🇨🇳 Shenzhen Files (no. 12/25) — In Shenzhen, the incubator shortens the time.

There is a word you hear often in French incubators: “support”. Another you hear often in Shenzhen: “access”.

This is not a semantic detail. It is a complete difference in philosophy about what an incubator should be. And this difference explains, more than we think, why hardware startups progress two to three times faster in Shenzhen than elsewhere.

The question is not which of the two models is “better”.

It is understanding what each model optimizes. Support optimizes vision, structuring, preparation for fundraising. Access optimizes the speed between the idea and the object. In an environment where competition eliminates those who hesitate too long, this second optimum is not a luxury. It is a condition for survival.

Accompaniment or access.

The incubator model we know in Europe is a services model: a workspace, mentors who drop by once a month, introductions to investors, workshops on the business plan. It prepares founders to pitch. It helps structure a vision. It builds a network.

What it does not do, in general: shorten the time between an idea and a physical object.

In France, Station F is often cited as a reference. It is the largest incubator in Europe. A remarkable building, a dense community, prestigious partners. But Station F does not have a shared production line in the adjoining building. It does not have fifty 3D printers available for free without booking. It does not have component suppliers twenty minutes away who come to prospect the teams in residence. This is not a criticism. It is a description of what each model has chosen to optimize.

That is exactly what Shenzhen’s incubators understood, and redesigned.

When access to tools is immediate.

I spent two days at InnoX. What struck me first was not what I was told. It was what I saw.

Workshops. Machines. More than fifty 3D printers, available immediately, free of charge, without booking online three weeks in advance, without mandatory safety training, without a safety officer to convince. A 3D printer available now. A machine shop available now. The parts you need, ordered and delivered the same day because suppliers are twenty minutes away. And a shared production line, in an adjoining secured room, accessible to the teams in incubation.

The list is mundane to write. It is revolutionary to experience for a European founder used to scheduling a prototyping session the way you schedule an important meeting.

In Shenzhen, you do not schedule. You go downstairs, you build, you come back up with an object. And that object teaches you in one hour what three meetings could not have told you. The cost of hesitation has become higher than the cost of testing. That reversal is what explains the speed, not the culture.

At the time of my visit, InnoX was hosting 93 startup projects in the building, nearly 900 entrepreneurs. The incubator was officially founded in 2021: four years of existence, and already a hundred projects launched and taken to market.

Forbes described it as the Chinese Y Combinator. The team accepted it without claiming it.

What distinguishes InnoX from a well-equipped workshop is the human density around the machines. 900 entrepreneurs in the same building, working on hardware products at different stages of development. A founder stuck on a mechanics problem can knock on the door of a team two floors up who solved it six months earlier. A founder looking for a sensor supplier gets three recommendations before lunch.

The machine does not only produce prototypes. It produces circulating information, in real time, between peers.

An academy before an incubator.

What distinguishes InnoX from a classic incubator is the upstream. InnoX does not start by welcoming startups. It starts by training talents who want to become entrepreneurs.

Throughout the year, bootcamps of different types: the Winter Camp, open to all profiles including non-technical ones; the Geek Camp, reserved for engineers; the Teenage Camp, for middle school students. Children of 12 build functional products in three to four days. A teenager designed a sensory alarm system because he did not like waking up to an alarm sound. Another developed a portable medical device. These products are not mock-ups. Some became companies.

Professor Li Zexiang, founder of InnoX, has a formula that summarizes his vision of entrepreneurial segmentation: if you are under 30, do B2C. If you are over 30, do B2B. The reason is simple: before 30, you have no money or network, but you know what young consumers want because you are one. After 30, you have accumulated resources and connections, but you have lost instinctive contact with trends. This segmentation is not anecdotal. It structures the cohorts, the programs, the types of products incubated.

90% of InnoX projects are B2C. 90% of entrepreneurs are under 30. The coherence is not accidental.

Selection is rigorous but its criteria differ from what is practiced in Europe. They are not looking for a founder with an MBA and a five-year market vision. They are looking for someone with passion for innovation, experience in robotics competitions, an ability to build with their hands. The prototype precedes the pitch, here too in the admission criteria.

The joint program with KAUST (the King Abdullah University of Science and Technology in Saudi Arabia) illustrates the international reach of the model. Master’s students in entrepreneurship come to Shenzhen for four months precisely to access what they cannot find at home: the supply chain, the 3D printers, the responsive suppliers. They arrive with ideas. They leave with prototypes.

What Saudi Arabia sends its students to find in Shenzhen, Europe could do too. It does not yet.

The market decides before the prototype.

This discipline of iteration goes further than manufacturing speed. It touches the order of operations: in Shenzhen, you validate the market before building the product.

A founder incubated at InnoX described his method with a formulation that summarizes everything: “it reduces the difficulty of decision-making. You don’t need to spend so much time doing theoretical analyses to finally find that it doesn’t work. You take the component at a very low price, very quickly. And you test it with your hands. Instead of staying in the lab thinking about it, you do it.”

This shift, testing before analyzing, is made possible by the near-zero marginal cost of the test. When a component costs a few euros and arrives in two days, prior theoretical analysis becomes a waste of time. The object decides faster than debate. This is not recklessness. It is a rational response to a different cost structure.

