What the 2025 Smart China Expo Tells Us About Where Manufacturing Is Heading

Every year, after walking the expo halls, the strongest impression is not that "technology has advanced again" — it is that the boundary of manufacturing has shifted outward by another notch. At the 2025 edition, two booths never emptied: humanoid robots and flying cars.

If you read these two only as showpieces, you miss the real signal. They point to two underlying capabilities of intelligent manufacturing: flexible assembly capability, and the ability to restructure logistics radius.

1. Humanoid robots: from walking to working

In earlier years humanoid robots mainly solved "walking steadily"; this year the focus moved to "working accurately". Several booths demonstrated grasping, insertion and screw driving in unstructured environments, with noticeably better tolerance than the previous generation.

The significance for manufacturing is not "should we buy a humanoid robot", but the technology path it validates: once vision, force control and decision models mature, robots no longer need to be bolted to a fixture — they can enter operations that previously required human hands.

Precision assembly has always been one of the hardest operations to automate. Parts have tolerances, incoming material varies, and assembly has a feel to it — all of which demand real-time judgement. The reason humanoid robots keep being demonstrated on assembly tasks is precisely that they compress sense, decide and act into a single closed loop.

For a factory, the practical landing point is not whole-robot replacement, but these capabilities being pushed down into dedicated equipment. Over the next three to five years you will see more assembly cells "with eyes and a sense of touch", rather than a fleet of humanoid robots walking the shop floor.

2. Flying cars: redefining the logistics radius

What is displayed is the vehicle, but the real industrial value sits on the logistics side. Once a ground journey measured in hours becomes a corridor measured in minutes, the logic of plant siting, inventory and even supplier layout shifts with it.

A simple piece of reasoning: if critical components can reach you across cities within two hours, then "keeping safety stock in a single warehouse" is no longer necessary. The coupling between regional distribution centres and the assembly plant loosens by one notch, and a company can hold the same delivery performance on less inventory.

That is why the low-altitude economy keeps coming up. It is not one more vehicle — it is one more transport network, which for manufacturers means one more lever to compress inventory and shorten lead time.

3. Where does the 40% efficiency gain come from

In industry estimates of rising intelligent manufacturing penetration, the figure "efficiency up more than 40%" appears often. It is not invented out of thin air; it breaks down into three parts.

The first is equipment utilisation. Once data is connected, changeover time, breakdown downtime and waiting for material all become visible and compressible. OEE typically gains several to a dozen points.

The second is output per head. Non-value-adding work — repeated handling, paper records, manual reconciliation — is taken over by systems, so the same people can run more equipment.

The third is decision speed. Scheduling used to rely on experience and exceptions on meetings; now it can be resequenced against real-time status. This part is hardest to quantify, and usually where the biggest gap lies.

One caveat: 40% is a ceiling-level description at industry level, not an average every company will collect. For a plant with weak basic management, gaining 10% in the first year is already a good result.

4. What exactly is "manufacturing as a service"

Another phrase heard repeatedly at the expo is "manufacturing as a service". It means the customer is not buying a machine or a batch of parts, but capacity — the ability to process.

For a plant with flexible lines, this means idle capacity can be rented out on demand; for a brand owner, it means expanding product categories without building a factory. The precondition is that production data must be transparent and traceable — otherwise nobody dares hand orders to capacity they cannot see.

So manufacturing as a service looks like a business model innovation, but underneath it is spillover from digital capability. Without a data foundation, the model does not run.

5. What China, the US and Europe are competing on

The focus of competition has shifted over the years. Early on it was single-machine precision; now it is three things: how deeply AI models are fused with industrial know-how, how well industrial software coordinates with equipment, and who holds the say over standards and ecosystems.

The first decides whether you can work smartly, the second whether you can work reliably, the third whether you can replicate at scale. Among the three, Chinese companies hold advantages in equipment and scenario data, and still have ground to make up in the core of industrial software.

For a specific manufacturer, the meaning of these grand narratives is this: the technology vendors, software ecosystems and standards you depend on will keep shifting over the coming years, so selection and architecture should leave room to be replaced.

6. What SMEs should actually watch

If you run a plant of a few hundred people, the thing to do after the expo is not to approve a robot purchase, but three small things.

First, identify the three operations on your line that depend most on a veteran's feel. Those three are the main battlefield for automation and intelligence over the next five years, and the yardstick by which you judge whether the technology is mature.

Second, put equipment data connectivity on the agenda. You do not need to do it all at once — start with uptime, downtime reasons and output counts on key machines, and run the data chain through once.

Third, recalculate your inventory and delivery cycle. If external logistics costs keep falling, your safety stock policy may need adjusting earlier than you think.

An expo shows you the trend; a factory competes on rhythm. Get the trend right but the rhythm wrong and you are still out of the game; hold the rhythm and being half a step behind on technology is not fatal.