Original author: Zane Hengsperger (@zanehengsperger)

Translation: ShenChao TechFlow

Introduction

This is a post that has gone viral on foreign websites.

The author is Zane Hengsperger, an entrepreneur related to American industry. He proposed a six-step roadmap titled "How America Can Beat Shenzhen," addressing the gap between China and the U.S. manufacturing sectors, from reversing the cultural disdain chain, building large factories, turning energy into a weapon, to using AI software to fill capacity gaps...

The core judgment of the Americans is that Shenzhen's success comes not from cheap labor, but from density, speed, and social status, which they believe America can replicate. The article carries a strong flavor of industrial nationalism, yet the data is solid and logical, providing valuable insights for understanding the underlying logic of America's current "reindustrialization" narrative.

Another industrial revolution has arrived.

Shenzhen was a fishing village in 1980. Today, it exports more electronic products than the entire United States. There, you can complete the entire process from CAD files to injection-molded prototypes within 48 hours. In the U.S., this takes six weeks and involves coordinating with four suppliers.

This is not because Chinese engineers are smarter, nor because labor is cheaper. Shenzhen wins with density, speed, and a culture that regards "making tangible products" as the highest status job.

All three factors are replicable. The United States has not replicated anything in the past thirty years, opting instead for outsourcing. In 2000, the U.S. accounted for 25% of global manufacturing value added, while China accounted for 6%. By 2023, China is at 29%, and the U.S. is at 17%. This is a choice. We can make different choices.

Below is a roadmap to follow in order. The order is important.

1. First, fix the culture

I genuinely believe that culture is everything. Culture is business, culture is family, culture is community, culture is industry, culture is nation.

The culture in America is not great, but it can be fixed.

This is the bottleneck beneath all bottlenecks. You can pour concrete and buy machines, but you cannot buy a generation willing to work in factories.

Only 6% of American high school students view manufacturing as a career direction. For every five skilled workers who retire, only two newcomers enter the field. The average age of welders is 55 years. Meanwhile, in Shenzhen, a 26-year-old running a CNC workshop is the person everyone in his friend circle wants to become.

For decades, we have told every smart kid the same thing: leave the workshop, go to the office. Learn to code. Manufacturing is something for your grandfather's generation. The result is 570,000 unfilled manufacturing jobs, and by 2030, there will be one trillion dollars' worth of output sitting on the table without anyone to take it.

The solution lies in status, not subsidies. Let machinists be the stars again. Factory workers should be cool.

What does it look like in practice? Technicians walk around the workshop shooting videos and posting online. Defense technology has made hardware cool for the first time since the Apollo era. Trade schools pay students to participate in real project training, securing jobs before graduation. A 22-year-old sees that a licensed welder earns more than a market analyst. Workshops in high schools instead of just distributing career brochures. To be honest, Instagram is good for this. I see too many cool young people working on machines; this is becoming cool again.

Culture moves before capital. No one will build factories in a country where no one wants to work in them. Once this is resolved, all the other numbers will become easier. Skip this, and the rest does not matter.

2. Build 100 times more factories

What we should pursue is "factory abundance."

Shenzhen's superpower is density. Within a 50-mile radius, there are thousands of factories. Your PCB factory, injection molding factory, CNC machining shop, and assembly line are all neighbors. Iteration cycles are measured in hours.

America is exactly the opposite. Our industrial base is dispersed, aging, and shrinking. Six primary aluminum smelters, four of which are partially or completely shut down. Annual steel production is 79 million tons, while China produces 1.005 billion tons. The delivery cycle for base metals ranges from 8 to 30 weeks.

You cannot counter density with a few super factories and a press release. The way to combat density is through scale. Hundreds of new rolling mills, foundries, machining shops, sheet metal shops, and assembly factories need to be consciously concentrated.

Cluster them in places where the infrastructure already exists: Detroit, Houston, Phoenix, Carolina. Land is cheap, freight rail is ready, and workers have muscle memory. Bring the supply chain together so that parts travel no more than a day from one step to the next. This is the truth of Shenzhen: not one giant factory, but ten thousand small shops operating like an organism.

And design energy into the factory from the very beginning. Aluminum and steel are energy monsters. On-site power generation and next-generation nuclear power have finally reached the implementation stage. A factory that generates its own power cannot be undercut by China forever.

3. Turn energy into a weapon

U.S. electricity demand has been flat for twenty years, growing an average of 0.1% per year from 2005 to 2019. Then AI came along. The electricity demand of U.S. data centers is expected to double by 2027, reaching 66 GW from 31 GW in 2025. Data centers will account for 8.5% of summer peak demand by 2027, up from 4% two years ago. In Virginia, they have already consumed over one-quarter of the state's electricity. Texas's grid load grows 10% annually.

Steel mills and data centers are competing for the same batch of electrons. Aluminum smelting is one of the most energy-intensive processes on earth, and it is bidding against large-scale cloud operators with trillion-dollar balance sheets that are willing to pay any market clearing price. Under existing power contracts, aluminum plants always lose. Missouri's smelter was shut down this way, while just a few kilometers away, a server farm is breaking ground.

The losers complain about this. The winners flip the script.

Energy-intensive industries can no longer be customers of the grid; they need to become assets of the grid. Steel mills should provide self-generated power: using natural gas for now, then switching to next-generation nuclear and geothermal when they come online. Steel mills act as flexible loads: running at full capacity when electricity prices are low and feeding 100 MW back to the grid when data center demand spikes, earning money doing it. A smelter that can listen to load shedding commands is a virtual power plant that also produces metal. Place steel mills and data centers around shared generation and grid access points, because the queue for new grid connections is already "years," and a spot in line is worth more than the land beneath your feet.

