a) Delivery
Since 2025, Alibaba used massive subsidies to compete with Meituan.
b) Micro-loan facilitation
In 2025, the industry faced regulatory crackdown on high fees or high APR, while Alibaba’s Ant Group operates at lower APR segment.
c) OTA
Trip.com (previously CTrip), the leader in China OTA, is targeted by regulators recently for antitrust issues.
d) PDD
In e-commerce, the industry faced some scrutiny and PDD is being probed.
Please see previous post on Old economy is new and new economy is old.
Here is another reason why this happened.
In the era of AI, robots and space, the TAM is beyond human beings.
The electricity is not only consumed by human, but will also by AI agents and robots.
Robots and AI agents will be the “new population” in physical and virtual forms.
Robots and AI agents will have their own identity, wallet, etc.; robots will need additional space to house.
If we will become multi-planet, we will be building a lot more, which will require a lot of manufacturing capacity, materials, power etc.
Thus, old can become new.
Meanwhile, the human being population may not increase that much.
Those human consumption categories will not increase with the “new population”.
Consumer internet probably won’t have dramatically increased TAM, as the ultimate demand comes from human beings.
Thus, new can become old.
Some will keep being “old”.
E.g. Robots and AI agents will not consume alcohol/baijiu.
Part of the previous “old economy”, such as industrials (e.g. heavy manufacturing), material (e.g. mining), utilities (e.g. electricity) are becoming the new “new economy” due to AI and related investments.
Indeed, AI isn’t just an app-layer story; it’s a capex + physics story.
They’re on the critical path of AI expansion, and they’re supply-constrained enough to earn economic rents.
Meanwhile, part of the previous “new economy” is becoming old, like software, e-commerce etc.
Software is less scarce, with AI-assisted coding and commoditized building blocks
E-commerce matured in many markets. China is the prime example. China’s gov reported e-commerce (physical goods) growth has been below 10% for 4 consecutive years since 2022.
To build a brand,
1/ need to find who you are first – define the mission/spirit or whatever value you want to express, then build the product, not the other way around
2/ need to do things that convey the value, not just talk about the value
Recently, 犒赏经济 has become a hot topic in China. The related articles try to show resilience in consumption and suggest a way to lift consumption.
While I agree with the necessity of this concept, as consumption in China needs to upgrade to “quality consumption” as some may say, I think 犒赏经济 is also trying to avoid some other key issues.
1/ key examples of 犒赏经济 are also lipstick effect.
Usually these articles argue that the rise in blind box toy sale like Popmart is a form of 犒赏经济.
However, if you think about it, Popmart toy is also like high-end brand lipsticks – people are replacing large item luxury purchases (handbags etc.) with smaller items ($20).
The desire to buy luxury products still exist during a bad economy, but people choose to buy stuff that have less impact on their financials.
One common use case of Popmart toy is to attach it to luxury handbags. Adding the “attachment” makes people feel that the handbag is “new” , thus somehow replacing the need to buy a new one.
Other examples of 犒赏经济 can also be lipstick effect.
Buying a nice dessert on the way back home? That’s a replacement for a much more expensive dinner out.
2/ 犒赏经济 tolerates other negative effects on overall consumption like stress or off-times.
Some part of the 犒赏经济 is not to celebrate in my opinion.
The mental stress is usually mentioned as a cause of rise in 犒赏经济, but is that a good thing? Are economists going to argue that in order to drive 犒赏经济, more people need to feel the stress?
Plus, these articles avoided discussions of long working hours and short holidays.
Long working hours is limiting dinner consumption and other 夜间经济.
In most companies in China, young people only get 5 days of annual leave per years. In additional, many companies will ask why you take a leave, and there is no such thing as getting paid for unused leaves. I bet many European people would say that like hundreds of years ago.
