Yiwu sourcing interpreters: the 214 Canton Fair exhibitors filed under another city

Checked on 20 September 2026. We are an interpreting company in Guangzhou (Guangzhou GuanRi Translation Service Co., Ltd.) — not an audit or inspection body, not a law firm, and not the Canton Fair organiser. The service this article recommends is the one we sell; read it on that basis. This is not legal advice and it is not compliance advice. Nothing below decides anything for anybody: the company decides whether you get in, and the organiser decides who enters the Pazhou halls. Wherever a figure comes from our own data rather than from an official document, we say so in the passage that uses it.

Our Canton Fair exhibitor directory holds 26,270 companies. It has a city column. Yiwu is not in it.

That is not a bug and it is not an accident: Yiwu is a county-level city administered by Jinhua, so the 214 Yiwu exhibitors we can find sit inside a cell that reads "Jinhua, Zhejiang" — next to 259 companies from the rest of Jinhua that are, on every measure we ran, a different population. This article is about what those 214 records support and what they do not, and about what an interpreted day in Yiwu is actually buying. It is not legal advice and it is not compliance advice.

1. Which row is you

Every row below ends somewhere you can act, and names who actually decides.

2. What this data is, and what it is not

The numbers below come from our own Canton Fair exhibitor directory: 26,270 companies, sourced from the 139th session's full exhibitor list dated 16 April 2026. Everything in it was entered by the exhibitor at registration. We did not survey anyone.

Four limits matter for this article specifically.

The registered address is a registered address. It is where the company is on file, not where a production line is. A company registered in Yiwu can run its line in another county or another province, and the directory has no column for a plant address.

The geographic columns are not copied through as they stand. The source table has only province, city and address — there is no district column at all — and where the province or city is blank or holds a generic value, we backfill it: from the address, then from a county-level-city lookup, then from the company name. The district column is entirely ours, cut out of the address. That backfill is exactly why Yiwu ends up inside Jinhua, and it is why section 3 exists.

We do hold registered addresses and contact details, in a separate file. To be exact about it: the validated company landline and website have been published since 10 September 2026, while the street address, the contact person's name, the mobile number, the fax and the e-mail are not — that material was collected together with personal information, and the address is a registration address rather than a plant address. We use it here only to count, never to publish a single company's details.

The attended_139 flag is derived from each company's exhibiting history, not from an attendance list at the door. Of the 214 Yiwu records, 153 carry it. The other 61 are not absences: 30 have histories ending at the 138th session, 28 at the 136th, 3 at the 137th. Do not read it as an attendance rate.

3. The directory has no Yiwu value, so we rebuilt one

473 records carry the city value "Jinhua". None of the three geographic fields — province, city, district — says Yiwu. That is not something our build lost: we went back to the 26,272 rows of the source table, and the city column simply has no value "Yiwu" in it anywhere; of the 221 source rows whose address or name mentions Yiwu, 207 have an empty city and 10 say Jinhua.

The district column needs saying plainly, because it reads like something an exhibitor failed to fill in. The source has no district column at all; ours is cut out of the address, by stripping the province and prefecture-city names and then looking for a district, county or town at the front of what is left. Yiwu is a county-level city, so what is left begins "Yiwu City" and the rule finds nothing — all 214 blanks are a mechanical consequence of that rule, not an observation about what exhibitors registered.

So we rebuilt the set from the registered address and the company name:

• 208 of the 473 have "Yiwu" inside the registered address.
• 167 have "Yiwu" inside the company name; 6 of those have an address that does not mention it — addresses like a road name in Beiyuan or Nianersanli, which only resolve to Yiwu because the company is called Yiwu something.
• The union is 214 records, 45.2% of Jinhua's 473.

The boundary of this set has to be stated exactly: it is the records whose city field says Jinhua and whose address or company name says Yiwu. Six further records have "Yiwu" in the address with some other city in the city field, and we read all six. They are not the same thing:

Four are false positives of a plain text search — two in Khorgos on the Kazakhstan border, where the address names a wholesale complex itself called the Yiwu International Trade City; one is a street called Yiwu Road in Shuyang, Suqian; one is a market called Yiwu Shangcheng in Zibo. The other two are not false positives at all: they are plain Yiwu street addresses, filed under Xianning and Ningbo. One reads "Yiwu City, Jinfuyuan Block C 608-613", the other "Zhejiang, Yiwu, Beiyuan, Qiushi Road 113" — and inside our own 214 two companies are registered in that same Jinfuyuan complex (B3-68 and Block B Gate 1) and four more at Beiyuan, Qiushi Road 113.

