum’s face and learns her every mood into my home until there are real rules for that data. This is emotional AI: it reads more than twenty emotional states and gets to know you over time. That is not a gadget. It is the most intimate record a person has, handed to a company, in the one place you are meant to be able to drop your guard.

And that is exactly why the data question cannot wait. If these were a passing fad, we could argue the ethics at our leisure. They are not. The adoption curve is already bending, and the demographics only point one way. The easy response is to call it creepy and look away. I think that is a mistake. While the market is still small, we should be deciding what a company may do with the record of how you feel.

The one thing I can’t put in a newsletter is what it’s like to stand in front of one of these: a face that softens when it reads yours, an engineer explaining how it learns you, and the people down the road who decide what any of it is allowed to remember.

That’s why I run the China Innovation Tour. In December I’m taking a small group into Shenzhen and Guangzhou, the hardware heart of all this, where the robots, the factories and the policy sit in the same few square miles. Six days inside the machines and the people building them, months before any of it reaches the West.

If this piece made you want to see it rather than read about it, come with me.

6–11 December 2026 · Shenzhen & Guangzhou · capped at 25 leaders.

Claim A Seat

The old way, that is weeks of work. Combing supplier listings, emailing factories, asking about minimum order sizes, waiting on samples, going back and forth on timelines. Mike had run that routine for years. He knew exactly how slow it was.

This time he didn’t email anyone. He opened an app, described the flashlight he wanted, and let it go find the supplier. A month later he relaunched it at US$2.50 a unit, down from the US$17 it used to cost him.

The app was Accio, Alibaba’s sourcing agent. You tell it what you’re trying to make or buy, and it searches suppliers, compares them, and handles the first rounds of negotiation for you. By March it had passed 10 million monthly users. Roughly one in five people on Alibaba now ask an AI where to source their products, instead of hunting for it themselves.

I watch China’s innovation up close for a living, and I look at a lot of shifts in any given week. This is the one I would put in front of a Western business owner before all the others. Not because it is the flashiest. Because it is the one almost nobody outside China is getting ready for.

Everyone’s watching the wrong half

The story you have probably already read is the consumer one. In China, shoppers have started asking an assistant what to buy instead of searching for it, and the answer comes back as a single recommendation instead of a page of blue links. That is real, it is everywhere, and plenty of people are writing about it.

I am far more interested in the half nobody is watching: procurement. The buying that contractors, small operators, and purchasing teams do every single day. The least glamorous corner of business there is. In most of the world it still runs on phone calls, email chains, and quote forms someone keys in by hand.

In China, that corner started to feel like shopping. Mike’s flashlight is the small, personal version of it. The bigger version is a contractor reordering a pallet of supplies the way you would reorder a pair of sneakers: swipe the feed, tap once, ship direct from the factory. The screen really is that simple. Underneath it, an agent is doing the sourcing and the haggling a person used to grind through by hand.

The part that stays with me is this.

Why I would tell a Western seller to move now

This shift hits you in two places: how you buy your own stock, and how your customers buy from you. Start with the selling side.

Let me be clear about where I land on this, because it is not where the doom headlines land.

Plenty of the coverage frames this as a threat: distributors watching their margins compress, brands going invisible to an algorithm that only reads structured data. That part is real. But I read the balance the other way. The distance between how buying feels in China and how it still feels for most US sellers is the widest opening anyone in this business has been handed in years. Whoever builds the easy version first, in their own market, becomes the standard everyone else gets measured against. That is a good place to be standing.

Here is the scale, so this doesn’t sound like a niche Chinese habit. Analysts call this shift agentic commerce, and China is furthest along on the business side of it. Gartner expects that by 2028, AI agents will handle 90% of business-to-business buying worldwide. That is more than US$15 trillion in purchases moving through software that researches, negotiates, and places the order. It is a forecast about the whole economy, not one country. China is simply where you can already watch it running.

Your buyer is being retrained right now

Think about what this does to the person on the other side of your invoice. Once someone has reordered stock by tapping a feed and letting an agent handle the rest, the old way starts to feel broken. Call the rep. Wait for the quote. Read three PDFs side by side. Nobody unlearns the easy version once they have felt it, and they start looking for it everywhere. That includes looking for it from you.

Most Western sellers can’t offer it yet. A Chinese platform folds finding, comparing, and ordering into one feed and one chat. A lot of US distributors are still working off a phone, an inbox, and a form somebody types up by hand. I know which of those I would rather buy from. So, increasingly, does your buyer.

And don’t assume business buyers are too rational for this. The person who reordered something with one tap last night is the same one who opens your quote form on Tuesday morning, and consumer apps have already reset what buying is meant to feel like. Your process gets measured against that now, whether you offer it or not.

There is a quieter cost sitting under the obvious one. When an agent does the buying, it can only choose from what it can actually read. Give it clear data, clear terms, clear availability, and your product is in the running. Leave those blank and the agent simply cannot see you, with no salesperson in the room to make your case.

This cuts both ways

That was the selling side. The buying side is the one Mike was actually on when he opened Accio. If you are the one doing the buying, the same shift is a gift. Learn to let an agent do the sourcing and negotiating, and you get faster and sharper on every order you place. Keep doing all of it by hand and you end up negotiating against people who aren’t, which is its own slow way of falling behind. The skill is worth building on either side of the table.

Which brings me back to Mike and his flashlight. He didn’t win by being a genius. He won a few weeks of his life back because he tried the new tool before he was forced to. That is the whole opportunity, sitting right there in one small story. The window to build your own version of this is open now, and the people who move while it is early get to set the terms for everyone who arrives late.

Watching this early is a lot of what my work actually is. ChoZan’s China research and consulting exists so the companies I work with see these shifts before they surface in the West, while there is still room to do something with them. And the ones who close this gap first won’t do it from reading a newsletter. They will do it by getting close to the people who already built the thing: the platforms running it, and the regulators drawing the lines around it.

Most of them sit in one city. In September, so do we.

I mean that literally. The one thing I can’t fit into a newsletter is what those rooms feel like: a founder walking you through it before the West has heard the name, a policy team showing you what ships next.

That’s why I run the China Innovation Tour. Four days in Beijing with a small group of leaders, inside the real rooms: Zhongguancun, the labs, the enterprise teams, the people who draw the policy lines. This is the serious side of China’s innovation engine, where research and policy turn into what actually ships.

If this piece made you want to see it rather than read about it, come with me.

6–9 September 2026 · Beijing · capped at 25 leaders.

Claim A Seat

Bernard Arnault didn’t go to Louis Vuitton. He didn’t go to Dior. He went to Songmont — a Chinese label most of his own executives had probably never heard of, where a bag sells for around US$400, the price of a nice dinner in Paris. He bought two, quietly, at the brand’s store in Shanghai. No entourage announcement. No press release. Just the man who has spent four decades teaching the world what luxury means, buying from a brand built to compete with him.

He didn’t say much about it at the time. He didn’t need to.

Late last year, Bloomberg reported that Chinese brands are displacing European luxury across cars, handbags, and jewellery. The trend is real. The headline was already behind.

Arnault saw it first. Most Western brands are still catching up.

Here’s the one number that says it all: Songmont’s online sales rose about 90% last year. Gucci’s online bag sales in China roughly halved. Same market, same shoppers, opposite direction.

The easy read is “Chinese brands are winning, Western brands are losing.” That read misses what’s actually happening.

Chinese households are sitting on more than US$22 trillion in savings — a record high. The wallet is full. What changed is who the spending is for.

For two decades, luxury in China meant showing something to other people — the logo, the status, the proof that you’d arrived. That instinct hasn’t disappeared. But a new one has grown up alongside it: spending on yourself, for yourself, with no audience required.

The clearest evidence isn’t in handbags. It’s in who’s buying.

Men’s average online spending in China has now overtaken women’s — 10,025 yuan a year per capita, according to Boston Consulting Group and Alibaba’s research institute. As marriage and homebuying have slowed, the money that used to go toward courtship and status is going somewhere else: grooming, fitness, gear, collectibles. Things bought for how they feel, not who’s watching.

That same logic explains the small, strange purchases people assumed were beneath a man like Arnault. A US$400 bag is not a lesser version of a US$8,000 one. It’s a different kind of purchase entirely — smaller, lower-commitment, easier to live with if your sense of identity shifts next year.

This is why the “luxury is shrinking” story is incomplete.

Personal luxury goods were the only major discretionary category to contract in China in 2025. Everything next to it grew. Travel. Wellness. Medical aesthetics. Designer toys. Pets.

You can see it on the ground before it ever shows up in an earnings call. Two years ago, the prime ground-floor spaces in Beijing and Shanghai’s biggest malls were lined with EV showrooms. Walk those same floors today, and many of them are gone — replaced by a Pop Mart store doing brisker business than the car dealership ever did, or a humanoid robot from Unitree doing a hip-hop dance to a crowd three rows deep. No analyst report flagged that shift in advance. You’d have had to be standing on the floor to see it coming.

A European brand losing share in China today isn’t just losing to a Chinese competitor. It’s losing to a weekend trip, a skincare clinic, and a collectible figurine that costs less than a dinner out — all fighting for the same kind of spend: personal, emotional, and entirely unconcerned with who’s impressed.

Why this isn’t just a China story

Here’s the part that should actually worry a Western executive: this isn’t a China-only phenomenon. It’s quiet luxury. It’s the experience economy. It’s younger buyers everywhere choosing a trip over a logo. China just got there first — and faster — which makes its mall floors less of a foreign curiosity and more of an early screening of a film about to open in your own market.

The brands that will win the next decade in London, New York, and Paris are the ones paying attention to what’s already playing out in Shanghai — not the ones waiting for next quarter’s report to tell them what they could have seen on the ground floor today.

Arnault, to his credit, seemed to understand this before most of his peers did. He wasn’t slumming it in a US$400 handbag shop. He was doing research.

The brands still waiting for the next quarterly report to tell them what’s happening are going to keep finding out a year late — the same way the rest of the world just found out about Arnault’s handbags.

The shifts reshaping China’s mall floors today are previewing what’s coming to yours. Our China Innovation Tech Tour puts you on the ground to see it before it shows up in someone else’s earnings call.

Get The Full Mission Brief For Our Next China Tech Tour Below

China Innovation Tour Beijing

He was wrong, and he knew it within the week.

What he was actually watching wasn’t a lesson in gambling. It was a lesson in money — taught early, taught lightly, taught the way most Western parents teach their kids to ride a bike. Win a little, lose a little, learn what risk feels like before it costs anything that matters. By the time that boy is running a business, he won’t fear capital. He’ll have grown up fluent in it.

Not the chips. Not the policy. Not even the engineers. Siu co-founded Animoca Brands, the Hong Kong firm behind The Sandbox, where he is now executive chairman. It holds stakes in more than 400 Web3 companies and is pursuing a Nasdaq listing through a reverse merger, reportedly targeting a valuation of up to US$5 billion. For the last year, Siu has been making a case most Western tech coverage keeps missing: AI agents will need to transact, negotiate, and hold value on their own. Whoever is most comfortable handing that job to a machine wins the next decade. And comfort with money, he says, isn’t evenly distributed.

He has a small, slightly embarrassing story to prove it. New in Hong Kong, he once asked European friends what their rent or property cost. Standard small talk, where he was sitting. They didn’t know. Or worse, they seemed to find the question itself a little rude — as if caring what something cost made you crass.

Ask someone their rent and you’re not prying. You’re connecting. Siu’s read on this gap isn’t that Asia is more materialistic — it’s that money, handled openly and early, stops being frightening. People who grow up asking “what’s your rent” without flinching are the same people who’ll hand an AI agent a wallet without flinching. People who were taught not to talk about money are the ones who’ll hesitate right when the agentic economy needs them not to.

This is also why, in Siu’s telling, Western tech platforms increasingly resemble the command economies he watched collapse as a child in Cold War Europe. Facebook’s ad auction sets a price nobody can question or negotiate. The platform simply decides what your reach is worth — the same way a Soviet ministry once decided what a loaf of bread should cost. He’s not being cute with the comparison. He means it. Closed system. Opaque pricing. No real market underneath it. The agentic economy, in his pitch, is what finally breaks that — agents negotiating directly, in the open, the way a market should work.

None of this is abstract for much longer.

Siu personally runs close to 300 AI agents today, handling tasks he’d have hired a small team for two years ago. He expects most of the connected world to be managing dozens within a few years. The agents won’t sleep, won’t take holidays, and won’t hesitate to move value at 2am if the deal is good. The only question, in his framing, is whether the people building and managing them are already comfortable with money — or whether they’re still working through it while the agents move on without them.

