The dark-factory concept, also called lights-out manufacturing, has existed since the 1980s, but no documented case does without people entirely. What has changed is the scale: consumer-electronics plants now make ten million phones a year with a few hundred people on site.

Xiaomi runs a plant near Beijing that assembles a phone roughly every six seconds, with a stated 81% automation rate and about 220 people overseeing the whole line, not each unit (Euronews[1]). Fanuc builds robots with robots at a plant near Mount Fuji: in 2003, Business 2.0 counted about 50 robots per 24-hour shift and up to 30 days unsupervised (CNN Money[2]). Philips assembled shavers with 128 robot arms in Drachten, the Netherlands, in 2012 (The New York Times[3]). Below are the five cases with date and source, and what each says about the role people still play.

Key takeaways

  • The strict definition is a factory that runs full shifts with no human presence on the production line; the practical definition is softer: minimal presence only for occasional intervention.

  • Mature cases share three traits: few references with low variability, high volume, and an also-automated input supply chain.

  • The model breaks on variability: each reference change requires reconfiguring robotic cells, recalibrating vision, and adjusting trajectories, work that tends to be faster with human operators.

  • Cobots (collaborative robots without safety cages) are gaining ground faster than the full dark model in Europe: cheaper, reusable across tasks, and socially less disruptive.

  • Spain has automated factories but no strictly dark factories: high product variability, strong union consensus, and the economics of the last human links explain the gap.

What defines a dark factory

The strict definition is a factory that can run full shifts with no human presence on the production line. The practical definition is softer: a factory where human presence on night or weekend shifts is minimal and limited to occasional intervention. Almost no real factory runs completely dark 24/7, but they do during lower-demand shifts.

Turning the lights off is practical. Robots do not need visible lighting, and climate control built for people stops making sense with nobody on the floor. In 2003 a Fanuc vice president put it this way: "Not only is it lights-out, we turn off the air conditioning and heat too" (CNN Money[2]). The exact savings vary a lot by plant type and there is no reliable universal figure, but it is part of the model’s economics, even if rarely the deciding factor.

Which dark-factory cases are documented, with date and source

The cases quoted again and again have a specific date and source, and none of them is a plant without people. These are the five I could check against the original source, in date order:

  • Fanuc, Japan (report of June 2003): in one of its 40,000-square-foot factories near Mount Fuji, robots built other robots at about 50 per 24-hour shift and could run unsupervised for up to 30 days. They stopped when there was no room left to store the output (Business 2.0, on CNN Money[2]).

  • Philips, Drachten, the Netherlands (August 2012): 128 camera-guided robot arms assembled electric shavers. At the sister factory in China, hundreds of workers did the same job by hand (The New York Times[3]).

  • Siemens, Amberg, Germany (February 2015): machines and computers handled 75% of the value chain and the factory made about fifteen million Simatic products a year. Siemens itself stressed that people remained indispensable for planning production and handling unexpected incidents (Siemens press release[4]).

  • Xiaomi, Changping, Beijing (July 2024): the factory opened with a 2.4 billion yuan investment, 81,000 m² and 11 phone lines, for about ten million units a year, running around the clock (Yicai Global[5]).

  • Xiaomi, Beijing (April 2026): a phone leaves the line every six seconds, with a stated 81% automation rate and about 220 people handling inspections, anomalies and quality control (Euronews[1]).

Xiaomi appears twice because the two sources measure different things: the 2024 announcement is about capacity and the 2026 visit is about how many people are still on the floor. That gap between the headline and the real plant is the constant across every case.

What the cases that work have in common

Mature dark-factory cases share traits:

  • Few references with low variability: phones, flash memory, standardized electronics, serial mechanical parts. The more uniform the product, the easier to automate.

  • High volume, which amortizes investment in specialized robotics.

  • Input supply chain also automated: materials arriving on continuous conveyors or autonomous vehicles, not boxes someone has to open and sort.

Xiaomi’s Changping factory fits all three traits: 11 lines dedicated to one kind of product, with most of the equipment developed by Xiaomi itself (Yicai Global[5]). The end product is always a phone with small variations in color and memory, which keeps unit-to-unit variability manageable for the robots.

Fanuc is the oldest case: already in 2003 it built industrial robots with industrial robots, at about 50 per 24-hour shift. The line could run up to 30 days straight with no one intervening (CNN Money[2]). Siemens in Amberg is the useful counterexample: with 75% of the value chain automated, it still needs people for the unexpected (Siemens[4]).

Where the promise breaks

The model breaks on variability. Any process that handles dozens or hundreds of distinct references, with different geometries, heterogeneous materials, or frequent design changes, does not fit the full dark-factory model. Each reference change requires reconfiguring robotic cells, recalibrating vision, and adjusting trajectories, and that tends to be faster with human operators.

Clear examples are European automotive plants that mix models on the same line. Volkswagen and Stellantis build different models on the same highly automated line, but not in dark mode. The reason is not lack of technology but economics: frequent reference changes and product complexity make the marginal cost of removing the last operators exceed the savings.

Another break point is internal supply-chain reliability. In a well-running factory, a tilted part or a jammed screw is a rare incident an operator resolves in seconds. Without operators, that same incident can stop a full line for hours until someone arrives.

Industrial automation line with robotic arms palletizing bread, showing the type of repetitive, low-changeover process where the dark factory model makes the most economic sense

The cobot question

An interesting trend complicating the dark-factory definition is the rise of cobots, collaborative robots designed to work near humans without safety cages. Universal Robots, FANUC CRX, and ABB’s collaborative lines have normalized in mid-size factories, especially in Germany and the Nordics. The global cobot market is estimated at 1.42 billion dollars in 2025, growing to 3.38 billion by 2030, a roughly 18.9% annual growth rate (MarketsandMarkets[6]). I covered this hybrid model in more depth in the 2026 collaborative-robotics balance sheet.