Separately, 16 of the 259 remaining Jinhua records cannot be placed in any county at all; 13 of those say "Jinhua" in the address and 10 name a development zone or industrial park inside Jinhua proper, so they are probably not Yiwu, but "probably" is the honest word.

Which means: 214 is a floor, not a census — and we already know the floor is at least 216. Those two records are the live proof: a Yiwu company can land in another city's pile because the city field was filled in wrongly, and a search keyed on "Jinhua" will never reach it.

We did sweep the Jinhua pile itself: run the remaining 259 records against the name of every Yiwu town and subdistrict, and nothing comes back — not one of them names one. So what we are missing is not in the Jinhua pile; it is in other cities' piles. It is also easy to see how it gets missed: three of the six name-only records above give a road name and nothing else, and no method keyed on place names can see that shape. Every number in this article is a number about those 214.

For you the practical form of this is short. If someone tells you a supplier is "a Yiwu company", that is not something they read out of this directory, because this directory does not say it. The place it is actually written is the registered address on the business licence. Ask for the licence, check the unified social credit code on it against the code on your contract and your invoices, and if you want to check that code independently it is searchable on the National Enterprise Credit Information Publicity System (gsxt.gov.cn), and the lookup is free. One caveat: that site blocks automated access hard — three script requests to it on 19 September 2026, and three more on 20 September, all came back HTTP 521 — so open it in a browser, and we cannot confirm for you whether it opens from your network. A supplier is entitled to decline to hand over the licence; a refusal is not evidence of anything, it only means you cannot settle the ownership chain before you travel, and the trip should be valued on that basis.

4. One city cell, two different populations

Here is the split that makes the reconstruction worth the trouble. Same province, same city field, exactly opposite shape:

Trading companies are 58.4% of the Yiwu records and 20.1% of the rest of Jinhua — 38.3 points apart; the manufacturer column runs the other way, 28.5% against 53.7%, 25.2 points apart. Rank all 69 city groups in the directory that have at least 100 exhibitors, and the two halves of this one city field land 44 places apart: Yiwu 11th on trading share, the rest of Jinhua 55th.

So "a supplier in Jinhua, Zhejiang" is not a description. Until you know which of these two you are dealing with, the sentence carries no information at all — and the city column will not tell you, because it gives both of them the same value.

5. Yiwu leans trading — but not as far as you think

The market-city reputation runs ahead of the data in three specific ways, and all three are worth having in front of you before you book anything.

First, Yiwu is not the extreme. On trading-company share among those 69 city groups, it ranks 11th at 58.4%. Ten groups are ahead of it. The top five are Xi'an at 72.1% (197 exhibitors), Nanjing 71.2% (340), Hefei 70.7% (338), Zhengzhou 66.2% (201) and Wuhan 64.3% (238); the other five are Dalian 63.6%, Shanghai 63.5%, Shijiazhuang 63.2%, Xiamen 62.0% and Beijing 59.0%. Those ten include inland provincial capitals, coastal cities and municipalities — and not one of them is known for a wholesale market. The median of the 69 is Linyi at 33.1%; the bottom is Chaozhou at 3.4%.

Second, 61 of the 214 — 28.5% — register as manufacturers, and a further 28 as manufacturer-and-trading. That is more than two in five carrying a production side (89 of 214, 41.6%). Whatever you have read about Yiwu being wall-to-wall traders, these records do not say that.

Third, the reputation gets one more thing wrong, and this is the sharpest number in the section. Everyone pictures the wholesale market. In the registered addresses of all 214 Yiwu companies, seven name the international trade city complex or a stall in it. In the entire directory of 26,270 companies, only six addresses anywhere contain the word 商位, the term for a numbered stall inside a wholesale market — five of them Yiwu, one in Linyi. The address column simply does not record who sits in a market and who does not. If that is the distinction you need, it has to come from the supplier.

6. Two signals you can use as a prior, and where each one fails

Two fields in these records do carry real information about what kind of company you are looking at. Both are self-declared at registration, so treat them as a prior — the thing that tells you which question to ask first, not the answer.