That kind of comfort isn’t something you pick up from a distance. You don’t absorb it by reading about Sha Tin. You absorb it by being somewhere it’s normal. In Berlin, that conversation is faintly taboo. In Hong Kong, it’s just Tuesday. Twenty minutes from a border that sets the pace on payments, manufacturing, and everything else. There’s a real distance between “knowing about” this shift and being fluent in it. Our China Innovation Tour is built to close that distance.

ChoZan has been helping global businesses learn from China since 2016, through innovation research, boardroom briefings, and immersive China study tours into the country’s tech and consumer landscape.

This is part of what I get to do a few times a year as part of ChoZan’s ongoing China innovation tour — sit down directly with the people actually building this future, not just read about them.

Want to learn more? Get the full mission brief below

China Innovation Tour Beijing

Since 2014, President Xi Jinping has delivered an annual New Year address to China’s 1.4 billion people: each speech a carefully calibrated signal of government priorities. This year’s December 31, 2025 address stands out. Bloomberg called it Xi’s “most upbeat New Year’s Eve address in recent years,” and the tone shift is strategic.

What’s different? Xi omitted the “external uncertainties” language that dominated his 2025 speech. Instead, he declared 2026 as the launch year for China’s 15th Five-Year Plan (2026-2030), the blueprint that will direct over $2 trillion in capital allocation toward green energy, AI, and advanced manufacturing.

The timing creates a decision window. President Trump visits Beijing in April 2026, following the October 2025 Busan summit that produced a fragile tariff truce. For executives making China investments, market entry, supply chain, or partnership decisions in the next 90 days, understanding these signals before policies solidify is essential.

👉 Ready to see it all first hand? Join us in April for an immersive China Innovation Tour. More details below https://chozan.co/china-inspiration-tech-tour/

In this issue, we’re breaking down 5 key signals:

1. Economic milestones and the US$19.7 trillion GDP threshold

2. AI and chip breakthroughs - plus how Chinese AI is going global in three ways

3. Mega-infrastructure investments in space, energy, and defense

4. Robotics and drones scaling from demos to industrial deployment

5. The 15th Five-Year Plan’s concrete milestones for 2026

Below, we break down these key signals from the speech—what Xi said, what it means, and what you should do about it.

1️⃣ Economic Scale Hits New Heights: 140 Trillion Yuan GDP

📜 “Our economic output has crossed thresholds one after another, and it is expected to reach RMB 140 trillion yuan this year.”

ChoZan Analysis: Xi’s confidence isn’t just rhetoric. Despite tariff speaking at 145%, China’s export machine found new markets across ASEAN and the EU. The speech frames this as validation of “Dual Circulation,” Beijing’s strategy to boost domestic consumption while maintaining manufacturing dominance.

What’s absent: No mention of the property crisis, deflation, or youth unemployment. Beijing’s 2026 narrative will emphasize strength over structural challenges.

🎯 Your Decision Trigger

Who: CFO, Head of Strategy, China GM

The choice: Companies still positioned for mass-market volume will face margin compression. Those repositioned toward premium segments with localized operations can access green incentives worth 10-15% cost advantages.

Action: Brief your board on two scenarios: continued bifurcation vs.stabilization post-April. Build H2 2026 budget flexibility for either path.

2️⃣ AI and Chips: Self-Reliance Becomes Reality

📜 “Many large AI models have been competing in a race to the top, and breakthroughs have been achieved in the research and development of our own chips.”

Key data points:

• DeepSeek’s January 2025 breakthrough delivered powerful AI despite US chip curbs.

• Semiconductor exports rose 24.7% year-over-year.

• HarmonyOS now runs on 1.2 billion devices worldwide.

• Meta acquired Chinese-founded Manus AI for US$2 billion in December 2025.

ChoZan Analysis:

Chinese AI is globalizing through three vectors. First, US tech giants are acquiring Chinese AI startups. Second, Hong Kong IPOs are bypassing US capital restrictions. Third, Alibaba Cloud is expanding AI SaaS across Southeast Asia and the Middle East.

Your “China-only vs. global” AI infrastructure assumptions may need revision.

🎯 Your Decision Trigger

Who: CTO, Chief Digital Officer, Head of Procurement

The choice: If you operate in China with US AI infrastructure, stress-test what happens when Chinese alternatives reach feature parity. Alibaba and Huawei Cloud pricing is already 20-30% below AWS in-region.

Action: Map your current AI and cloud dependencies across China vs. global operations. Identify single points of failure.

👉 Want to see this transformation firsthand? Join Ashley in China this April. Early bird has flown, but the opportunity to shape your 2026 China strategy hasn’t. 👉 More here https://chozan.co/china-inspiration-tech-tour/

3️⃣ Mega-Infrastructure Projects: “New Quality Productive Forces” in Action

📜 “Construction of the hydropower project at the lower reaches of the Yarlung Zangbo River started... The Tianwen-2 probe began its star-chasing journey... China’s first aircraft carrier equipped with an electromagnetic catapult system was officially commissioned.”

Key data points:

• The Yarlung Zangbo mega hydropower project costs US$170 billion.

• This is four times the cost of the Three Gorges Dam.

• At 60 GW capacity, it will generate enough power to run the entire UK.

• Completion is expected in the mid-2030s.

• China’s data center power demand is projected to increase 3.5x by 2030.

ChoZan Analysis:

This infrastructure trifecta shows how “New Quality Productive Forces” translates to capital allocation. The hydropower project solves a real constraint: AI data centers and advanced manufacturing require massive, stable electricity. Aerospace and defense mentions signal dual-use technology with long-term commercial applications.

🎯 Your Decision Trigger

Who: Chief Supply Chain Officer, Head of Energy Procurement

The choice: China’s clean energy buildout creates both opportunityand risk. Western green tech firms have a partnership window now. But energy-intensive industries in China will soon access subsidized power that competitors cannot match.

Action: Model the energy cost impact on 10-year manufacturing competitiveness, both for your China operations and your Chinese competitors

👉 Ready to see it all real time? Join us in April for a China Innovation Trip. More details here https://chozan.co/china-inspiration-tech-tour/

4️⃣ Robotics and Drones: From Demos to Deployment

📜 “Humanoid robots did kung fu kicks, and drones performedspectacular light shows.”

Key data points:

• Chinese firms target 1 million+ humanoid robots in factories by 2027.

• DJI holds roughly 70% of the global drone market.

• China exported 6.5 million EVs in 2025, up 1 million units from 2024.

• Robotics is projected to contribute US$1 trillion to GDP by 2030.

ChoZan Analysis:

Xi’s playful framing undersells the strategic intent. Robotics addresses China’s most pressing challenge: a workforce shrinking by roughly 5 million people annually. Beijing’s solution is automation, not immigration. Chinese robotics firms now offer turnkey factory automation at prices 30-40% below European and Japanese competitors.

The question isn’t whether Chinese robots will be competitive. It’s when.

🎯 Your Decision Trigger

Who: COO, Head of Manufacturing, Workforce Planning

The choice: Labor cost arbitrage, the traditional reason to manufacture in China, erodes as automation scales. But automation itself creates new arbitrage through aggressive Chinese robotics pricing.

Action: Request a competitive analysis of Chinese robotics solutions (UBTech, Unitree) against your current automation vendors. Reassess whether “low-cost labor” still belongs in your China value proposition.

5️⃣ The 15th Five-Year Plan Launches

📜 “The year 2026 marks the beginning of the 15th Five-Year Plan...We should focus on our goals and tasks, boost confidence, and build momentum to press ahead.”

Key priorities emerging:

• Boost Domestic Consumption: Target an increase in consumption share of GDP from 38% to 60%.

• Scale R&D Investment: Raise R&D spending from 2.7% to over 3.2% of GDP by 2030.

• Expand Clean Energy: Generate 500+ TWh of clean electricity from Tibet by 2050.

• Revive Foreign Investment: Initiate selective reopening following a 15% decline in FDI.

Critical window:

FYP drafts are still circulating for industry consultation through early 2026. Companies that engage now can shape provisions. Those who wait will adapt to rules written without them.

🎯 Your Decision Trigger

Who: CEO, Head of Government Affairs, Corporate Development

The choice: Your industry association may have input channels to 15th FYP sector provisions. If you’re not engaged, you’re already behind competitors who are.

Action: Ask your China government affairs team one question: “What is the 15th FYP consultation timeline for our sector, and have we submitted input?”

Xi’s speech signals a China confident in its trajectory, perhaps overconfident given the challenges left unmentioned. But for executives, the message is clear: Beijing believes it has weathered the tariff storm and is ready to launch an ambitious new cycle.

Three characteristics of companies that will thrive:

1. Localized operations that access domestic incentives rather than fight them.

2. Flexible strategies ready for either continued tension or post-April détente.

3. Active policy engagement that shapes 15th FYP rules before they solidify.

The question isn’t whether to engage with China’s transformation. It’s whether you’ll shape it or merely react.

👉 Join us in April to see it all firsthand. More details here https://chozan.co/china-inspiration-tech-tour/

As we wrap 2025, the gap between the headlines we read and the reality on the ground is real.

When you step off the plane in Shanghai or Shenzhen today, something feels different. Not “catching up” different—more like “10 years forward” different. You pay with your face. Robots deliver room service. Your taxi has no driver. The subway uses AI to predict crowding. And everyone seems genuinely excited about it.

Yet beneath the optimism: youth unemployment’s at 17%, property markets dragging, consumer confidence fragile. The same country installing more robots than the rest of the world combined can’t create enough jobs for its graduates?

This is China’s paradox: a nation simultaneously building the future and struggling with the present.

But here’s what visiting reveals that reports miss: frontier entrepreneurs in AI, robotics, and biotech are thriving. The same ecosystem that went from 1 novel drug approval in 2005 to becoming a biomedical powerhouse is now racing ahead in efficiency-focused AI, humanoid robots, and global EV expansion.

That’s why we’re taking you there > https://chozan.co/china-inspiration-tech-tour/

Meanwhile, we analyzed expert interviews, policy announcements, and industry forecasts from top China analysts. These 20 trends have the highest conviction—and the biggest implications for global business.

In this final issue of 2025, here are 5 things to watch in China tech in 2026.

1️⃣ AI Deployment Accelerates: From Robots to Factories

🤖 What to Watch:

Humanoid Robots in Homes, Hotels & Hospitals

→ Shanghai building humanoid robot training base for 1,000 robots by 2027

• The eldercare robot market is growing 15% annually through 2029.

• Market size will double from US$1.1 billion in 2024 to US$2.2billion by 2029.

• This is the fastest-growing service robot segment in China.

• Robots will handle companionship, health monitoring, medication reminders, and emergency response.

→ Mass-market pricing hits mainstream

• Chinese humanoids cost US$5,900–$20,000. Western competitors like Tesla Optimus charge US$20,000-30,000.

• Over 150 startups are competing, which drives costs down through “involution”.

• Beijing warned the sector is “drunk” and there’s a real bubble risk.

• Robots are entering eldercare facilities, hotels, and retail stores.

🧠 What to Watch:

Practical AI Everywhere (Not Just Chatbots)

→ Conversational AI becomes the default interface

• Shenzhen Metro has been running AI robots for deliveries and guidance since August 2025.

• Shanghai is testing over 900 autonomous vehicles on public roads.

• 80% of Chinese brands now use AI-first customer service. TheUS sits at 65% and the global average is 54%.

• Beijing and Shenzhen are following Shanghai’s playbook.

→ Mass deployment from factories to villages

• Smart factories: Hyper-automation is cutting costs by 50%.

• Companies that stay manual will pay a 15-20% cost premiumby Q4 2026.

• Hospital AI: 59 AI medical devices have been approved. 50+ national standards are coming by 2026 for imaging and dosing.

• Rural areas: 5G and fiber reach almost all villages now. US$70AI devices are being used for farming and small manufacturing.

2️⃣ Next Mobility: Robotaxis & Flying Cars Get Real

🚕 What to Watch:

Robotaxis Expand Beyond Pilot Cities

• Baidu Apollo Go robotaxis are 30-70% cheaper than human-driven taxis.

• Base fares are US$0.55 versus US$2.50 for traditional taxis.

• Commercial operations are expanding beyond test zones.

→ 2026 watch: Expansion to more cities

• Robotaxi zones will expand beyond Wuhan, Shenzhen, andBeijing.

• Government support and EV supply chains provide cheap sensors that enable scale.

✈️ What to Watch:

“Low-altitude economy” becomes policy priority

→ China is targeting electric vertical takeoff and landing (eVTOL) aircraft for urban mobility

• Infrastructure is being built: vertiports, air traffic management systems, and charging networks.

→ 2026 milestones to watch:

• First commercial urban air mobility routes will likely launch on the Shenzhen-Guangzhou corridor.