And the surge in power demand itself is a tailwind. AI is forcing America to expand its power generation capacity at a pace not seen in two generations. Every gigawatt of power built for computing capacity is muscle for heavy industry to rely on for the next 40 years. The cheapest stable electricity wins heavy industry, period. The prosperity of data centers is pulling us in that direction, regardless of whether Washington has a plan. Build steel mills next to them.

4. You can get anything you need here

I want to be able to say "I need 8 different parts with 8 different manufacturing processes," and get them all in the U.S.

The real product of Shenzhen is not cheap labor, but distance. Need a gearbox, a coating, a set of molds, a custom bracket by 4 PM? There is someone within an hour who can do it, and it will be at your dock the next morning. This is the truth behind the entire 48-hour iterative closed-loop. It is not one supply chain, but every supply chain, within reach.

The American version does not need to be one hour; it needs to be within the borders of the country. Anything a factory needs can be sourced domestically in just a few days. From ore to alloys, casting to coatings, fasteners to complete machines. This is the standard. We are still far from this standard.

Using honest standards today, almost everywhere fails: can you create a complete product, an engine, a missile, a machine tool, without letting a single purchase order cross the Pacific? China refines about 90% of the world’s rare earths, while each F-35 contains 920 pounds of rare earths. Our machine tools come from abroad. Our aluminum relies on imports. There are cutoff switches in the supply chain, and we do not control the switch.

So, list every key raw material and fill in the gaps one company at a time. Rare earth refining, forged and cast parts, special alloys, machine tools—these are the unsexy middle tiers that no one tweets about but on which everything depends. Then overlay a layer of software across the entire system, allowing an Ohio shop to type "6061 aluminum sheet, anodized, 500 pieces, Thursday," with goods coming from 300 miles away instead of from a container ship.

The ultimate promise is a sentence you can print on the wall: Whatever you want to make, you can get it here. When this statement is valid, the return of manufacturing is no longer a patriotic slogan but an obviously correct business decision. At that point, the flywheel will begin to turn on its own.

5. Leverage technology; this is our unfair advantage

Code is a good thing; write more code. But production capacity comes first.

This is where everyone gets it wrong. China builds capacity first and then stacks technology on top. America tries to jump over capacity and relies solely on the technology layer. That game was lost. But it left us with something valuable: the best software, AI, and robotics talent on the planet.

The correct approach is not to use software to replace factories, but rather to let software grow within the factory from day one.

As of 2023, China has surpassed us in robot density, with the number of operational industrial robots about five times that of the United States. But their software stack is not better than ours. Ours is better. We just never pointed the gun at the shop floor.

A new factory built in the U.S. in 2026 should be AI-native, just like a startup is cloud-native. Quotes should be in hours, not days. Real-time scheduling, not whiteboards. Predictive maintenance, not $2 million per hour in unplanned downtime. Collaborative robot units handle repetitive tasks, allowing certified welders to do what robots cannot. Every repair, every weld gun adjustment, every manual override, all that data should be captured, because the TIG welder with 30 years of experience has tens of thousands of micro-decisions to make, and he’ll be retiring in four years.

Run the floor and gather data. Data trains the software. Software makes the floor run faster. This loop will yield compounding growth, and vertically integrated companies on both sides will ultimately win. A brand-new American factory using 2026 software can outperform a Chinese factory dating back to 2005 using 2015 software. We can skip their legacy code. That is the entirety of the advantage of starting late.

6. Get the government working in the same direction

I think the U.S. is doing okay on this, with a basic bipartisan consensus, but the government needs to play a big role.

Shenzhen did not happen by accident. In 1980, Beijing designated it as an economic special zone, then let go of everything aside from capital, land, and speed. The state clears the road, and the market races ahead.

The U.S. government is doing exactly the opposite. Approvals take years. Environmental reviews take longer than product iteration cycles. The SBA loan limit is $5 million, while machines cost $8 million. The financing system is designed for real estate, not for CNC machining units.

What Washington should do, and only should do:

Speed. Industrial approvals in 90 days, not 5 years. If it takes longer to approve a smelter in China than to approve a parking lot here, we will lose just on paperwork.

Demand. The Pentagon is the best anchor customer on Earth. Multi-year, large-volume procurement has already been proven feasible by the recent ammunition expansion case. 2027 is the deadline to expel Chinese rare earths from U.S. weapon systems; this is the correct type of forced propulsion. Set multiple deadlines like this.

Capital. Provide loan guarantees for industrial equipment and facilities, insured by those who understand what output a machine tool delivers. "U.S. Manufacturing Loan Guarantees" is a start; scale it by ten times.

Then stop. Do not pick winners, do not create five-year industrial plans, do not let bureaucrats design factories. Clear the road, anchor demand, guarantee capital, and then step aside.

Thread throughout the text

Shenzhen took 45 years. We do not have 45 years, nor do we need 45 years because we are not starting from a fishing village. Our starting point is the deepest capital markets on earth, the best software talent, the strongest demand signals globally, the return of defense plus manufacturing, combined.

Fix the culture so people want to make things. Build factories so there are places to work. Secure energy so no one can win in the electricity bidding war. Complete the supply chain so that everything you need is already here. Fill the factory with technology so that production lines run faster than any in Guangdong. Have the government focus on speed, demand, and capital; do not touch anything else.

Shenzhen is not magic. It is density, speed, and status. All three are replicable, and all three can be defeated.

Go build factories.