The company name. 105 of the 214 Yiwu companies — 49.1% — have "import and export" or "trading" in the name. Of those 105: 97 are registered as trading companies, 8 as manufacturer-and-trading, and zero as manufacturers. Not one counter-example in 105. Across the whole directory the same test leaves 1.9% manufacturers (102 of 5,492), and in Foshan 2.2%, Shenzhen 2.9%, Guangzhou 0.6% — so the signal is real everywhere and unusually clean here.

The reverse direction is much weaker, and this is where people go wrong: of the 109 Yiwu companies without those words, 61 are manufacturers (56.0%) but 28 are still trading companies (25.7%). A plain-sounding name tells you very little.

The registered headcount band. Half the Yiwu set — 109 records, 50.9% — is in the "under 50 employees" band. Of those 109, 88.1% are trading companies and 5.5% are manufacturers. And the manufacturers sit somewhere else entirely: of the 61, 38 are in the 150–500 band and 10 in 500–1,000, with only 6 under 50. Where this signal fails has to be said too: close to 12% of that 109-record band are not trading companies, and 6 of the 61 manufacturers are registered under 50 employees — a small headcount does not mean there is no line, it usually means so in these records. And the band, like the company type, is self-declared at registration.

Now the warning that has to travel with all of it. The company-type field is what the company registered itself as, and "trading company" does not mean "no factory". A manufacturer very commonly registers a separate trading entity and exhibits under that, and the directory holds one row per exhibitor, not one row per group. The prior tells you the first question is which legal entity owns the line, and where is it — it does not tell you the answer, and it is not evidence of anything about a specific company.

7. The product column here is too thin to carry a phase reading

In the Shenzhen article we tried to read a city's fair phase off its registered products. It did not survive its own test: capping our keyword list at an equal depth per zone moved the answer, so we withdrew the claim. For Yiwu we are not going to start, and the reason is visible before any method is applied.

Only 51 of the 214 Yiwu records filled in a product list at all — 23.8%. Among the 69 city groups, exactly two come in lower: the group whose city field is empty, which is not a city, at 23.79%, and Shenyang at 23.62%. One caveat on that: Yiwu is at 23.83%, four hundredths of a point above second place, so the ranking itself is fragile — what holds is the order of magnitude. By comparison Shenzhen is at 37.4% and Foshan at 40.7%.

The keyword column is better filled — 212 of 214 — but thin in absolute terms: 616 keyword instances, 550 distinct terms, 2.88 per company, and 505 terms (91.8% of the distinct ones) appear exactly once. Only 15 terms appear three times or more, and the most frequent is a tie at five occurrences each between "socks" and "trade" — one is a product and one is not a category at all. In fairness to the data and against our own case: 12 of those 15 are genuine product categories (seamless underwear, plush toys, picture frames, mirrors, paper bags, stationery, shirts, toys, hardware, tools, apparel and socks) and only 3 are not (trade, import and export, OEM). And on dispersion Yiwu is not unusual at all: on the share of all keyword instances held by the top ten terms, Yiwu's 5.8% ranks 34th of the 69 groups, squarely mid-pack, and slightly more concentrated than Shenzhen at 3.0% — and Shenzhen is the city where we did attempt the mapping. So our reason for not mapping is not that Yiwu is unusually scattered. It is the two below.

First: with an average of 2.88 keywords per company and three quarters of them not filling in a product list at all, spreading that across a fixed grid of 54 official zones leaves too few records per zone to look at. Second, and this is the heavier one: what such a count mostly measures is how many words we had written for each zone. We learned that the expensive way on 19 September and we are not repeating it here: no city-level phase lean for Yiwu, and no district-level one either. The route that does work is in the next section, and it does not run through us.

8. The 140th session's calendar, and which days it does not take from your supplier

Start with what the organiser has published — noting that the original says the session is planned to open on that date. The booth application announcement for the 140th session sets the dates (the lead line and the three phase lines, quoted as one block):

第140届广交会计划于2026年10月15日开幕，分三期举办线下展，其中： 第一期：2026年10月15—19日 第二期：2026年10月23—27日 第三期：2026年10月31日—11月4日

The same announcement states that the session has 54 zones, and its Annex 1 assigns each zone to a phase: 19 zones in Phase 1, 14 in Phase 2, 21 in Phase 3. Your supplier's phase follows from its zone, and the zone was settled when it applied for its booth. So ask the supplier which phase it is in and what its booth number is. If it will not say, the one route that does not depend on the supplier is the organiser itself — the channel it published in the 139th session's accompanying-personnel notice, in its own English (one sentence, followed by the contact lines; the two versions lay them out differently — two lines in Chinese, three in English):

For inquiries, please contact the Canton Fair Customer Service Center: Domestic: 4000-888-999 Overseas: (0086) 20-28888999 Email: info@cantonfair.org.cn

Whether the organiser will tell you which phase a particular exhibitor is in is its decision, and we do not speak for it.

