This guide settles an account left open in another text. In the guide on solar, batteries and electric cars we wrote, in passing, that the lazy explanation for Chinese competitiveness is cheap labour and that it has been wrong for at least a decade. A sentence like that cannot be left without proof. Here is the proof, and it holds for far more than energy: it holds for the tool, for the home appliance, for the electronic product and for almost everything that crosses the Pacific inside a container.

The number almost nobody looks at

China's National Bureau of Statistics publishes the average wage by sector every year. In the May 2026 release, with figures referring to 2025, this was the picture of manufacturing:

Segment Average annual wage in 2025 Change on 2024
Manufacturing, urban non-private units 113,594 yuan +5.2%
Manufacturing, urban private units 76,055 yuan +6.4%
National average, all sectors, non-private 129,441 yuan +4.3%
National average, all sectors, private 71,590 yuan +3.0%

Three readings come out of that, and none of them fits the idea of a low-wage country. The first is the value: something close to 9,500 yuan a month in the larger formal segment and 6,300 in the private one, in an average that mixes the rich coast with the poor interior. The second is the direction: it went up. The third is the most awkward for whoever repeats the myth: manufacturing rose faster than the national average, in both segments. This is not a sector being squeezed, it is a sector competing for people.

Anyone comparing by hour worked gets to the same place along another road. The estimates circulating in the sourcing market put the Chinese factory hour at around six to seven dollars, against something close to three in Vietnam and four and a half to five in Mexico. Those figures are not official and vary a great deal by province and by sector, so they are no basis for a spreadsheet. They serve the only conclusion that matters here: on the isolated variable of hourly price, China lost, and lost a long time ago.

The line to take away: if the wage were the reason, the container would have got more expensive every year for the last fifteen years. It did not.

What should have happened, and did not

The forecast has been repeated since around 2010 and always with the same reasoning: when Chinese wages go up, production migrates to whoever charges less, and China becomes an expensive country the way Japan did. Part of that forecast came true, and it is worth acknowledging. Clothing, footwear, simple toys and low-value manual assembly did leave, and they went to Vietnam, to Bangladesh, to Indonesia and to Cambodia. Anyone working with those products watched it happen in their own catalogue.

The other part did not come true. China's share of world industrial production did not fall: it stands at around 28% to 30% of the planet's manufacturing value added, more than the next countries combined. And there is one indicator that says better than any other where things are heading: in 2024 China installed 54% of every industrial robot sold in the world, some 295,000 units of the global total, according to the annual survey by the International Federation of Robotics.

So both things happened at the same time. The wage rose and the country's position got stronger. When two variables move together in a direction a theory forbids, the theory is wrong. Cheap labour really was a Chinese advantage in the 1990s and 2000s, but it was the easiest advantage to copy, and for that very reason the least important of the ones the country built.

The right calculation is not wage per hour, it is cost per piece

Here is the reasoning error that produces bad buying decisions. Wage per hour is not cost of production. What matters is how much labour costs per finished piece, and that calculation has two halves: how much is paid for the hour and how many pieces come out of that hour. A worker who costs twice as much and produces three times as much is cheaper, not more expensive. It is the only way to explain how a German or Japanese factory manages to export industrial goods while competing with someone paying a tenth of the wage.

Now bring that to your own order. Open the FOB price of whatever you import, be it a drill, a blender, a light fitting or an electronic board. Labour is a fraction, and hardly ever the largest one. The bulk sits in raw material and components, in mould amortisation, in energy, in packaging and in the inland freight to the port. Halving the wage on an item where it is a small part of the cost barely moves the price. Having the component supplier two hours away by car instead of two weeks away by ship moves it a great deal.

That is what replaced the low wage. There are four of them, and the reason China keeps winning is that the four are only built together, and they take decades:

What replaced the cheap wage What it solves in your order Why it is not copied in three years
Supplier density A change of colour, connector or packaging without stopping the line It depends on hundreds of small neighbouring firms, not on an investment decision
Automation and capital Repeatability across units, and a cost that does not rise with the wage It needs capital, process engineering and someone doing maintenance in the same city
Scale and speed An amortised mould, a large batch and a short cycle between sample and production It needs constant volume, which only comes from a huge domestic market
Infrastructure and engineers Ports, power and technical people in quantity, all of it near the factory It is public works and the training of people, measured in decades

1. Density: the neighbour who solves it in two hours

This is the most underrated of the four, and the one that shows up most in the daily life of a buyer. The value of a Chinese industrial hub is not in the factory that signs the order, it is in the hundreds of firms within a few kilometres that make the mould, the injection moulding, the stamping, the electroplating, the screen printing, the board, the cable, the foam and the cardboard box.