• Regulatory frameworks will be finalized.

• EV supply chains (batteries and motors) are being repurposed for eVTOLs.

→ The advantage: China’s existing EV supply chain enables cheap eVTOL production

• The same batteries and motors used in EVs work for flying vehicles.

• This gives China a head start in the low-altitude economy.

To understand China, you need someone on the inside.

Ashley Dudarenok isn’t a China analyst. She’s a naturalized Chinese entrepreneur who lives, builds, and thrives in this ecosystem every day.

3️⃣ AI Infrastructure & Applications: The “Superhighway” Advantage

🛣️ What to Watch:

AI Becomes a Rental Utility, Not Owned Infrastructure

→ China’s “AI superhighway” model is different from the West

• China deployed massive computing power, 7 times the original target, across 250+ data centers.

• Many data centers are located in robotics industrial parks.

• These parks rent GPU processing to nearby factories by thehour.

• Small manufacturers don’t buy expensive hardware. They rent AI like electricity.

→ AI is reaching China’s remotest villages

• 98% of villages now have 5G or fiber internet.

• US$70 edge-AI devices plug straight into this network.

• These devices enable instant inventory checks and crop monitoring.

• 2026 watch: Widespread AI rollout in remote agriculture and small manufacturing with real-time decision-making tools.

→ Why centralized infrastructure beats fragmented approaches

• US model: Big Tech companies like OpenAI, Google, and Anthropic are building private computing empires. This creates innovation but also fragmentation.

• China model: Government-coordinated infrastructure enables faster nationwide deployment.

• 2026 test: Who will embed AI in healthcare diagnostics and logistics networks first?

• Prediction: By 2027, China will capture 25% of global AI server demand. The centralized approach wins on speed for industry-specific AI tools.

→ US-China safety collaboration moves from talk to test

• AI safety dialogues are continuing through APEC 2026.

• Both countries regulate AI domestically and share best practices. This is called “safety in parallel”.

• Shared safety benchmarks may emerge through broader collaborations.

Subscribe now

4️⃣ China Going Global 2.0: Exporting Ecosystems, Not Just Products

🚗 What to Watch:

EV Assembly Moves Overseas

→ China’s battery overcapacity is forcing a new strategy

• Battery production is 2× domestic demand and 1.2× global demand.

• Idle capacity equals 5-10 million vehicles per year.

• Solution: Move final assembly abroad to dodge 25-45% tariffs.

→ Where they’re building in 2026:

• BYD is opening plants in Hungary (Q1 2026), Brazil, and Turkey.

• NIO is scouting locations in Spain and Mexico.

• The strategy: Ship 75% of parts from China, assemble locally, and stamp products “Made in Hungary” or “Made in Mexico”.

• 2026 watch: Exports will surge beyond 6 million vehicles annually by the late 2020s.

→ What this looks like in practice:

• BYD’s Changzhou plant builds a Seagull EV in 90 minutes.

• Labor costs are 48% below Tesla’s Fremont factory.

• The same automation is heading to BYD’s Hungary plant.

📦 What to Watch:

30-Minute Delivery Infrastructure Goes Global

→ “Chuhai 2.0” means exporting complete business models

• Ultra-fast delivery systems perfected in China are expanding to Southeast Asia, Latin America, and the Middle East.

• This isn’t just hardware. China is exporting automated warehouses, routing algorithms, and operational playbooks.

• 2026 prediction: Major logistics firms will launch full ecosystems abroad and reshape local retail.

→ Chinese service robots are flooding overlooked markets

• Service robot exports grew 50%+ in 2025.

• Shipments are increasing to Mexico, Turkey, and Nigeria.

• Chinese service robots are priced around US$300, which is up to 40% cheaper than Korean competitors.

• These are markets Western firms ignore. That’s China’sopportunity.

→ 2026 indicators to watch:

• SF Express and JD Logistics are opening international branches.

• 30-minute delivery will launch in new countries.

• Watch for hybrid models that blend Chinese tech with local adaptations.

5️⃣ Full‑Stack Self‑Sufficiency: From Software to Cancer Cures

💻 What to Watch:

Software Independence Accelerates

→ Beijing is pushing software independence under the 2026-2030 plan

• Provincial government clouds are increasingly running domestic alternatives.

• Currently 60% use homegrown systems, up from 35% last year.

• Kylin OS and Alibaba’s Longsoar database are replacing Microsoft and Oracle.

• Switching to homegrown software is yielding significant cost savings by cutting Microsoft, Oracle, and Red Hat licenses.

→ Chip design software is going domestic

• SiCarrier’s tools were launched in October 2025 and have been adopted by 20,000+ engineers.

• Orders for domestic chip design software are growing significantly.

• AI chip output is tripling.

• Why the urgency: Trump flipped chip export rules 3 times in 4 months. CEOs now assume “If it’s US-made, it can vanish overnight”.

• 2026 prediction: AI chip production could triple despite US restrictions.

🧬 What to Watch:

Biotech’s “DeepSeek Moment”

→ China surpassed the US in active clinical trials in 2024

• China’s pharma out-licensing deals surpassed US$100 billionin 2025 alone. That’s up from US$28 billion in 2022.

• Beijing is targeting a US$280 billion biomanufacturing sector by 2033. This aligns with Healthy China 2030 goals.

→ US-China collaboration opportunities are expanding

• China offers patient scale. The US brings capital. Together, this means faster cures.

• AstraZeneca’s China collaborations are growing. For example, there’s a US$100 million deal with Jacobio.

• 2026 watch: Joint clinical trials and shared investments are accelerating.

🌱 What to Watch:

Climate Tech “Talk Less, Do More”

→ China focuses on concrete capacity, not policy speeches

• China produces 80% of solar panels, 75% of EV batteries, and 60% of wind turbines globally.

• Beijing avoids “climate leadership” rhetoric and focuses on building.

• 2026: Aggressive expansion into Latin America, South Asia, and Africa with affordable tech.

→ Next frontiers to watch:

• Green hydrogen infrastructure buildout.

• Smart grid upgrades that enable renewable energy integration.

• Prediction: Renewables will hit 33%+ of the global power mix by 2026. China’s tech is enabling transitions worldwide.

→ Why this matters:

• Industrial capacity compounds over decades.

• This is leadership by accumulation, not proclamation.
  

China feels 10 years ahead, and it’s not hype. While others debate strategy, it deploys at scale. Reading about it isn’t enough. You have to see it firsthand. Join me in April.

• Policy Worked: Government stimulus drove consumer spending to 2019 highs. State-backed sectors delivered—DeepSeek cut AI costs 30%, BYD exports surged 150%, 150+ humanoid companies emerged. Innovation was real and measurable.

• Markets Punished: Tech earnings forecasts crashed from +3% to -9%. Premium positioning failed—NIO, Li Auto, Western retailers all missed badly. Beijing warned against robot overcapacity.

The paradox? Innovation accelerated while profits collapsed.

Hello there,

December 2025 was the deadline Beijing circled in red — the finish line for “Made in China 2025”. (Read More)

On the surface, it was a victory lap. Official channels declared near-total industrial independence, citing dominance in EVs, batteries, and AI cost innovation. But beneath the headlines, the reality is more nuanced.

While official narratives claim 100% self-sufficiency, three evident challenges emerged:

• The “Robot Bubble”: Over 150 humanoid robot companies flooded the sector, triggering official warnings of “blind expansion” and overcapacity.

• The Semiconductor Gap: Advanced chips remain import-dependent. This week’s twist: Trump reversed his chip ban, allowing Nvidia H200 sales to “approved” customers (25% to US coffers). Beijing’s silence speaks volumes—do they even want US chips anymore?

• The Profit Paradox: Tech earnings forecasts have crashed (from +3% to -9%) even as AI valuations soar, a dangerous divergence between story and substance.

Strip away the noise, and 2025 proved three absolute truths: Scale won. Exports won. Cost efficiency won. Meanwhile, premium strategies and legacy brand power were quietly crushed.

As we turn toward 2026, the landscape resets. New tariffs, AI agent regulations, and the 15th Five-Year Plan will redefine the rules. This year-end newsletter separates signal from noise:

• The Tech Reality: Who actually made money in 2025?

• The Consumer Shift: Why “Premium” is a dirty word.

• The 2026 Playbook: 5 Dates you need on your China calendar in 2026.

What Happened in China Tech This Year?

🗺️ Big Picture Shifts

The big winners weren’t the premium brands—they were the companies that scaled globally and cut costs aggressively.

📊 The Numbers:

• BYD shipped 410,000+ vehicles overseas (+150% year-over-year)

• CATL captured 37% of global battery market share

• DeepSeek launched AI at 30% below Western competitors

• NIO delivered 55,000 vehicles vs 70,000 target (margins dropped 2.6 percentage points)

• Li Auto cut guidance from 480,000 to 380-400,000 units

🤔 What It Means:

1. Export Scale Survived
   BYD makes sub-8% margins in China but charges 1.8x prices in Europe—global reach saved them. CATL dominates by being everywhere, not by being premium. Companies that went international had pricing power. (Read More About China’s Industries That Scaled Globally)

2. Domestic Premium Failed
   NIO and Li Auto both missed targets badly as Chinese consumers shifted to “good enough” options at lower prices. Xiaomi sold 20,000 EVs monthly at 60% of Tesla’s price. Premium positioning without export strategy = margin death.

3. “Boring” Tech Won Big
   China is betting on mature semiconductors (cars, appliances, not cutting-edge AI chips) and mass-market robots ($9,000 humanoids vs $50,000+ Western models). They’re dominating commoditized technology, not breakthrough innovation. Your supply chain is becoming more Chinese whether you planned for it or not.

💡Key Takeaway

1. Everything just got cheaper – AI, chips, batteries.

2. They’d rather sell abroad than at home – so import from China, don’t try to sell into it, unless you’re moving millions of units.

Stop reading about China’s tech transformation. Experience it firsthand with Ashley Dudarenok:

While tech companies battled margin compression and tariffs, Chinese consumers were rewriting the rules of what “premium” means. Here’s what changed at the consumer level:

How Chinese Consumers Changed In 2025

📊 Big Picture Shifts

Consumers stopped aspiring to premium Western brands and started optimizing for value + technology. “Emotional value” (products that improve daily life, calm stress, or save time) replaced status symbols.

📊 The Numbers:

• Tmall Double 11: AI assistants used 1.5B times, outdoor sports sales +52%, delivery times cut 25% （Read More About China’s Consumer AI)

• PDD/Temu: 869M active buyers, revenue $20.4bn (+58% YoY)

• Gen Z: 37% of China’s discretionary spending; 61% buy for “emotional utility” not status

• Pet economy: US$10.6B (+22% YoY)—more pets than kids under 4

• Repair boom: Phone repair +35.7% YoY; 40% of Gen-Z sold second-hand in 2025

🤔 What It Means:

1. Tech + Value Beats Brand Prestige
   Anta sold 2.1M smart yoga pants by embedding sensors, not logos. PDD hit 869M buyers with AI-driven bargains. Tmall used AI 1.5B times during Double 11 to personalize shopping and cut delivery 25%. If your product doesn’t use tech to save time or money, consumers ignore it.

2. Western Brands Fail Without New Value Props
   Walmart and Gap closed dozens of stores. But Sam’s Club thrived (11 new warehouses, 30% bigger baskets) by offering bulk value + experience. Premium pricing for Western logos alone is dead—you need a reason beyond heritage.

3. Gen-Z Rewired Consumer Behavior
   They’re 37% of spending but prioritize emotional utility (61%), circular economy (40% sold second-hand), and pets over kids (US$ 10.6 B pet market). If you’re not addressing stress relief, sustainability, or companionship, you’re missing the biggest spenders.

💡Key Takeaway

Chinese consumers still spend, but only if you save them time, money or emotional calories—brand logos alone no longer count as currency.

Four Strategic Pivots That Will Define 2026

🤖 1. AI Agent Store Wars

The Shift: In March 2025, Tencent launched Yuanbao—an AI assistant built directly into WeChat. Within one week, it overtook DeepSeek to become China’s #1 AI app. Why? WeChat added a one-tap download button for its 1.3 billion users. Standalone AI apps can’t compete with platform distribution.

Now Beijing is drafting rules (expected March 2026) to force all super-apps to open their APIs, turning WeChat, Alipay, and Douyin into AI “app stores”.

❓Your Question: Are you building for WeChat, Alipay, or Douyin? In China, owning your own distribution doesn’t work—you need to play inside the platforms.