Between the phases there are two gaps. The trade services zone announcement of 4 September 2026 names them in one sentence about its own opening hours (1 sentence):

服务时间为每天9:30—18:00，其中10月20—22日、10月28—30日换展期间闭展。

So the changeover days are 20–22 October and 28–30 October. Do not read "changeover" as "days this supplier is free". A changeover is the previous phase moving out and the next phase moving in, so the window immediately before and the window immediately after a given phase are exactly when that phase's exhibitors are in the halls handling freight. Counting it the other way round is what you need: what the fair takes from a supplier is its own phase plus the changeover window adjacent to it.

• A Phase 1 supplier is on stand 15–19 October and moves out 20–22. From 23 October the fair's calendar has no further claim on it.
• A Phase 2 supplier moves in 20–22, is on stand 23–27, moves out 28–30. It is free of the fair on 15–19 October, and from 31 October.
• A Phase 3 supplier moves in 28–30 and is on stand 31 October – 4 November. The thirteen days from 15 to 27 October are entirely untouched — both earlier phases and the changeover window between them fall inside it — and so is everything after the close.

Those three lines cover only the two changeover windows the organiser has published. There is also a move-in period before opening day, and no official notice gives its dates, so we are not going to fill them in. A Phase 1 supplier must be in the hall building its stand at some point before 15 October: the fair's calendar is empty on those days and the supplier's is not, so if you want a visit before 15 October, ask the supplier when its move-in starts.

That is our reading of the published dates and the closed-for-changeover sentence, not a statement the organiser has made about any one company; which day actually works still has to come from the supplier. And one thing those three lines do not cover: "the fair is not using them" is not the same as "the plant is running." Whether a given week includes a weekend the plant works, and how many days it takes for the National Day holiday in early October, can only be answered by the plant. The holiday schedule is published separately each year by the authorities, and we do not put dates in that we have not seen in an official notice.

And if you are not coming for the fair at all, none of this section binds you: those three lines describe whose time the fair takes, while your dates depend only on whether the plant is working — which only the plant can tell you.

9. How the day is priced, and what you are entitled to ask us

Yiwu is in Zhejiang. Our interpreters are in Guangzhou. So this is a travelling day before it is anything else, and we will not tell you how long the journey takes or how far apart two addresses are. Anyone quoting those numbers from a desk should be treated with suspicion: we do not have your supplier's plant address, the registered address in our file is not a plant, and conditions are not constants. What we can give you is an order of operations.

Ask for the street address in writing, door number included. In the 214 Yiwu records, 176 addresses (82.2%) carry a house number and 108 (50.5%) name a floor, room or building — but 10 give only an industrial-zone or park name with no identifier of any kind, and in 2 more the address field literally repeats the company name — 12 records, two shapes of the same problem. If what you are handed looks like either of those, you do not yet have an address. (Keep one distinction: an address that runs down to a numbered market stall, or to a unit in a named building, is not a missing address but one you can walk to; there are 2 of those in the set.)

Then ask the supplier to state the door-to-door route itself, and check it yourself on a map. Then ask which process steps the visit actually covers and how long their side takes — that number is theirs, not ours, and we will not estimate it from company size. Then add the return leg at the real hour you will be travelling. Only then do you know whether you are booking a half day, a full day, or two days.

One reason not to compress this: the 214 companies are not in one place. Matching the names of Yiwu's towns and subdistricts anywhere in the address — without requiring the words "subdistrict" or "town", because many addresses give only the place name — 159 records match, 74.3%, and they name 13 different places. The three most common are Futian (32 records), Beiyuan (21) and Fotang (16) — together 69 records, 32.2% of the set. Those are different towns and subdistricts, not one street. How far apart they are is a question for your map and the actual door numbers, not for us.

Two questions you are entitled to put to any intermediary, ourselves included — answered here for us, because it would be cheap to ask them of others and dodge them ourselves:

• How is travel time charged? It is charged, and it is a separate line from interpreting hours — as are transport, meals and any overnight stay. Our published floor is ¥200 per hour or ¥800 per full day for consecutive and escort interpreting, from our own FAQ page, depending on language pair and subject — and a travelling day outside Guangdong will not come in at that floor.
• How much of it reaches the interpreter? We are an intermediary. The quote covers the interpreter's fee plus our screening and the responsibility we carry; we do not publish the split.