The validation logic sometimes goes even further. A team incubated at InnoX described their method: generate an image of the product, post it on a resale platform, and measure real interest before even having a physical prototype. Buyers contacted the team to order. They had to refund and explain they were doing validation. “It’s a very crude method, but it validates the market well.” The prototype no longer precedes validation. It follows it.

Once the prototype is validated, the sequence continues without interruption: supply chain mentors to connect teams with suppliers, go-to-market support, launch strategy. This year, InnoX took 40 teams to CES in Las Vegas. The Pilates reformer incubated there launched its Kickstarter campaign from Shenzhen and raised several thousand dollars in a week. Not for the funding: to validate that the product would sell internationally, and signal to investors that there was real demand. Kickstarter as a proof tool, not a funding tool. The distinction matters.

DJI, which dominates more than 70% of the global civilian drone market, practices an institutionalized version of the same logic: cryptic teasers on Chinese social networks, showing almost nothing, letting fan communities speculate for weeks. The audience’s reaction measures appetite before the production commitment is final.

The startup posting an image on a resale platform and DJI publishing a fifteen-second video on Weibo are doing the same thing at different scales: letting the market decide before building. This discipline is not reserved for giants. It is taught from the first bootcamp.

The 30-minute innovation circle.

InnoX does not describe its location as an address. It describes it as infrastructure.

“We call it the 30-minute innovation circle: you can access all the key resources in half an hour.” DJI next door. EcoFlow next door. Venture capital funds next door. Universities next door. The supply chain next door.

Two years ago, InnoX was based in the Pingshan district, thirty minutes by train from the heart of Nanshan. The team deliberately moved to join this density. This cannot be called geographic luck, it was a deliberate strategic choice.

This move says something important about how InnoX thinks about its own role: the incubator alone is not enough. It is its position in the network that determines its value.

An isolated incubator, however well equipped, remains an island. An incubator anchored in a dense ecosystem becomes an accelerator of circulation, of information, components, capital, talent. The difference is not in the machines. It is in what surrounds the machines.

Behind the density, there is a funding logic that one of my InnoX interlocutors formulated with rare clarity: “the government does not directly fund startups, but puts money into the ecosystem to make life easier for startups.”

This is the exact inverse of the French model.

In France, subsidies go to company A, to company B, separately, case by case. BPI France funds projects or sectors. SATTs fund transfers. Regions fund isolated incubators. The result is a constellation of scattered support points, rarely interconnected, in an industrial fabric that has not been densified.

In Shenzhen, public money goes into the network’s nodes: shared infrastructure, common spaces, connections between actors; and startups benefit from the ecosystem thus built. Funding the nodes rather than the individual members: that is a political choice, not a cultural property.

HAX illustrates what this model is worth in the eyes of the world. Founded in 2011 in Shenzhen, the world’s first accelerator specializing in hardware, HAX developed its method in the Shenzhen ecosystem for ten years. More than 257 startups completed the program. Then, in 2021, following geopolitical developments, New Jersey signed a 25 million dollar deal to transfer the model to Newark: 35,000 square meters of equipment, mechanics, electronics and chemistry labs, built from scratch because Newark did not have what Shenzhen had naturally around its incubators. An American city paid to import what forty years of concentrated industrialization had made obvious in Shenzhen. This is not an anecdote. It is a confession.

Europe has incubators. It doesn’t have the rest.

Leaving InnoX, I noted an observation my interlocutor had made without attaching particular importance to it. I had pointed out that no French university figured among the program’s international partners. She had replied: “Oh, maybe you can help.”

Europe has excellent incubators. It has FabLabs, makerspaces, serious acceleration programs. Station F exists and is a strength for the French ecosystem. Hundreds of support programs exist.

What it has not yet managed to assemble is the embedding between the incubator and the surrounding industrial fabric. Suppliers twenty minutes away. Precision mechanics workshop thirty minutes away. Plastic injection an hour away. All of it interconnected by logistics that deliver within the day.

This is not a lack of will. It is rather a lack of accumulated industrial density. Newark built 35,000 square meters of equipment to compensate for the absence of a dense industrial ecosystem within cycling distance. It is the only possible response when the ecosystem is not there: internalize everything. But internalizing is expensive, does not adapt, and does not produce the network effects that Shenzhen’s density spontaneously generates.

An incubator in Shenzhen is a node in a network. An incubator elsewhere is a well-equipped island.

This density cannot be created by decree. It is the product of decades of concentrated industrialization. Shenzhen has forty years of head start.

French startups that come to prototype in Shenzhen arrive alone, without institutional bridge, without facilitation. They leave with their product but without anchoring in the ecosystem.

No French university among InnoX’s partners means not only a missed visibility opportunity. It means no French structure plays the role of systematic connector between French founders and this ecosystem. Every founder has to rebuild the path from scratch. This entry cost is not prohibitive. But it is real, and it is paid in time.

Understanding exactly what we are trying to build, before announcing we already have it.

Shenzhen Files is a series of 25 articles, drawn from 25 days spent in China (mainland and Hong Kong) meeting founders, investors, engineers, lawyers and researchers.