Japanese Firms in China Are Quitting Customers and Keeping Factories
Teikoku Databank counts a record-low 10,118 Japanese firms in China, down 22% in two years. But the share calling China their top production base fell only 5 points since 2019, while the share calling it their top sales base fell 16. Japan is writing off the Chinese customer and keeping the Chinese plant.
The headline count is dramatic. Teikoku Databank's ninth survey of Japanese companies in China found 10,118 with a presence there in June, the fewest since the series began in 2010, and a record 4,137 exits or disappearances in two years. Taken at face value, the numbers say Japan is leaving. The better reading says Japan is giving up on selling to Chinese customers while keeping the Chinese factory, which is a different bet with different winners and losers.
Ask the surviving firms where China matters most and the split shows. Since 2019 the share of China-present companies naming China their most important production base slipped from 49.0% to 43.7%, a drop of about a tenth. The share naming it their most important sales base went from 47.7% to 31.3%, a drop of a third. China is still about five times as important as Vietnam as a production base, where 8.7% of firms put their main plant. Capital is staying where the supply chain is. What is leaving is the ambition to sell.
The arithmetic of the count is worth seeing. Exits of 4,137 less new entrants of 1,221, the fewest of any survey and below even the 1,352 of the Covid years, gives the net fall of 2,916, so the drop is as much a story of thin arrivals as of heavy departures: few Japanese firms of any size now think China is where a new venture should start. That fits a market that Teikoku Databank describes as having moved from the world's factory to a giant market, where entry costs more and rewards less. The composition of the departures says the same thing. Small companies are shrinking as a share of the total, with firms under ¥100 million in sales now 4.8% of the count and those between ¥100 million and ¥1 billion 24.3%, while firms above ¥1 billion are rising. The firms that left were mostly the ones without the balance sheet to wait.
Look at which industries emptied. Manufacturing fell from 5,139 firms to 4,206, down 18%. Wholesale fell from 4,218 to 3,320, down 21%. Services fell from 1,803 to 1,213, down a third, with IT and software services the largest single decline. The people who sell, distribute and advise went faster than the people who make things. Textile makers fell from 310 to 218.
Geography agrees. Shanghai, where foreign firms keep regional head offices and sales teams, lost 21.5% of its Japanese companies in two years, falling to 3,961. Beijing has half as many as in 2022. Jiangsu, home to the factory belt, lost only 7% and still has 1,517. Guangdong has 1,409. Shanghai's offices and Beijing's service desks emptied while the plants carried on.
The reasons sit on the demand side. Teikoku Databank lists a property slump, state-backed Chinese competitors pricing low in the same segments, worsening relations between Tokyo and Beijing, and rare-earth export restrictions it describes as economic coercion. Every item on that list makes a Chinese customer harder to win, and none of them touches a plant that exports. The count peaked at 14,394 in 2012, the year of anti-Japanese protests over the Senkaku islands, so Japanese business has lived with political risk in China for a long time and has learned to price it.
Where did the sales ambition go? Thailand doubled as a sales base, from 4.8% to 9.7%. The United States is at 9.4%. India rose from 1.4% to 4.9%, and Indonesia from 1.8% to 3.1%. None is large beside China's 31.3%, but all four are growing from small bases, and they are the places where Japanese managers now expect to find customers. Southeast Asian sales channels and distributors gain most.
A record exit number looks like the strongest evidence for the exodus story, and Teikoku Databank itself says consolidation and closures may increase over the medium term. They may, and the composition tells you who. A count that falls by a fifth because trading houses, software outfits and consultants close their Shanghai offices is a different event from one that falls by a fifth because factories shut. Jiangsu barely moved. The researcher's own conclusion is a decision to keep the China business while lowering dependence on it.
The tariff survey the firm ran in October 2025 supports that. Among China-present manufacturers, 58.4% expected lower profit for the year because of US tariffs, against 33.4% of all respondents. Plants in China carry export risk that offices in Shanghai never did, and the answer firms gave was to diversify the buyers, which is what the sales-base numbers show.
For investors the consequence is a sorting rule. Japanese suppliers of components and materials sold into Chinese factories look safer than Japanese brands sold to Chinese consumers, because the first group follows the plant and the second follows household demand, which is where the property slump bites. A parts maker with customers in Jiangsu loses nothing when a Tokyo consumer brand gives up on Shanghai. Companies that report a falling China sales share beside a stable China production line are following the same plan, and that pattern is a plan, not a fire sale.
The sales side has no near-term rescue. The restrictions on rare earths, the revised anti-espionage law and the price war with state-backed rivals all remain in force, and the relationship between Tokyo and Beijing offers no sign of a thaw that would send Japanese brands back to Chinese shoppers before the next survey.
Teikoku Databank repeats the survey every two years. Our bet is that the 2028 edition shows fewer than 9,000 Japanese firms in China, a further fall of about 11% and half the pace of this one, with Jiangsu and Guangdong holding up better than Shanghai and Beijing. The share of surviving firms calling China their main sales base will be below today's 31.3%, and the plants will still be running.