💰 2. From Coupons to Cash (Stimulus 2.0)

The Shift: The 2025 voucher programs (phone trade-ins, appliance rebates) have ended. For 2026, Beijing is switching tactics: instead of vouchers, they’re testing direct cash drops. Pilot cities got ¥1,000-2,000 loaded onto digital wallets for appliances, EVs, sports gear, and eldercare tech.

The March 2026 National People’s Congress will finalize a 4% deficit plan with US$ 8.33 billion in bonds aimed at putting money directly into consumers’ hands.

❓Your Question: Are your products in the right categories—green appliances, EVs, sports tech, eldercare tech—to catch this wave?

🌍 3. Make It in China, Sell It Abroad

The Shift: Chinese companies invested more overseas than domestically in 2025 for the first time ever. They’re escaping sub-8% margins at home and dodging 45% EU tariffs (100% in the US) by moving final assembly abroad.

Battery-makers are leading—CATL, Envision, and Gotion are building gigafactories in Hungary, Turkey, and Morocco. Assembly plants are multiplying: Chery in Spain, XPeng partnered with Magna in Austria, JAC building in Mexico. The playbook: ship parts from China, turn the final screws locally, stamp it “Made in Mexico/Spain,” avoid tariffs.

Beijing is accelerating this. Starting January 2026, companies need export permits for finished EVs—the message is clear: assemble abroad, not at home.

❓Your Question: Have you modeled a 20-30% tariff hit? Should you be scouting Hungary, Mexico, or Turkey for the same assembly strategy your Chinese suppliers are already booking?

🎯 4. The 40% Price Rule

The Shift: Chinese competitors are proving they can deliver 75% of your product’s performance at 40% of your cost across almost every category. Unitree makes humanoid robots for $9,000 (Boston Dynamics charges $75,000). Huawei’s Mate 60 Pro costs 65% of an iPhone 15 Pro. CATL’s new sodium-ion batteries are 40% cheaper than lithium.

This isn’t just happening in tech anymore—it’s spreading to appliances, medical devices, and industrial equipment.

❓Your Question: Which of your products could a Chinese competitor copy at 75% performance and 40% of your price in the next 18 months?

(All figures are the latest available as of 8 Dec 2025; YoY = year-over-year.)

📅 Key Dates for 2026

March 5-11 — National People’s Congress announces 15th Five-Year Plan details and new consumer stimulus measures

Late March — Ministry of Industry releases AI agent implementation guidelines and app store rules

April — EU announces final ruling on BYD tariffs, US likely announces reciprocal EV duties

June — DeepSeek launches R2 model with lower costs and better reasoning

July — Zhejiang and Guangdong provinces roll out first humanoid robot purchase vouchers

2025 clarified the new rules. Premium positioning failed, but export pricing worked. Tech costs dropped across every category. Consumers rewarded utility over brand logos. The winners weren’t guessing, they were watching what actually moved.

2026 brings clear timelines and opportunities. Tariff rulings in April and July will reset pricing power. March stimulus targets specific categories you can plan for now. AI platform rules (late March) will define distribution for the next three years. Companies that align their strategy to these known milestones will capture disproportionate advantage.

Share

• China is the first country to operate 2 million industrial robots in its factories. This massive stock advantage sets the new global manufacturing benchmark, accelerating cost efficiency and domestic innovation.

• For decades, foreign firms controlled China’s market (over 70% share). The new 57% domestic market share marks a complete power flip. This pivot confirms China’s self-sufficiency, giving local firms full control over supply chain, cost structure, and the proprietary machine data generated by these robots.

Hey there!

Amazon’s recent holiday strategy made global headlines: deploying up to 600,000 robots for warehouse operations. But these are traditional industrial robots—arms that sort, scanners that track, wheels that transport. Efficient, yes. Revolutionary? Not quite.

Meanwhile, in Shenzhen this November, something fundamentally different emerged. UBTech just announced the mass delivery of its Walker S2 humanoid robots to BYD and Foxconn. These aren’t mechanical arms bolted to factory floors. They are bipedal workers that walk production lines, swap their own batteries, and make autonomous decisions.

With orders surging past 800 million yuan (US$113 million) this month alone, the “pilot phase” is officially over.

Here’s the critical distinction most miss: China has led in volume for years. But this shift to humanoid robots represents a move from Tools (machines we operate) to Agents (machines that operate themselves).

In this issue, we decode the two-track robot revolution reshaping global manufacturing—and why China’s aggressive deployment creates a data advantage the West may never catch.

🤖Your Questions About the Robot Revolution, Answered

⚙️Traditional Robot Workers (The Known World)

Q1: How much can robot workers actually save companies?

Traditional industrial robots deliver 20-40% cost reduction through reduced labor, fewer errors, and continuous operation. ROI typically hits within 2-3 years. China’s robot density now exceeds 470 per 10,000 workers—nearly double the US rate.

Q2: What’s driving mass adoption now?

Three converging factors: AI makes robots smarter (computer vision for quality control), costs have dropped 50% since 2020, and labor shortages have hit critical mass (China faces a 30 million worker shortfall by 2025).

Q3: Which industries lead deployment?

Automotive (932 robots per 10,000 workers globally) and electronics (413 per 10,000) dominate. Amazon’s 600,000 robots handle repetitive tasks, while China’s Foxconn uses 40,000+ robots for iPhone assembly.

🤖Humanoid Workers (The New Frontier)

Q1: Will humanoid robots take my job in the next 5 years?

Not directly. Current humanoids handle specific tasks—materials transport, quality inspection, basic assembly. They are tools augmenting human work, not replacing it wholesale. However, Goldman Sachs projects that by 2030, 30% of manufacturing tasks could shift to humanoids.

Q2: How “smart” are today’s humanoid robots?

They are competent assistants, not skilled workers. Current models like the Walker S2 handle basic tasks (moving boxes, simple quality checks) but struggle with spatial intelligence (judging fit, adapting grip).

The breakthrough: Physical AI. Every robot deployed in a factory generates critical data on failures and edge cases. They get smarter through collective real-world experience, not just lab simulations, making deployment the key to intelligence.

Q3: Why are humanoids game-changers if they’re still limited?

Flexibility solves the Data Collection Paradox. To make robots truly capable, you need massive, real-world data. Humanoids solve this by operating in unchanged human workspaces—same doors, same tools, zero factory retrofitting required. China’s aggressive strategy isn’t waiting for perfection; they are using imperfect robots to quickly build the massive, essential training dataset for the next generation.

So we have a paradox: China’s robots are simultaneously too dumb to replace humans and smart enough to transform industry. How does that work?

The answer lies in understanding that China and the West are playing different games entirely.

The Chinese Playbook vs. Western Strategy

BYD’s Triple Integration

• Deploy: 500+ Walker S1 robots handling materials, qualitycontrol

• Develop: In-house robotics team creating custom solutions

• Democratize: Robot-as-a-Service model for smaller suppliers

Tesla’s Laboratory Approach

Optimus remains in controlled testing, targeting 2026 deployment. Perfectionist engineering, limited real-world data collection.

The Critical Difference: Data Advantage

China’s factories provide the world’s richest terrain for training and iteration—vast, diverse, and increasingly robot-friendly.

Every Walker S2 operating in BYD’s factory generates operational data. Multiply by thousands of robots across hundreds of facilities. China is building an insurmountable dataset for training EmbodiedAI—real-world scenarios Western labs can’t replicate.

The Convergence Play: Chinese EV companies repurpose autonomous driving tech (sensors, AI, batteries) for robotics. Same supply chains, same engineers, accelerated development. It’s ecosystem leverage.

Five Developments That Will Define 2026

Next 3 Months (Q1 2026):

1. Mass Production Milestone: Watch for Chinese manufacturers hitting 10,000+ humanoid units deployed. While UBTech targets 500 by December, competitors like Unitree and Agibot are scaling rapidly.

2. Cost Breakthrough: Unitree’s R1 humanoid just hit US $5,900 (39,900 yuan) —cheaper than a used car. It’s performing cartwheels and handling simple industrial tasks. When robots cost less than three months’ salary, the economics of automation fundamentally change. Government subsidies could push effective costs below US $3,000 by year-end.

Next 12 Months (Through 2026):

1. Service Sector Invasion: Humanoids will appear in Chinese hospitals, hotels, and retail stores. Healthcare and elder care will lead to adoption, moving from “novelty” to “necessity”.

2. Western Partnership Scramble: Major Western manufacturers will announce Chinese robotics partnerships.The choice is stark: partner now, or compete against competitors with 24/7 operations and 50% lower labor costs.

3. The “ChatGPT Moment” for Robotics: China’s advantage in “Physical AI”, models that understand physics and gravity, will create robots that don’t just follow scripts, but truly understand and adapt to their environment.

Whether we’re ready or not, the age of robotic and humanoid workers has begun.

Hello there!

The other week, Presidents Trump and Xi eased trade tensions—rare earth access extended, tariffs cut, tech cooperation discussed. Markets rallied. Executives breathed easier.

By next year, these terms will change again. The “stable state” everyone’s waiting for? It’s not coming.

But here’s what most executives miss: While Western companies pause investments waiting for clarity, Chinese tech companies have been thriving under constant regulatory uncertainty for years.

Gaming regulations shift overnight. Data laws rewrite the rules. Export controls block supply chains. Yet ByteDance, Alibaba, BYD, and CATL keep innovating faster than Western competitors.

The Trump-Xi deal isn’t about who won or lost. It’s a reminder that regulatory uncertainty is the new operating environment—not a temporary phase.

What You’ll Learn In This Issue:

1. What was actually in the “tech cooperation talks”

2. Three scenarios for 2026-2030 about how to build your tech strategy

🤖 What Was in the “Tech Cooperation Talks”

✅ What Was Discussed

1️⃣ Nvidia AI Chips (No Conclusion Reached)

• Trump confirmed they discussed semiconductors and said Nvidia would negotiate directly with China

• However, he explicitly ruled out the advanced Blackwell AI chip: “We’re not talking about the Blackwell”

• Positioned himself as “arbitrator” between Nvidia and China

The translation: Kicked the decision down the road. No concrete chip approvals, no timelines—just an agreement to keep talking.

2️⃣ End of Investigations Targeting US Chip Firms

• China agreed to terminate dumping and antitrust investigations targeting Nvidia, AMD, and Intel

• Eases immediate operational pressures for US chip firms operating in China

• However, broader export restrictions on advanced chips remain in place

The translation: This stops the bleeding but doesn’t create cooperation. It’s damage control, not partnership.

3️⃣ TikTok Structure Tacitly Endorsed

• US gets 6 of 7 board seats, algorithm control

• China provides unofficial approval of the structure

The Translation: THIS is what “tech cooperation” actually looks like—US operational control, Chinese IP ownership. This becomes the template for other Chinese tech entering US markets.

✅ Why Vagueness Matters

Three truths for digital transformation leaders:

1. No Framework Exists You’re navigating sector-by-sector uncertainty. Your e-commerce strategy needs different rules than your AI strategy.

2. Nvidia = Bellwether Watch Nvidia-China chip talks—it predicts everything else. Limited approval (60% probability) = selective cooperation. No approval (25%) = decoupling. Full approval (15%) = broad cooperation.

3. April 2026 Is the Real Test Trump visits China then. Green lights: joint ventures announced. Red flags: visit canceled, new export controls.

Western governments just secured a 1-year reprieve on rare earth exports, celebrating “breathing room” while betting on 2027-2030 supply MOUs.

Meanwhile, CATL, the global battery leader, is already in mass production of batteries that eliminate rare earth needs entirely (e.g., Sodium-ion). They are not waiting for stability; they are building optionality beyond the negotiating table.

The Trump-Xi deal is temporary. The 1-year rare earth extension will expire. Regulations will shift again. New tensions might emerge.

Here’s how to build strategy when stability isn’t coming back.

🤔Three Scenarios for 2026-2030

SCENARIO 1: HIGH UNCERTAINTY

What this scenario looks like:

• Regulations change frequently (every 6–12 months) and markets fragment with incompatible rules.

• Companies are forced to choose and commit to the US tech stack OR the China tech stack for critical systems.

• No comprehensive US-China tech cooperation framework ever emerges.

Your strategy in this scenario:

Immediate actions (2025-2026):

• Implement deep geographic diversification (no single market should represent more than 30% of critical operations).

• Build modular product architectures that can operate independently using different regional technology foundations.

Long-term positioning (2027-2030):

• For Tier 1 (critical) products, choose and commit to a single, secure ecosystem; use vertical integration where diversification is impossible.

Optimize for options, not efficiency: prioritize having three reliable suppliers at 80% efficiency over one highly efficient supplier.

SCENARIO 2: SECTOR-BY-SECTOR FRAGMENTATION

What this scenario looks like:

• Most likely scenario. Consumer tech stabilizes, but Defense, AI and Semiconductors remain highly volatile.