In 1936, France introduced law for 2 weeks of paid leave for all workers. This is on top of 9 days of national holiday at that time. The 2 weeks was further raised to 3 weeks in 1956, to 4 weeks in 1969, and to 5 weeks in 1982.
Wonder why concert is more popular than traveling? Because concert is usually in the city or a weekend trip that doesn’t involve taking a leave.
Let me just stop here.
Overall there are huge potentials in consumption in China I believe, and the quality consumption is the way to go. But some limiting factors need to be addressed first.
Midea has over 400mn billion rmb revenue in 2024.
Xiaomi is climbing over 400 billion rmb revenue in 2025.
Midea’s typical gross margin is ~25%+, Xiaomi is also over 20%.
Midea net income margin is higher at ~9-10%, Xiaomi is now ~8-9%.
Both have over CNY 40bn net income in 2025.
Both have diversified business lines. And Xiaomi is entering the core home appliance segment globally.
Both have potential in future robotics business. [Xiaomi’s robot] [Midea’s robot]
Due to Xiaomi cars, Xiaomi is revenue growth is fast, projected to be ~30% in 2025. Midea revenue growth rate in 2025 is ~10%.
Xiao may compound faster, as
1/ Midea has ~70% dividend payout, while Xiaomi doesn’t pay any.
2/ Xiaomi has higher growth momentum, driven by cars.
3/ Xiaomi employees are more software and AI-focused.
Meanwhile, Gree has a smaller revenue base of ~below 200 bilion rmb, less than half of Xiaomi or Midea.
Gree does enjoy a higher net income margin of high-teens in 2024, but that may turn out to be a sign of underinvestment in talents and future development.
What you are seeing in 3q25 capex spending…
What you are hearing from mgmt.
MSFT – “now expects capital investment growth in fiscal 2026 to exceed that in fiscal 2025”.
Alphabet – “expect significant increase in 2026 capex” / 2025 capex to be in a range of $91 billion to $93 billion” vs $85 bn guided in 2q25.
Meta – “capital expenditures would be notably larger in 2026 than 2025”. / 2025 capex to be $70-$72 billion vs $66-72 billion guided in 2q25.
Amazon – “we expect our full-year cash CapEx to be ~$125 billion in 2025, and we expect that amount to increase in 2026” vs ~$118bn guided in 2q25
My rule of thumb is that technology is deflationary.
Via technology improvement, people get more for the same price.
This is evident in chip technology, where the same amount of money buys you newer chips with better performance each year. It’s more obvious when you notice that the “best” chips of previous years will always cost less if you wait.
It’s not always the case in other industries.
The same can of coke will cost you more over the years!
Pharmaceutical and biotech can be technology.
One example is the genome sequencing cost, which has dropped significantly.
Over the years, a drug normally will see its price drop significantly, thus patients shall pay much less for the exact same product.
However, during the period of a few years, drug price may increase, before the price erosion. This period is coke-like.
Last year, Huarong, now China CITIC Financial Asset Management (2799.HK), bought into Bank of China H-share, becoming a over 3% shareholder. And it took one non-executive board seat of Bank of China right after.
Great Wall AMC, became a over 3% sharehoder of Minsheng Bank and got 1 non-executive board seat.
Xinda, via purchasing convertible bonds, became a over 3% sharehoder of Shanghai Pudong Development Bank.
These shouldn’t be taken lightly, as banks are heavily regulated in China. So there were political decisions made.
For the same destination, Baidu’s Apollo robotaxi in Shenzhen will charge RMB 125 (before coupon) vs RMB 40 on Didi express (affordable tier, before coupon) and RMB 50 for regular taxi.
What’s also interesting is that Baidu’s robotaxi estimates that it will take 79 minutes!
Meanwhile Didi estimates it’s about 31 minutes, which is in-line with other map apps’ estimates.
Baidu robotaxi charges more than 2x the taxi price and takes more than 2x the time…
Well done.
Attaching the breakdown of Baidu robotaxi fare (before coupon)