The practical effect is banal and decisive. You ask on Tuesday for a colour variation and a different fitting, and by Thursday a sample exists. Not because someone worked through the night, but because the tool shop is twenty minutes away and has done that three hundred times. It is the same mechanism that makes a product variation in 45 days possible, and it is what the map of the industrial clusters draws city by city.

The counterexample is even more eloquent. Assemble the same product in a country without that surrounding web and watch what happens to the shopping list: the component still comes from China, now by air whenever the deadline tightens. Leaving China ends up importing from China with a stopover in the middle, and the importer pays the freight twice. Anyone who has tried to move an electronic product out of the country without moving the chain with it knows that arithmetic.

2. Automation: two million robots, and an average that misleads

The International Federation of Robotics figures for 2024 say two things that look contradictory and are not. The first is size: around 2 million industrial robots in operation in China, the largest stock in the world, roughly 4.5 times that of Japan, which is second, and more than half of everything the world installed that year.

The second is density, which measures robots per 10,000 manufacturing workers. On that cut China shows up with 166 units and 22nd place in the world, well behind South Korea, with 1,220, Singapore, with 818, Germany, with 449, Japan, with 446, and the United States, with 307. The world average is 132. It is worth recording that this position fell relative to earlier reports for a statistical reason and not an industrial one: China's own statistics office revised its manufacturing workforce upwards, and the denominator of the calculation grew.

The correct reading of the two together is this: China is not the most automated country per worker, it is the country that automates the most, every year, and by a wide margin. And it is also where automation costs least, because it is there that much of the robot, the servomotor and the welding cell is made.

What that means at the negotiating table, and hardly anyone uses it: an average of 166 across an industrial base of that size means there are factories at both extremes, and plenty of them. Two quotations for the same product, at the same price and with the same specification sheet, may come from an automated line and from a shed with a bench and a screwdriver. The price does not give the difference away. Repeatability across units does, and it only shows up later, in the whole batch, when it is already too late to change supplier.

Anyone who wants to see this same industry from the finished-product side, rather than the process side, will find the portrait in the guide on the robot that already works: what made service robotics cheap is exactly what made factory automation cheap.

3. The scale and the speed

A mould is a fixed cost. It costs the same for a thousand pieces and for a million, and what changes is how many pieces it is divided by. A factory turning over constant volume amortises its tooling on a timescale a small-batch factory never reaches, and that is why it can offer a price that looks impossible to anyone looking only at the bill of materials.

Speed is the second effect, and it is the one the importer feels first. A line running the same product family without stopping does not have to be learned again with every order: the setup is known, the jig exists, the operator already knows where the product tends to fail. What you buy with the same money is not a cheaper hour, it is a shorter cycle between the idea and the container.

There is a price for this, and it comes up in the conversation about minimum order. The same scale that pushes unit cost down is what makes a good factory turn down a small order, and the minimum it asks for is rarely its own: it is the sum of the minimums of everyone standing behind it in the chain.

4. What is not the factory: port, power and engineer

The fourth piece is the one that comes up least in sourcing conversations and weighs most on the lead time. A container leaves Shenzhen, Ningbo or Shanghai for anywhere in the world every week, with a frequency and a choice of shipping line that almost no other country in the region offers. Industrial power is abundant and reaches where the factory is. And there are process engineers in quantity, trained in the thousands every year, and it is that professional who gets a new line running steadily in weeks rather than months.

None of that is a company decision, it is a country decision taken thirty years ago and paid for with public works. It is also why the conversation about replacing China rarely includes the hard part: it is relatively simple to open a shed where wages are lower, and very difficult to put a port, a power grid and a technical university next to it.

Where labour still decides, and it is fair to say so

It would be dishonest to close the argument without the other half. There is a whole family of products in which the wage remains the main variable, and it shares one trait: the more of the final price is human hands, the more the country of origin matters.

It is exactly that family that migrated, and keeps migrating. The other side of the rule holds too, and it is the part that matters to anyone importing industrial goods: when the price is dominated by components, mould and machine, the wage saving is small and what is lost on the other side is large. Where those who leave China go, and what changes in cost and risk in the Chinese interior and among its neighbours, belongs to a guide of its own, because the answer is different from this one.

What this changes in your order

Now the practical part. If labour is not the lever, some things done out of habit in a negotiation start doing harm instead of good.

