After watching WAIC, it feels like the AI space right now is a huge battlefield of rivals.

What’s good is that our private capital has been mobilized—without the country wasting much money—but the job got done.
Small and medium-sized enterprises are as numerous as strands of hair; although more than half of them die, the people still get tempered.
The best part is that we’ve turned AI into infrastructure for export.
Now things like electronic components and optical modules—foreign buyers have to line up to buy them.
Now, let me complain about what doesn’t sit right.
The robot application scenarios are too single-minded—everything is full of performers.
People go to the booth to watch—dancing, serving coffee—but none of this makes money!
A service robot sells for several hundred thousand, and to hire someone in central provinces you can pay $3,000 to $4,000 per month—then they can work for ten years.
Instead of thinking about replacing high-risk industries, you focus on chewing up bottom-level “NPCs”—did you narrow the road for yourself?
In places where robots are truly needed—high-altitude work, flood prevention and disaster relief—there’s not even a decent product to be seen.
At the end of the day, it’s because founders are too homogeneous; they only copy what others do.
In the next one or two years, the bubble is definitely going to burst.
But that might not be a bad thing.
The AI bubble is much bigger in the United States; what we lose over here are the high-stakes gamblers—rich people and institutions.
Once this round of hot money burns out, the real treasures will be the remaining depreciated equipment and the outstanding programmers.