• Periodic “mini-deals” occur every 12–18 months (tactical, not comprehensive stability).

• Innovation accelerates in “neutral ground” regions like ASEAN, attracting both US and China investment.

Your strategy in this scenario:

Immediate actions (2025-2026):

• Implement the Tiered Tech Stack approach, segmenting products by risk profile (Tier 1 vs. Commercial).

• Establish ASEAN ‘swing capacity’ to serve both US- and China-aligned markets depending on regulatory shifts.

Long-term positioning (2027-2030):

• Maintain distinct sector-by-sector strategies (e.g., Defense follows Scenario 1, Commercial follows Tier 2).

Monitor the 2027–2028 inflection point (new US-ASEAN supply chains coming online) to decide on pivoting away from licensed Chinese IP.

SCENARIO 3: REGULATED COLLABORATION

What this scenario looks like:

• Sustainable frameworks for coexistence are established (e.g.,the TikTok board model expands).

• Joint ventures emerge in climate tech, AI for healthcare, and agricultural technology.

• Regulations are predictable with multi-year visibility(changing every 2-3 years).

Your strategy in this scenario:

Immediate actions (2025-2026):

• Deepen Chinese tech partnerships, moving beyond licensing to co-development arrangements.

• Stop over-investing in redundant supply chain capacity and redirect that capital to innovation.

Long-term positioning (2027-2030):

• Integrate supply chains deeply to optimize for efficiency over security (assuming stable material and technology flows).

Leverage Chinese scale and rapid iteration speed to reduce costs globally, but maintain 20-30% alternative supplier capacity as a critical risk hedge.

• CATL didn’t wait for last week’s Trump-Xi deal to know what to do. They were already building sodium-ion production lines regardless of what happened with rare earth access negotiations.

• ByteDance didn’t wait for TikTok regulatory approval. They had already built modular legal entities that work independently in different markets.

• BYD didn’t wait for clarity on EU tariff decisions. They had already built factories in three different regions to hedge three different scenarios.

The companies that win over the next 5 years won’t be those that successfully “picked the right side” of US-China relations.

They’ll be the companies that built adaptability into their core operating model.

• Poverty Reduction (1981-1990):Under the 5th and 6th Five-Year Plans, China slashed extreme poverty by 167 million people—from 542 million to 375 million—accounting for over 100% of net global reductions that decade, as poverty rose elsewhere, fueling the “China Shock” in trade and migration.

• Green Technology Dominance (2011-2015):The 12th Plan propelled China past Germany to become the top solar producer in 2015, with installed capacity hitting 43 GW. Today, it commands 95% of global polysilicon/wafer production and 90% of rare earth processing—critical for EVs, batteries, and tech—locking in green dominance.

• Technological Self-Reliance (2021-2025): The 14th Plan’s national security push targets self-sufficiency in AI, semiconductors, and biotech, sparking U.S.-China frictions. One fallout: China’s rare earth stranglehold means each F-35 jet needs 920 lbs of these materials, hobbling Western defense supply chains.

Hello there,

Everyone talks about China’s five-year plans. Some fear them as proof of an unstoppable economic juggernaut. Others dismiss them as Soviet-era relics with no relevance to modern markets.

Here’s what both camps get wrong:

They assume the question is whether China can plan the future. The real question is what these plans reveal about where Beijing will deploy capital, regulatory support, and state resources over the next five years—and whether your business is positioned to benefit or get squeezed out.

This week, Beijing is finalizing its 15th Five-Year Plan (2026-2030), with President Xi personally steering discussions. The stakes cover everything from semiconductor policy to foreign investment rules.

For corporate leaders navigating China strategy, the temptation is to either treat the plan as gospel or ignore it entirely. Both approaches leave money on the table.

Before diving into what’s next, let’s examine what actually happened during the 14th Five-Year Plan (2021-2025). The results reveal a predictable pattern that separates winning corporate strategies from wishful thinking.

🇨🇳 The 14th FYP Scorecard (2021-2025)

✅ WHAT’S OVERDELIVERED

1️⃣ Green Transition & Technology: The Supply-Side Dominance

• Renewable capacity doubled;

• Computing power 2nd globally;

• 4.55M 5G base stations;

• 60% of global AI patents

The Secret: The U.S. tech fight provided the mobilization fuel. Success came from building complete ecosystems (subsidies, protected markets) around existing industrial strengths. Speed of deployment beat perfection.

👉 We will explore specific tech breakthroughs in the next section.

2️⃣ GDP Growth: The Headline Win That Hides the Problem

• 5.5% annual growth (2021-2024);

• US $4.9 trillion economic increment

• Domestic demand contribution: 86.4% of GDP growth (up from 78%)

The Reality Beneath: Yes, China hit targets despite COVID, property crisis, and trade tensions. But growth was propped up by exports and infrastructure, not the desired consumer spending. The economy got bigger, but prosperity for average citizens barely budged.

3️⃣Employment: Creating Jobs While Eroding Confidence

• 59.2M urban jobs added by Aug 2025 (exceeded 55M target)

• 12M+ jobs/year despite global headwinds

• Urban unemployment: Stayed near 5.5% target

The reality beneath the numbers: State-owned enterprises, infrastructure projects, and the digital economy created millions of jobs. But quality matters. Youth unemployment hit record levels (over 20% before Beijing stopped publishing the data). Jobs created didn’t match the aspirations of graduates, fueling the “lying flat” movement.

4️⃣People’s Well-Being: Checkboxes Ticked, Problems Persist

• Pension coverage: 95% achieved

• Life expectancy: On track for +1 year

• Education: Average years increased as planned

The reality beneath the numbers: Coverage ≠ adequacy. Pension/healthcare costs rise faster than income; underlying anxiety over the future remains high.

❌ WHAT DIDN’T DELIVER

1️⃣Consumption: The US $4.9 Trillion Paradox 💸

• The Miss: Consumption share stalled at 56% (Target: 65%).

• Why? You can’t subsidize away insecurity. The property crisis wiped out household wealth (70% of family assets in real estate). Families hoard cash due to weak social safety nets(for medical/elder care).

• Vicious Cycle: Businesses face involution—falling prices and shrinking margins—so they can’t raise wages, meaning consumers can’t spend more.

2️⃣Demographics: Policy vs. Culture 👶📉

• Reality: Continued collapse to 9.54M births in 2024.

• Why the Miss? Cash can’t overcome structural life changes. Educated women prioritize careers; the 996 work cultureleaves no time for parenting. The cost of housing and education makes having two or three children financially terrifying. Cultural norms have shifted.

3️⃣ Disposable Income & Confidence Crater 📉🏘️

• The Miss: Disposable income growth slowed dramatically(Target: 6%).

• Why?Loss of property value made people feel poor even with a flat salary. The link between effort and reward severed for white-collar workers due to layoffs and stagnant wages.

4️⃣Foreign Investment: The Unintended Irony 🏭⬇️

The Setback: FDI dropped 33% (from US$ 182B to US$112B).

• Why the Miss? Beijing’s success created the problem.

• Domestic champions (EVs, solar) outcompete foreign firms (the “countdown phenomenon”). Strategic sectors tightened with data rules and security reviews.

• The Irony: Rhetoric of “high-level opening-up“ diverged from a reality that increasingly favors domestic players. Foreign firms stopped expanding.

The ultimate pattern is clear: Beijing won on supply-side mobilization and lost on demand-side engineering. State capitalism proved unparalleled at coordinating capital and resources to build infrastructure, industrial capacity, and technological systems. However, it is structurally ill-equipped to engineer human confidence, struggling to make citizens feel secure enough to increase consumption or reproduce.

The lesson for the future is to expect a doubling down on the successful, top-down approach—a further, aggressive technology push built on the successes of localization and industrial dominance.

🌍The Technology Push (2021-2025)

🚗Case Study #1: NEV

14th FYP Target: 20% of new vehicle sales by 2025

Results:

• Target hit by mid-2022 (2.5 years early)

• 45%+ penetration by August 2025

• 538% ownership increase (4.9M → 31M+ units)

Why it worked: Consumer demand existed (pollution, urban restrictions) + subsidies + infrastructure (60% of global charging network) + industrial policy (battery supply chain)

The mechanism: Not creating demand from scratch—amplifying existing momentum with state resources

💻Case Study #2: Computing Power

• 14th FYP Priority: Build digital infrastructure to support AI and data economy

Results:

• Computing power: 2nd globally

• 4.55M 5G base stations by June 2025

• 60% of global AI patents

• HarmonyOS: 900M+ devices across ecosystem

• AI token consumption: 100B daily (Jan 2024) → 30 trillion daily (June 2025) = 300x increase

Why it worked: U.S. restrictions created urgency + state funding + complete supply chain approach (not just R&D, but manufacturing, applications, infrastructure)

The mechanism: Made in China 2025 playbook—massive subsidies, protected domestic market, forced localization, scale advantages

🛜Case Study #3: 6G Technology

• 14th FYP Priority: Leverage 5G leadership to set global 6G standards and secure next-generation telecom dominance

Results:

• Early leadership in 6G research and standards-setting (positioning before commercial rollout)

• Building on 5G infrastructure advantage (4.55M base stations create testbed)

• China filing significant portion of early 6G patents (extending telecom IP dominance)

Why it Worked: The existing 5G head start provided a real-world testing ground. Crucially, state coordination allowed universities, SOEs (Huawei, ZTE), and research institutes to align on national standards early to avoid fragmentation.

The Strategy: Beijing mobilizes the “national team” model, pooling resources across government, SOEs, and private giants. The goal is to lead on standards, patents, and use cases from the beginning, making the Chinese 6G architecture the global default.

🍃Case Study #4: Green Energy & Nuclear Innovation

• 14th FYP Priority: Achieve carbon peak preparation while maintaining energy security; lead global clean energy manufacturing

Results:

• Renewable capacity doubled from 2020 levels

• Energy intensity: ↓11-12% (on track for 13.5% target)

• 1 in 3 kilowatt-hours now from green sources

• World’s first commercial SMR nears activation (expected to power 500K+ homes)

• China now produces majority of global solar panels, wind turbines, batteries

Why it worked: State banks financed projects, provincial governments fast-tracked permits, SOEs built infrastructure, subsidies made clean energy economically viable

The mechanism: When Beijing says “go,” capital flows, regulations bend, and local officials compete to overdeliver

The aggressive technological push above, from 6G dominance to ecosystem building, provides the immediate context for China’s next great policy roadmap.

While the final 15th Five-Year Plan document is not slated to drop until March 2026, the priorities guiding resource allocation and strategic direction are already clear from the intense plenum discussions held this October.

🤔What’s Coming in the 15th FYP (2026-2030)

1️⃣ Self-Sufficiency in Strategic Technology

The Central Focus: Reduce foreign dependency in key sectors (core response to geopolitical rivalry)

China’s AI Strategy: Deployment Over Breakthroughs

Unlike Western focus on frontier models, China pursues a “deployment-first” strategy:

• Embedding AI across logistics, manufacturing, surveillance, urban planning

• Building the “logistics of intelligence”—organizing data, compute, hardware, and applications into functioning systems

• Developing the “low-altitude economy” (drones, robotics)

The real competition: Not who invents AGI first, but who deploys “good enough” technology at massive scale most effectively.

💰Core Investment Areas:

• Chip Manufacturing and advanced design.

• AI Foundation Models and industrial applications of AI.

• Commercial Aerospace capabilities.

2️⃣: Domestic Consumption (With a Security Lens)

The Goal: Boost consumption in ways that strengthen domestic supply chains (”dual circulation” strategy)

What’s Coming:

• 💵 Subsidies for big-ticket goods (vehicles, appliances)

• 👴 Income boosts for rural and elderly populations

• 🏥 Services expansion (elderly care, healthcare, childcare)

The Business Angle: Consumer opportunities are real, particularly in services and big-ticket item trade-in programs, but the preference will lean toward goods that rely on domestic manufacturing and supply chains.

3️⃣: Green Transition Acceleration

The Stakes: 2026-2030 is crucial for meeting China’s 2030 carbon peak target

Strategic Shift:

From controlling total energy use → Controlling total carbon emissions

📋 New Regulations:

• Stricter carbon accounting & disclosure requirements

• Carbon labeling certification systems

• Product carbon footprint management

💰 Investment Surge:

• Renewable energy & energy storage

• Energy efficiency solutions

• CCUS (Carbon Capture, Utilization, Storage) tech

Action Item: Green tech receives strong state backing, but all companies must prepare for increased environmental monitoring and compliance costs.