The cobot is the philosophical opposite of the dark factory. Instead of removing the operator, it augments them with a robotic partner that handles repetitive tasks while the human handles exceptions. The resulting model is not dark but hybrid: one person supervises two or three collaborative cells that together do the work of five traditional operators.

In practice this hybrid model is gaining ground faster than the full dark model in Europe. The reason is economic and sociopolitical at once. Cobots are cheaper than complex robotic cells, reusable across tasks, and their deployment does not require the same social consensus as closing a factory entirely to personnel.

Computer vision: the quiet enabler

The technical advance most pushing dark-factory viability in recent years is computer vision. Industrial cameras combined with deep learning models have dropped a lot in price. Entry-level AI visual-inspection systems now run 3,000-10,000 euros and mid-tier ones 20,000-50,000. Cost reductions versus manual inspection are at least 40% (Jidoka[7], AMD Machines[8]).

This changes the math because quality inspection was one of the last jobs where humans were hard to replace. Recent Vision Transformer-based models have closed that gap category by category. In sectors like food processing, where product variability is intrinsic, this was the main bottleneck. That is changing, though not overnight.

The Spanish reality

Spain has highly automated plants, especially in automotive, but I have not found any documented as a dark factory in the strict sense. The reasons match the rest of Europe: high product variability, strong union consensus, and economics where the last human links are cheaper than automating them.

If unstaffed shifts appear, expect them in the sectors that fit the three traits above: electronics and serial components with low variability. Siemens in Amberg marks the documented European ceiling: 75% of the value chain automated, with people for the unexpected.

The transition toward more darkness in Spain will likely pass through cobots and robotized cells in sectors with high cost pressure and low variability. We will not see a Spanish automotive plant running 100% dark this decade, but we will see more autonomous cells inside hybrid plants.

My take

The dark factory is a real model, mature in certain sectors and geographies, but exaggerated in general applicability. Painting it as the universal future of manufacturing ignores that product variability, changeover economics, and limits of automated fault detection still condition where going human-free pays off. The documented cases, from Fanuc in 2003 to Xiaomi in 2026, are real but specific to high-volume, low-variability sectors, and all of them keep people on site.

For an average Spanish manufacturer the reasonable horizon is not dark factory but intelligent automation with smaller teams and integrated cobots. That transition is already underway and is far less disruptive than the full-darkness promise. Investing today in collaborative cells and computer vision is concrete progress toward the future.

What matters to understand from Spain is that Chinese dark factories already compete with us on production cost for certain products. Responding by trying to replicate their exact model makes no sense, but understanding where labor cost is the dominant factor and where it is not does. In sectors where labor is 40% of cost, the Chinese dark factory is unbeatable medium-term. In sectors where labor is 10%, the difference is nearly irrelevant and there is room to compete on design, quality, and flexibility.

Smart industrial strategy identifies which is which instead of trying to win a race that cannot be won.

This article is also available in Spanish: Fábrica oscura: robótica y automatización hoy. I have also written about the digital thread, the traceability a plant needs before it cuts human presence.

Sources:

  1. Euronews: Inside Xiaomi’s Beijing factory: the machine behind a smartphone made every 6 seconds[1]
  2. Business 2.0 (CNN Money), 1 June 2003: Fade to black, on the Fanuc plant[2]
  3. The New York Times, 18 August 2012: Skilled work, without the worker[3]
  4. Siemens, 23 February 2015: German Chancellor Merkel visits Siemens’ digital factory in Amberg[4]
  5. Yicai Global, 9 July 2024: Xiaomi starts mass producing phones at new smart factory in Beijing[5]
  6. Wikipedia: Lights out (manufacturing)[9]
  7. MarketsandMarkets: Collaborative robot market to be worth 3.38 billion dollars by 2030[6]
  8. Jidoka: AI visual inspection system pricing guide for manufacturers[7]
  9. AMD Machines: AI quality vision reduces inspection costs 40%[8]

Frequently asked questions

Are there dark factories in Spain?

I have not found one documented in the strict sense. Spain has highly automated plants, especially in automotive, with continuous human presence. The reasons are high product variability, strong union consensus, and economics where the last human links are cheaper than automating them. Even the reference cases keep people: Siemens in Amberg automates 75% of the value chain and Xiaomi keeps about 220 people on its Beijing line.

What kind of production fits the dark-factory model?

Mature cases share three traits: few references with low variability (phones, flash memory, standardized electronics, serial mechanical parts), high volume that amortizes specialized robotics, and an input supply chain that is also automated. The model breaks on variability: each reference change requires reconfiguring robotic cells, recalibrating vision, and adjusting trajectories, which tends to be faster with human operators.

How much does an AI visual-inspection system cost today?

Entry-level systems run 3,000-10,000 euros and mid-tier ones 20,000-50,000, with cost reductions versus manual inspection of at least 40%. That price drop matters because quality inspection was one of the last jobs where humans were hard to replace. Recent Vision Transformer-based models have closed that gap category by category, including sectors like food processing where product variability is intrinsic.

Sources

  1. Euronews
  2. CNN Money
  3. The New York Times
  4. Siemens press release
  5. Yicai Global
  6. MarketsandMarkets
  7. Jidoka
  8. AMD Machines
  9. Wikipedia: Lights out (manufacturing)