What still cannot be settled from a distance

Notice what comes out of all this. If price has stopped separating factories, what separates them is precisely what is not in the quotation: which side of the automation average the line sits on, whether the plant that made the sample is the plant that will produce the order, whether the company signing is a manufacturer or a trading company inside the hub, and whether the surrounding suppliers really exist or whether the electroplating will be subcontracted 900 kilometres away.

None of those four questions is answered by a PDF, by a production-line video or by a catalogue, and all four decide whether the second order will come out like the first. They are answered standing up, inside the factory, looking at who operates and what is on the floor. That is what our team does at origin, and it is the reason the trip exists. For anyone starting there, the guide on Shenzhen from the inside teaches how to separate factory, trading company, agent and sales office, and the one on Yiwu beyond the market shows who is really behind the counter of a wholesale market.

Frequently asked questions

Is China still the cheapest place in the world to manufacture?

By the hour worked, no, and it stopped being that a long time ago. Vietnam, India, Bangladesh and much of Southeast Asia pay far less than a Chinese coastal factory. By the cost of the finished piece, delivered on the agreed date and with the same quality repeated across every unit, China usually still wins in most categories, and the reason has nothing to do with wages. It has to do with how many suppliers sit two hours away, how much of the line is already automated, the size of the batch the factory turns over each month, and how easy it is to put all that into a container.

How much does a Chinese factory worker earn today?

According to the official figures from China's National Bureau of Statistics, released in May 2026 with reference to 2025, the average annual wage in manufacturing was 113,594 yuan in urban non-private units and 76,055 yuan in private units, which comes to roughly 9,500 and 6,300 yuan a month. Both rose, by 5.2% and 6.4%, and both rose faster than the national average across all sectors, which was 4.3%. It is an average for a continental country: an industrial city on the coast pays well above that and an inland province pays below.

If wages rose every year, why did the price of the Chinese product not rise with them?

Because labour is a minority of the cost of most of what Brazil imports from China. The FOB price of a home appliance, a tool or an electronic product is dominated by raw material, components, mould amortisation, energy and logistics. Halving the wage on an item where labour is a fraction of the cost barely moves the final price, while gains in productivity, automation and scale move it a lot. And when a Chinese product really does get more expensive, it is almost never the wage: it is a commodity, the exchange rate or a tax decision, such as the 9% export VAT rebate that ended on 1 April 2026 and made solar panels more expensive without any supplier having raised its price.

Is it worth moving the order to Vietnam because of labour cost?

It depends on how much of the final price is human hands. In sewing, manual assembly and hand finishing, the wage gap shows up directly in the cost of the piece, and that migration is real and has already happened in clothing, footwear and simple toys. In a product dominated by components, mould and machine, the wage saving is small and tends to be swallowed by what is lost on the other side: the component still comes from China by air or by sea, the lead time grows, and the surrounding suppliers that used to solve a problem in two days are no longer there. The right question is not which country pays less, it is which part of your cost is made by hand.

Is China the most automated country in the world?

No, and the confusion here is common. By density, which measures robots per 10,000 manufacturing workers, China shows up with 166 units and sits 22nd in the world, behind South Korea, which has 1,220, Singapore, Germany, Japan and the United States. What it does have is the largest absolute stock on the planet, around 2 million robots in operation, roughly 4.5 times that of Japan, which is second, and in 2024 it installed 54% of every industrial robot sold in the world. In other words: it is not the most automated per worker, it is the one that automates the most, every year. For a buyer, the useful reading is that this average hides factories at both extremes, and two identical quotations may come from opposite sides of it.

How does this change the way you negotiate price with a Chinese factory?

It changes the question. Asking for a generic discount in a category where labour is not the lever pushes the cut into the only place still available, which is the material: a thinner cable, a cell from another brand, lighter sheet metal, plastic with more mineral filler. The price falls and the product changes without anyone giving notice. The productive negotiation is to ask in writing what changes in the specification at each price level, to compare samples instead of comparing tables, and to work out which side of the automation average the factory that will actually produce your order sits on.

Cost per piece Industrial hub Robots per worker Mould amortisation

BCVN — Brasil China Viagens e Negócios has connected Brazilian companies to the best suppliers in China and Asia since 2008. Find out about our services in commercial intermediation, quality inspection, factory audits and guided business trips to China's main trade fairs.

Keep reading

The map of China's industrial clusters Read article → Solar, batteries and electric cars Read article → Inside Shenzhen Read article → From idea to container in 45 days Read article →