🌾 Also Watch: Food Security

Food security remains a cornerstone, with focus on:

• AI-powered agricultural technology

• Diversified food sources

• Gene editing for higher yields

• Rural infrastructure investment

So, back to our original question: can China really predict the future?

Can China predict tech breakthroughs, economic cycles, consumer behavior?

❌ No. Nobody can.

Can China mobilize resources toward strategic goals better than most countries?

✅ Absolutely. That’s the core competency of the system.

Should you bet your business strategy on China’s plan being right?

❌ Never.

Should you understand China’s planning so you can adapt faster than competitors?

✅ Always.

🤔Where to Position Your Business

💰 Follow the Money

Here’s what sophisticated investors are watching:

🏥 Healthcare Specialization

With an extra 70 million people aged 60+ by 2030, the middle class is demanding specialized care (e.g., psychotherapy). Foreign expertise in specialized care and elderly services is highly valued (e.g., French nursing home operators already entering).

🌐 Services for Chinese Companies Going Global

As Chinese companies go global, their domestic legal and financial services can’t keep up. This creates a high-margin opportunity for experienced international firms in cross-border advisory, compliance, and legal services.

💡 “Good Enough” Tech for Global South

China is pioneering affordable, packaged solutions for emerging markets. This means investing in bundled offerings like renewable energy + AI grid management and building open-weight AI models adapted for local use.

✈️Nuclear Energy & Aerospace

China needs help closing high-reliability supply chain gaps. Despite its self-reliance push (like the C919 jet), essential partnerships (like those with Airbus) are still needed to acquire nuclear expertise and secure aerospace supply chains.

⚠️ The Risks Most Businesses Underestimate

⏱️The “Countdown” Phenomenon

A predictable pattern: Foreign companies get 5 to 7 years to profit by introducing expertise before domestic competitors—backed by the state—catch up, leverage scale, and ruthlessly overtake the market (e.g., chemicals, solar, EVs).

🤯The Involution Trap

Cutthroat competition with diminishing returns is a structural feature. Intense rivalry among over-leveraged players keeps margins razor-thin, leading to low capital efficiency and high youth unemployment despite overall GDP growth.

🚨 The Political Security Trap

The principle of “Political security above all” overrides market logic, causing:

• Information Bottlenecks: Local officials withhold bad news, crippling reliable market analysis.

• Regulatory Volatility: Abrupt crackdowns (on tech, gaming, etc.) based purely on political signaling, making long-term planning a high-risk gamble.

The future is being finalized this week, as the Central Committee reviews the proposals in Beijing.

While the full blueprint is months away, the strategic roadmap is imminent. The national mandate is set: Security, Technology Self-Reliance, and High-Quality Development.

Hello there! China’s open-source AI push proves the best competitive move isn’t protecting what you have—it’s making sure everyone needs what you’re offering.

This strategic shift is already redefining Silicon Valley. As Martin Casado of a16z recently revealed, 80% of U.S. AI startups now prototype with Chinese open-source models like DeepSeek. Why? They report 3x faster prototyping and simply can’t afford the proprietary alternatives.

As AI pioneer Kai-Fu Lee framed it: “If we look back in a decade at how DeepSeek helped China avoid falling behind the U.S. in AI, the answer won’t be its technology itself. The true contribution will be the open-source era for large models that it ushered in.”

That era is built on sheer scale. China’s contribution to global AI open-source has grown more than 10x in a decade, creating the world’s largest pool of active contributors, now over 2.2 million strong. This momentum is contagious: it’s fuelling India’s 17 million developers and accelerating massive AI market growth in places like Brazil. While the U.S. focused on proprietary stacks, the global balance of influence is rapidly shifting.

This newsletter examines why China doubled down on open source, whoelse is building alternatives, and what it means for you.

🇨🇳 Why China Chose the Open-Source Path

🎯 Turning Weakness Into Strength

• The Constraint: U.S. chip controls cut off China’s access to the advanced hardware needed for the expensive, all-out AGIrace. Couple with limited depth of its capital markets, China recognized they couldn’t compete dollar-for-dollar with the U.S. on compute, they strategically stopped trying.

• The Pivot: Software Efficiency First. Instead of chasing AGI,China focused intensely on practical applications (consumer products, robotics) and software efficiency. DeepSeek V3.2-Exp is the result, running at 50%+ lower API costs with 3x faster inference on domestic chips.

• The Result: This efficiency drive dramatically closed the performance difference. Proprietary models are now only 1.7% better than the most efficient open-source alternatives. China successfully turned a hardware weakness into a decisive software advantage.

• The Outcome: Global Economic Dominance: American startups are defaulting to Chinese models because the economics are irresistible. China is now winning the battle for consumer application dominance worldwide.

🌍 Free Beats Best in Emerging Markets

• The Playbook: Alibaba has open-sourced 300+ Qwen models, underpinning over 170K derivatives. When developers in Lagos,São Paulo, or Jakarta need AI, they pick what’s free and works.

• The Trap: Once they build on Chinese foundations, the ecosystem lock-in begins. Global forks of Qwen models contribute 2x more optimizations back to Alibaba than U.S.equivalents.

• The Calculation: China trades licensing fees today for platform dominance tomorrow. The code written today determines which platforms dominate tomorrow.

📊 U.S. vs. China: Different Games

• What U.S. Is Doing: Produced 40 notable AI models in 2024 vs.China’s 15. Emphasis on proprietary stacks for national security. Private investment leads with 26% growth in 2025.

• The Problem: China closed the performance gap while undercutting on cost—proprietary models now lag open-source by just 1.7% on benchmarks, per Stanford AI Index 2025. A 2025 cost management study found 85% of companies missed AI infrastructure forecasts by 10%+, often due to overestimating proprietary efficiencies and underpricing open-source scaling.

Would you rather own the best model, or own the ecosystem everyone builds on?

China chose the ecosystem. To truly understand the competitive implications of this strategy—the incredible efficiency, the cost advantages, and the speed of their development cycles—you need to see it up close.

🌍 The Global Build-Out

🇮🇳 India: The Global Talent Engine

India is rapidly emerging as the world’s new AI talent multiplier.

Massive Talent Pool: India boasts 17 million total developers andover 2 million active open-source contributors.

Future Leader: Projections suggest India is on track to surpass theU.S. in total developer numbers by 2028.

Young Workforce: This growth is fueled by a young tech talent pool,with the majority of the workforce under 35.

Localized Innovation: Indian developers are actively building localized AI tools for their 22 official languages through platforms like Bhashini, demonstrating practical, real-world adoption.

🌏 Regional Power Plays

UK → Top 3 globally for AI readiness. Emphasizing ethical AI andopen-source for business transformation.

Germany → Leads Europe in AI R&D. Industrial applications andopen protocols to prevent monopolies.

South Korea → Edge AI at 36.9% CAGR through 2030. Open-source accelerates semiconductor integration.

France → 1,000+ AI startups, $12B market size. Prioritizing customized open-source solutions.

UAE → #2 in AI computing. K2 Think open-source model (Sep 2025, 32B params) rivals OpenAI.

Canada → AI Strategy Task Force + US $5M Compute Fund (Sep2025). Sovereign R&D with open-source tools.

Singapore → US $1.2B public funding for 2025 AI strategy. 68% startup adoption; US $3.5B market by 2030.

Japan → New AI Strategy HQ (Sep 2025). OpenAI collab for public services; focus on open R&D and human-AI ethics.

Each region has different reasons: cost (emerging markets), sovereignty(Europe), integration flexibility (Asia). Same outcome: building outside theAmerican tech stack—non-U.S. open-source contributions hit 60% of global total in 2025.

🤔 What You Need to Know

💰The Agent Workforce is Here

Manny Medina, CEO of Paid, projects that “Half the workforce will be agents in five years.” AI isn’t just replacing tasks—it’s bundling entireoperations into autonomous agent workforces that run quietly in thebackground, charging based on value delivered rather than per-user seats. (Watch More)

This transition is already rewriting the economics of business:

• Cost Cuts: Open-source AI cuts costs by 50%+ in business units.

• Adoption Rate: 63% of companies now use open-source for cheaper, faster deployment.

• Productivity: Developer productivity is up 19% for open-source tasks due to better tools (early 2025).

• The Penalty: Companies without an open-source integration strategy face 3.5x higher software expenses.

💰 Business Model Transformation

The shift means AI agents become core business units, not just support functions. Companies treating AI as something to purchase will lose ground to those integrating, customizing, and evolving it. Open-source enables this constant evolution; proprietary APIs resist it.

🔮 Next 6 Months

Watch For:

• DeepSeek & Qwen: Major new model releases (V4, R2) from China’s most active open-source leaders.

• Security Focus: India’s open-source security initiatives via the Linux Foundation.

• New Revenue Stress: Multimodal and edge AI capabilitiesbeginning to strain proprietary API revenues.

• Quality Gap Shrinking: The “free” vs. “premium” distinction is fading fast as top open-source models (Falcon 180B, Mixtral8x7B) already handle business-critical tasks.

China isn’t giving away AI from generosity. They’re executing a decade-long strategy to control the foundations everyone builds on. While the U.S. guards premium models, China wins ecosystem dominance. The opportunity window is measured in quarters.

The question isn’t whether open-source AI matters. The question is: will you build with the new foundation, or watch your competition do it first?

Hi there! The global AI race is intensifying, driven by a dramatic tech rivalry between China and the US. The US chip ban, intended to slow Chinese development, has instead fueled a strategic counter-offensive. In a recent move, China has reportedly barred major tech firms like ByteDance and Alibaba from purchasing Nvidia’s AI chips, a decision that might signal Beijing’s confidence in its homegrown alternatives.

While Western markets cautiously debate AI’s future, a silent revolution is already underway in China, where artificial intelligence is being woven directly into the daily lives of 1.4 billion people. Consider this:

• For the People: Doubao, ByteDance’s AI app, now has 157 million monthly active users, accelerating mass adoption for everything from daily planning to personalized education. Meanwhile, Ant Group’s AI Healthcare app, AQ, serves over 140 million users, bringing telemedicine and diagnostics to underserved rural areas, proving AI is actively bridging the urban-rural divide.

• For businesses: Chinese SMEs are increasingly adopting AI, with over 90% now using AI tools and approximately 40% specifically employing generative AI.

In this issue, we will reveal what Alibaba, Tencent, and Baidu are betting big on, and how China’s consumer-first approach is reshaping the global tech landscape.

🎬 AI as Economic Medicine in a Slowing Market

Market Performance Tells the Story: Stock Surge

• Shanghai Composite: +17% YTD (through August 2025)

• Outperformed S&P 500 on AI innovation momentum

• Beijing’s $98 billion AI investment creates strong conditions for tech acceleration

I. AI as Economic Catalyst

Chinese companies pursue AI integration across entire business models while other markets focus on incremental improvements.

Generative AI Adoption:

Over half of Chinese merchants (52%) are already using generative AI for chatbots and logistics, showcasing a willingness to prioritize rapid deployment over gradual integration.

System-Wide Automation:

China is rapidly advancing its “AI Plus” strategy, aiming for a staggering 90% AI adoption across industries by 2030. Broader AI adoption metrics are also leading the world, with Chinese companies surpassing Western markets in areas like:

• AI-Powered Robotaxis 🚗 Baidu’s Apollo Go leads the global robotaxi market with over 14 million cumulative rides, a figure that tops Waymo’s reported 10 million paid rides. With fleets accumulating over 200 million autonomous kilometers, Baidu is scaling its safe, driverless mobility.

• AI Avatars for Sales 🤖 Virtual hosts are changing the face of e-commerce. A single livestream featuring an AI avatar of influencer Luo Yonghao generated US $7.7 million in sales in just seven hours. These digital salespeople are consistently outperforming human hosts, boosting conversions by over 30% and allowing companies to run pitches 24/7 without fatigue.

• AI driven logistics 📦JD’s autonomous delivery vehicles and drone networks has reduced delivery costs by 30% in pilot markets and enabled faster delivery times like same-day or next-day service in key international markets such as the U.S. and Japan.

II. 🤔 Real-World AI Deployments at Massive Scale

Alibaba’s Amap

AI-driven Amap now serves 40 million users with personalized location and consumption recommendations. These users actively engage with AI to make daily purchasing decisions. When AI can predict where you want to eat and streamline the ordering process, spending naturally increases. (Read More)

Alipay + Luckin Coffee

Alipay and Luckin Coffee have launched China’s first conversational AI payment system. By enabling customers to order and pay in a single, seamless conversation with an AI assistant, it eliminates the friction of being redirected to a separate payment page. This innovation, powered by Alipay’s new Model Context Protocol, is a key step forward in conversational commerce, setting a new global standard for a faster and more intuitive customer experience. (Read More)

Meituan’s Xiaomei AI Agent

Serves as an AI life assistant, it is a conversational interface that enables users to complete tasks like food ordering, restaurant booking, and personalized recommendations using natural language. The agent is part of a larger strategic shift to move beyond costly price wars. It’s not just improving efficiency in the US $80 billion delivery market; it’s expanding what consumers consider “orderable,” like complex group orders or time-sensitive deliveries. (Read More)

The Deeper Strategy

AI isn’t a separate product, it’s woven into every consumer touchpoint to reduce friction and increase spending velocity.

It’s tempting to see the AI race as a simple contest of who’s “winning,” but the truth is far more complex. The U.S. and China are running entirely different races, each with its own finish line.

East v.s. West: Two Radically Different AI Commerce Models

Open source 🇨🇳 v.s. AGI race 🇺🇸

China: The Open-Source Approach

China’s strategy prioritizes a collaborative, industry-wide effortusing open-source models to solve real-world problems. By allowing companies like DeepSeek and Alibaba Cloud to share their models, China facilitates rapid adoption and democratizes AI access.

💡This approach focuses on the immediate transformation of industries like healthcare, education, and gaming, creating a massive user base and a fast-moving ecosystem

U.S.: The AGI Race

The U.S. approach is a resource-intensive, winner-take-all race for artificial general intelligence (AGI). Firms like OpenAI and xAI are building massive data centers and consuming enormous amounts of energy to develop a single, super-powerful model.

💡This strategy is driven by the belief that the first company to achieve AGI will dominate the market, even if it comes at the cost of vast resource expenditure and a less-distributed impact.

Seamless ecosystem 🇨🇳 v.s. Fragmented Innovation 🇺🇸

China: The Seamless Ecosystem

China’s AI development is based on its existing reality of “super-apps” that have a billion-plus users. This provides companies with a seamless ecosystem and vast, unified datasets, leading to highly effective algorithms and hyper-personalization that are already a reality.

💡This integrated approach allows for systemic AI to be implemented across industries, from mobile payments to on-demand services, creating a cohesive user experience.

U.S.: Fragmented Innovation

In the U.S., AI development is characterized by fragmented innovation. Siloed applications like PayPal, Instagram, and Amazon limit the consolidation of data, reducing AI’s effectiveness. This creates a disconnect where AI is often retrofitted into existing systems rather than being a core part of the design from the start.

💡This approach can lead to slower integration and a less unified user experience across platforms.

China’s AI-first approach to consumer commerce is about reimagining how people shop, eat, travel, and live. With B2B digitization projected to exceed $35 trillion GMV by 2026, the implications extend far beyond China’s borders.

What This Means for Global Commerce 🌏

Ecosystems Over Applications

China’s unified, data-rich ecosystems prove that a network of seamless services is a more powerful competitive moat than a single, best-in-class application.

To compete, companies must break down internal data silos and prioritize unified, frictionless user journeys, integrating AI across their entire value chain—from supply to sale.

Application Trumps Theory

While the U.S. leads in the theoretical pursuit of AGI, China is winning the race for practical, profitable AI deployment.

The lesson is that immediate, large-scale application of even “less advanced” AI can generate a significant competitive advantage in sales, logistics, and efficiency right now. The real value is in putting AI to work, not just in building it.

Hi there, what if the world’s two largest innovation engines stopped competing and started collaborating?

While headlines focus on trade wars, a seismic shift happened at the SCO Summit in September 2025: Xi Jinping and Narendra Modi met for the first time on Chinese soil in seven years, with Xi calling their relationship a”Dragon-Elephant Tango”.

The numbers tell a better story: Bilateral trade hit US $128 billion in FY25, up from US $118 billion the previous year, making China India’s second-largest trading partner after the U.S. Direct flights suspended since 2020 are resuming. This warming appears tied to Trump’s punishing tariffs on both nations, forcing them to find common ground.

But here’s what most analysts are missing: When the world’s two largest populations decide to collaborate instead of compete, they’re not just reshaping trade routes — they’re creating the world’s most daring innovation experiment.

What This Co-existence Means for Innovation

From Trade War to Tech Partnership

India’s trade deficit with China hit a record US $99 billion in FY25, driven by electronics imports—but this dependency is becoming a strategic collaboration.

What’s Changing?

In Mumbai boardrooms, Indian entrepreneurs are making a calculated gamble: taking Chinese money while their governments publicly clash. Indian tycoons like Adani and Ambani are actively seeking Chinese tech transfers for EVs and renewables, betting that economic necessity trumps political posturing.

Policy changes now allow Chinese VCs up to 24% stakes in Indian startups (up from post-2020 restrictions).

What’s The Twist?

The challenge lies in navigating India’s strict FDI rules, which still require government approval for Chinese investments. This regulatory barrier is precisely what forces this partnership to be strategic, not reckless.

💡So What?

This is the foundation for the world’s largest innovation partnership, covering 2.8 billion people. But it’s also the world’s most fragile alliance, where one policy change could trigger a US $100 billion exodus.

The Global Unicorn Map

🇨🇳 China: Beijing (74), Shanghai (43), Shenzhen (32) = 149 total

🇮🇳 India: Bangalore (50), Delhi (39), Mumbai (17) = 106 total

Combined, these two nations control 255 startups in the global top tier—approaching San Francisco Bay’s 271 .

💡So What?

While Silicon Valley celebrates billion-dollar valuations, Asia is quietly building the infrastructure that makes those valuations possible. 73% of San Francisco’s “American” unicorns depend on Asian supply chains, developers, or manufacturing.

The Open Source Revolution

While the West built walled gardens, Asia is building open infrastructure.

🇨🇳China: A global leader in open-source contributes over 2.2 million developers.

🇮🇳India: The fastest-growing developer population, set to overtake the US by 2028 as the largest developer community on GitHub.

The Uncomfortable Truth: Your favorite American app probably runs on code written in Bangalore and Beijing.

When China’s state-backed depth meets India’s grassroots breadth, they’re not just collaborating—they’re quietly controlling the world’s digital infrastructure.

While China’s story is about a top-down, state-curated revolution, India’s is a completely different kind of epic. It’s a grassroots, bottom-up saga fueled by entrepreneurial hustle and the sheer audacity of building for a billion people.

But here’s the question that keeps Western executives awake: can two fundamentally different innovation philosophies merge without destroying each other?

Big Picture Comparison: Control vs. Chaos

🇨🇳China’s Top-Down, State-Driven Engine

💰Capital & Control: While private VC dropped to US $10.8B in H1 2025, the government deployed US $138B through state-backed funds - a strategic reallocation.

🧠The Hidden Advantage: China produces 11.8M graduates annually (40% in STEM) and retains over 90% of its top AI talent (The Global Artificial Intelligence Research Trends Report).

While India celebrates its diaspora, China builds a domestic depth that’s hard to poach. It’s also the second-largest AI talent hub, and Microsoft employing 29% of its global AI researchers in China.

⚙️Leading Industries: EVs (BYD), drones (DJI), AI (SenseTime) with a goal of 70% domestic chip content by late 2025.

The Result: while the West debates supply chain resilience, China already achieved it.

🇮🇳India’s Bottom-Up, Market-Led Ecosystem

💰Capital & Scale: The ecosystem is a global VC nexus, attracting US $13.7B in 2024. This global exposure allows for rapid scale, but also leads to volatility, with funding down 22% in 2025.

⚙️Leading Industries: India’s strengths are in software and services. It is a global hub for SaaS (Zoho), Fintech (PhonePe), and IT services, with rising strength in digital-first retail and solar manufacturing.

🧠Talent & Culture: India’s strength is its globally-integrated talent. It produces 2M STEM graduates and has 1.2M AI professionals, leveraging “brain gain” with 78% of scholars returning.

The work culture, while also characterized by long hours and hustle, is more flexible, with a focus on “Jugaad” (frugal innovation) and a growing push for a better work-life balance.

The Execution Paradox: When Different Models Collide

🇨🇳China’s Infrastructure Monolith

175 billion packages delivered in 2024 through top-down efficiency, with over 100 billion already handled in the first half of 2025 alone. Logistics Performance Index: 19th globally (2025).

The Model: control everything, optimize systematically.

🇮🇳India’s Gig-Powered Chaos

12M gig workers (growing to 23.5M by 2030) handling last-mile delivery through human-led solutions. LPI ranking: 38th but rising fast.

The Model: embrace chaos, adapt constantly.

The Collision Point: When these approaches meet in joint ventures, who wins? Early evidence suggests hybrid models outperform both pure approaches—but the cultural friction is intense.

From our experience, more Indian tech companies are eager to tap into China’s innovations in retail, livestreaming commerce, and micro-drama production.

Beyond joint ventures, they’re keen to learn and adapt from each other’s successes, highlighting the growing importance of collaboration in today’s fast-paced market.

Now here’s where it gets interesting. These two opposing models aren’t just learning to coexist—they’re actively learning from each other. And what they’re discovering could change everything.

What They Can Learn From Each Other?

China ➡️ India: The Self-Sufficiency Lesson

The Success Story: Cambricon’s 4,300% revenue growth by building Nvidia alternatives. The deeper lesson: when external pressure forces innovation, integrated ecosystems win.

What India’s Learning: Instead of celebrating software exports, Indian companies are building hardware capabilities. The risk: this could eliminate India’s cost advantage while playing to China’s strengths.

India ➡️ China: The Open Platform Revolution

The Success Story: UPI processing billions of transactions through open APIs. The deeper lesson: sometimes letting go of control creates more influence.

What China’s Discovering: Their “enter as competitor” global strategy hits walls. India’s “integrate with everything” approach opens doors. The opportunity: Chinese companies adopting Indian-style platform thinking could accelerate global expansion.

Both ➡️ World: The Hybrid Threat

Vision: China’s long-term, deep-tech focus

Execution: India’s agile, frugal approach Scale: Combined 2.8B population testbed

The Global Implication: When the world’s most strategic economy applies agility and the world’s most chaotic economy applies discipline, everyone else needs a new playbook.

Want to stay on top of the global race? Get in touch with our team at www.chozan.co

Hello friends!

The DeepSeek moment, world’s top high-speed rail system, EVs global lead, space program and more... You might already be amazed by China’s pace, but what’s even more important is how quickly some of its industries are turning into global billion-dollar powerhouses—offering replicable models for brands, media, and B2B players.

Take Pop Mart, for example. In just a few years, they’ve turned quirky collectable toys into a US$ 4.2 billion business. Now as the mini versions of Labubu will launch soon, PopMart founder predicts further growth – and you know this is just the beginning.

Or humanoid robots are making their way into the mainstream this summer, with China hosting the world’s first humanoid robot games and opening a robot shop where humanoids are sold for anywhere from US$ 278 to millions.

Meanwhile, in Xiong’an, China is constructing the “city of the future”, with cutting-edge projects in autonomous tech, AI, satellites, and urban planning.

In this newsletter, we’ll explore:

• The global rise of micro-dramas

• Other industries that are expanding globally from China (and fast!)

• How your business can participate or replicate their success.

First, let’s dive into the world of micro-dramas—the real game-changer in China’s entertainment industry.

• Market Growth: The micro short drama market in China is projected to reach US$ 7.2 billion in 2024, with expectationsto exceed US$ 9.5 billion in 2025 and US$ 14 billion by 2027.

• Investment Trends: Investment flow estimates have been adjusted downward due to increased competition, stricter regulations promoting quality, and a significant push for overseas expansion by companies seeking new revenue opportunities.

🎬 Everything You Should Know About Micro-Dramas

Booming Industry Growth: $7B and Climbing

In 2024, China’s micro-drama industry nearly hit US$ 7 billion, surpassing traditional cinema for the first time. By 2025, it’s projected to reach US$ 9.4 billion , fueled by 300 million MAUs and over 665 million global downloads in the first half of 2025.

Business Model & High Returns

Starting small, its production costs range from US$ 100K-$500K per series, generating revenue through in-app purchases (IAP), ads, and subscriptions. In H1 2025, IAP revenue across global apps like ReelShort and DramaBox surged to US$ 1.1 billion, marking a 5x year-over-year increase.

Global Appeal & Local Success

Micro-dramas have not only dominated China’s entertainment landscape but have also found global traction. In just Q1 2025, they reached over 370 million global downloads, six times last year’s figures. Their short, addictive format is ideal for platforms like TikTok, with U.S. users accounting for 49% of global IAP revenue in early 2025.

So, what’s driving this lightning-speed growth? Two key factors: effective marketing, and government support. Chinese firms are pushing micro-dramas concept and content on platforms like TikTok and Instagram through strategic ads, dominating top global apps downloads and creating joint venture opportunities with over 1,000 titles in distribution hubs.

Building on hits like Black Myth: Wukong and Labubu, which boosted China’s soft power, Beijing has launched a 2025 plan to support international expansion of micro-dramas, films, and animations. This includes easing content regulations, improving copyright protections, and encouraging the import of foreign works to expand China’s cultural influence.

Now, let’s explore how China’s micro-drama industry achieved its success and what key takeaways you can apply to your own business.

🤔 What Makes the Chinese Short Dramas Industry Stand Out?

Binge-Worthy Storytelling: Keep Them Coming Back for More

Beyond basic cliffhangers, micro-dramas master “emotional velocity”—delivering high-stakes twists, character arcs, and payoffs in just 1-2 minutes. This format leverages dopamine-driven anticipation, keeping viewers hooked. In early 2025, 576 million Chinese users (nearly 50% of internet users) watched these dramas, with 85% retention rates and top series averaging 70-80 million views per episode.

💡Takeaway

Micro-dramas capture attention with high-stakes twists, relatable themes (e.g., revenge, rags-to-riches), and dopamine-driven anticipation. Their bite-sized (1-2 minute) format suits mobile-first viewers, leading to 85% retention rates and millions of views.

AI & Tech Monetization: Revolutionizing Content Creation

AI is transforming micro-drama production—generating scripts, visuals, and predictive analytics for retention. In Q1 2025, AI-driven apps like Sereal+ and Kling generated US$ 20.8 million in revenue. Features like iQIYI’s AI iJump Talkout boosted engagement by 300%, setting the stage for profitability by 2027.

💡Takeaway

AI is transforming content production and personalization. By reducing costs and optimizing viewer engagement, AI-driven micro-dramas are scaling quickly.

Global Adaptation: The Power of Localization

Localization goes deeper than subtitles: AI analytics swap archetypes (e.g., CEO falls for maid in China becomes alpha billionaire rescues underdog in the U.S.) to boost relatability. In H1 2025, global IAP revenue surged to $1 billion, with 49% from the U.S. and 665 million downloads. Platforms like ReelShort tailored original content for specific markets, leading to 4x higher retention than generic versions.

💡Takeaway

Localization is a key growth driver for micro-dramas. By using AI analytics to adapt content for different regions, micro-dramas have achieved 4x higher retention globally.

Collaborative Growth: Strategic Partnerships & Alliances

Success hinges on ecosystem alliances that pool resources beyond simple co-productions—think shared IP libraries, talent databases, and cross-platform distribution to mitigate risks like content saturation.

💡Case Study

The Hengdian International Short Drama Alliance (HISDA), launched in 2024, has partnered with CCTV, Xinhua, and People’s Daily, offering over 1,000 titles and boosting 470 million global downloads. Initiatives like DreameShort’s script contests have reduced production times by 40%, sourcing fresh talent.

While micro-dramas exemplify China’s dominance in bite-sized digital entertainment, the country’s innovation engine extends far beyond content creation. Billion-dollar industries from robotics to gaming, low-altitude tech, and virtual humans are leveraging cultural exports, AI integration, and consumer trends to scale globally—often with unexpected cross-sector synergies, regulatory tailwinds, and hyper-local adaptations that drive outsized growth.

Here’s a deeper dive into each, enriched with 2025 data revealing non-obvious insights such as ecosystem collaborations, consumer behavior shifts, and future projections.

🔍 Other Chinese Billion-dollar Industries You Must Watch

🎮 Video Game Cultural Exports

Black Myth: Wukong has shattered records, selling 28 million unitsand generating US$ 1.3 billion in revenue by mid-2025. The game peaked at 3 million concurrent players on Steam, with 30% of players overseas. Its success has boosted Shanxi province tourismthreefold, and it sparked a Wukong-themed art exhibition in Hangzhou, drawing 100,000 visitors. China’s gaming exports grew 15% in H1 2025, driven by global interest in Chinese-themed IPs.

So what?

As seen with Black Myth: Wukong, these games create cross-industry opportunities, boosting tourism, art, and even social movements. Global brands should adapt storytelling, co-create experiences, and drive brand awareness through cultural content integration beyond just gaming.

🍜 F&B: Expanding Globally with Wellness Trends

Brands like Heytea and Mixue have expanded to 10,000+ storesworldwide by mid-2025. Heytea is valued at US$ 1 billion, and Mixuehas surpassed McDonald’s in store count, thanks to its efficient franchising model. The global F&B market is on track to hit US$ 683 billion by 2025, with health-focused trends like low-sugar and functional foods seeing 20% YoY growth.

So what?

The health-conscious shift in F&B is global consumer behavior. Mixue and Heytea have redefined franchising by blending local tastes with wellness trends—creating mass appeal. The opportunity for brands is to harness functional foods, low-sugar offerings, and localized marketing to dominate in emerging markets.

🚁 Low-Altitude Economy: Skyrocketing to Global Heights

China’s eVTOL/drones market has attracted US$ 10 billion in investments by mid-2025, with EHang receiving the world’s first commercial passenger license. The sector is projected to grow 8x by 2026, driven by regulatory support for drone deliveries and urban flying cars. Global demand for drone technology and flying cars is accelerating, particularly for logistics and mobility solutions in international cities.

So what?

Adapt fast to low-altitude tech in logistics or partnerships—if you’re not thinking of last-mile drone delivery, you’ll fall behind as the 8x growth by 2026 takes off.

🤖 Virtual Humans: A Global Surge in AI Influence

The virtual human market surged to US$ 6.6 billion in 2025, with 292% YoY growth. Chinese AI influencers like AYAYI and Luo Tianyi are driving global e-commerce, with live streams and virtual singers gaining traction worldwide. China is pioneering the use of XR and AR technologies, with global adoption of AI and virtual humans seeing explosive growth.

So what?

AI influencers drive millions in e-commerce and real-time brand interactions. Use AI to craft hyper-personalized experiences, virtual personalities, and create immersive brand interactions that resonate with a digital-first, experience-driven audience.Incorporate virtual humans into your strategy, and you’ll be leading a new wave of consumer connection.

China’s most explosive industries—from micro-dramas to robotics and retail—aren’t succeeding by chance.

Don’t wait to be disrupted.

🔍 Lessons from China’s Fast-Growing Industries

Chinese models thrive on agility: Low-cost entry (e.g., micro-drama budgets), data-driven iteration (A/B testing content), localization for global fit, and ecosystem partnerships (e.g., alliances for distribution). Government support accelerates via subsidies and regs, while cultural exports build soft power.

Key Takeaways

• ☑️Run high-return, low-cost pilots to test demand.

• ☑️Use AI for real-time personalization.

• ☑️Form joint ventures or alliances to enter new markets faster.

• ☑️Stay ahead by innovating faster than the disruptors coming your way.

Chinese Companies Are Coming.

Are you ready to defend or capitalize on this opportunity? Get in touch with my team at chozan(dot)co to chat about your project.

Hello there!

Jensen Huang’s recent viral video in China , where he praised local tech giants like Huawei and Xiaomi, underscores the growing global recognition of China’s role in the tech and AI race. “China is not behind … we are very very close”, Huang remarked, emphasizing the growing contribution of China’s tech companies for global development.

And he’s not alone. CEOs from Apple, Ford, and Google have also publiclyacknowledged China’s rising influence in shaping global innovation.

Despite rising geopolitical tensions, tech firms and MCNs are flocking to China to harness its cutting-edge tech infrastructure and build strategicpartnerships that offer a competitive advantage. The rationale is clear: AI, green tech, robotics, and other innovations are transforming industries, and to stay ahead, brands must engage with China’s evolving tech ecosystem.

In this issue, let’s dive deeper into these trends and explore how your business can innovate and stay ahead in this ever-changing landscape.

Get in touch with my team via info(at)chozan(.)co and visit www.chozan.co for more.

🤖So, What Global CEOs Think About China?

Nvidia’s deep collaboration with China’s tech ecosystem signals the country’s pivotal role in AI infrastructure. Despite political complexities, Nvidia recognizes China as a critical partner for building the next wave of AI infrastructure. (Watch More)

Despite rising tensions, China’s manufacturing expertise—especially in advanced tooling and precision engineering—remains crucial to Apple’soperations. It’s the skills and workforce in China that give it an edge. (Watch More)

Ford’s recognition of China’s leading-edge EV technologies and smart car innovations shows the company’s dependence on China’s advanced automotive landscape to stay competitive. (Watch More)

Google recognizes China’s rapid progress in AI, particularly in large language models and problem-solving capabilities. As China accelerates its AI innovation, it’s becoming a critical partner and competitor for global tech companies aiming to remain at the forefront of AI. (Watch More)

Despite the rising geopolitical tensions, MNCs are turning to China to harness its unparalleled tech innovation.

Why Now? China is home to cutting-edge innovations in AI, robotics, green tech,med tech, and digital infrastructure.

As global interest in China’s AI ecosystem continues to heat up, we just wrapped up China’s largest AI Expo, the 2025 World Artificial Intelligence Conference (WAIC) just wrapped up, this year’s conference brought forward some interesting developments.

Here are a few highlights for you.

🤖 Three Big Takeaways From WAIC 2025

🌍 Explosive Growth in Robotics

WAIC 2025 (China’s lharegets AI fair) saw a staggering rise from 18 robotic exhibitors last yearto 80 this year. Over 150 humanoid robots were showcased, with stand out innovations like Dobot Robotics’ AI-powered arm and Unitree Robotics’ R1 companion. This surge reflects a clear shift from concept to practical application, marked by breakthroughs in mobility and task precision. (Global Times, 2025)

💰 BAT’s Quiet AI Battle

Baidu, Alibaba, and Tencent showcased distinct AI strategies. Baidu focused on AI infrastructure, with 21 million developers using its deep learning platform. Alibaba led in user-friendly AI, powering 1 billion daily requests through AI tools. Tencent integrated AI models and robots, with their 3D model downloaded 230,000 times. Together, they are shaping a connected AI future. (NetEase Tech,2025)

🙋Smart Glass Market Surge

AI-powered smart glasses from brands like Rokid, XREAL, and Alibaba’s Quark Glasses were some of the most popular attractions at WAIC 2025. With features like real-time translation and photo-taking, these glasses are quickly evolving from a niche product to mainstream, driven by Chinese innovations like XREAL’s 57° field of view and Rokid’s multi-lingual capabilities. (36kr, 2025)

As we head into the second half of 2025, China and beyond are buzzing with tech events and expos you won’t want to miss.

Here’s a roundup of the biggest events that are shaping the future of tech, innovation, and sustainability in Asia.

🌏 The Must-Attend China Tech Events

World Robot Conference, Beijing, China

Date: August 9-12

Focus: Robotics and AI technology.

Why Attend: China’s largest robotics expo showcasing the future of AI-powered industries, with major brands and innovators leading the charge. (Read More)

Neaschina Shanghai, Shanghai, China

Date: August 13-15

Focus: New energy automotive tech, carbon neutrality, andautomotive chips.

Why Attend: Gain exclusive insights from top innovators and industry leaders on the rapidly advancing automotive technology sector. (Read More)

IFINEXPO HK—International Financial Ecological Expo Hong Kong Summit, Hong Kong, SAR China

Date: September 6

Focus: Financial technology trends and product innovation.

Why Attend: Join global financial elites as they share the latest trends, opportunities, and developments driving the future of fintech. (Read More)

Hong Kong International Biotechnology Convention 2025, Hong Kong, SAR China

Date: September 10-13 Focus: Biotechnology and life sciences.

Why Attend: Discover the latest breakthroughs in biotechnology, and engage with top experts to explore new business and investment opportunities. (Read More)

China International Industry Fair, Shanghai, China

Date: September 23-27

Focus: Smart manufacturing, sustainable supply chains, and cutting-edge equipment.

Why Attend: Explore innovations that are transforming global supply chains and smart manufacturing solutions, all under one roof in Shanghai. (Read More) IEEV China (International New Energy and Intelligent Connected Vehicles Exhibition), Beijing, China

Date: October 16-19

Focus: New energy and intelligent connected vehicles.

Why Attend: Explore cutting-edge innovations, connect with key industry players, and uncover new business opportunities in the rapidly evolving smart mobility ecosystem. (Read More)

CHINA HI - TECH FAIR, Shenzhen, China

Date: November 14-16

Focus: Market-oriented innovation, cutting-edge global tech trends, cross-industry integration.

Why Attend: Connect with global tech leaders, discover the latest innovations, and unlock new commercial opportunities in one of the world’s most dynamic technology marketplaces. (Read More)

Get in touch with my team via info(at)chozan(.)co to discuss your next China project, and visit www.chozan.co for more.