And a third question that only has a per-booking answer: will the person who turns up be the profile you were shown? We are not going to state a general company policy here that would apply to every booking, because that is a term of a specific engagement. What you should do instead is have it written into the confirmation — the named interpreter, how a substitution would be notified, and whether you may refuse one. Cancellation and rescheduling terms belong in the same place. You are entitled to ask for all of it in writing before you pay anything, from us or from anyone else.

10. Our pool: 812 profiles, none of them in Yiwu

The pool behind our Guangzhou interpreter directory holds 812 profiles. The home-city field on every one of them reads Guangzhou; that field has exactly one distinct value. Yiwu is in another province, so for a Yiwu day we are sending someone to you — which is what section 9 is about.

By language, counting profiles rather than people, since one profile can carry more than one language: English 631, Japanese 248, French 100, Spanish 65, Russian 56, Korean 51, German 32, Italian 17, Vietnamese 8, Arabic 7, Thai 6, Indonesian 4, Turkish 3, Portuguese 1.

Read the tail of that list literally. There is one Portuguese profile in the pool, three Turkish, four Indonesian. (The next ones up are not thick either: Thai 6, Arabic 7, Vietnamese 8, Italian 17.) We do not have a Portuguese team; we have one person, and one person has no backup — which is why these are the ones to ask about early. That sentence rests on those counts, not on any guess about demand. And "early" has a scale we publish ourselves: our FAQ says rare language pairs or large conferences should be booked 4–8 weeks ahead, while common pairs (Chinese to English, Japanese or Korean) can usually be scheduled 1–2 weeks ahead. The 140th session opens on 15 October, 25 days after this article was written on 20 September 2026 — so for Portuguese, Turkish, Indonesian, Thai or Arabic, a Yiwu day inside the fair window puts you at the very end of that lead time, and a day before the fair opens puts you past it. If you need one of them and we cannot staff it, there are three routes and each has someone who owns it: we can look outside the pool, and we will tell you what we found rather than promise before we have looked; you can work through a relay language, but only if someone on your side speaks it; or you can move the date, which only works if the trip is not pinned to the fair. When none of the three applies, the honest answer is that we cannot take it, and it should reach you early enough to arrange something else. One more thing said plainly: hiring an interpreter locally in Yiwu is a real route. We have no data on it and it is not our business, and all we can tell you is that nobody in our own pool lives there.

What an interpreter is not, on a day like this:

• Not an auditor or an inspector. An interpreter does not sign an audit report, does not judge a production line, and should not be asked to.
• Not a badge-holder requirement — but that sentence has a range. A private visit to a company in Yiwu needs no Canton Fair badge of any kind. The certificate rules people quote — a CATTI or NAETI certificate, or a foreign-language graduation certificate from a mainland institution — are conditions the fair attaches to the "interpreter" class of its accompanying-personnel badge, not conditions for visiting a company. And if you want the same interpreter inside the Pazhou halls, that is a separate document on a separate clock: under the 139th session's rules only the buyer may file it, at least two calendar days ahead, and buyers holding a Chinese passport or from Hong Kong, Macao or Taiwan may apply for an assistant badge only, not an interpreter badge. The 140th session's badge notice has not been published as of today. Do not plan around one person covering both. But do not read that as "nobody will ask for anything": your own audit protocol, your company's supplier rules, or the site's own access and confidentiality requirements may each require something, and none of those is ours to interpret or to waive.
• Not a witness and not an expert. Do not plan to rely on an interpreter's memory to prove what the other side agreed to. Write it down on the spot and have both sides accept the written version. That is practical advice, not a legal opinion about what counts as evidence.
• Not the person who decides at the door. Whether you get in, whether you may photograph a line, whether you may see a particular process, is the company's decision on the day. We do not make it and we do not promise it.
• Not a lawyer. Nothing an interpreter says about contracts, licensing or liability is legal advice.
• And the trap in the other direction: the supplier's own salesperson interpreting for you is not a neutral interpreter, however good their English. That is not an accusation of bad faith — they are simply one of the two parties in the room.

11. If it has already gone wrong

• You arrived and it is an office. Say so on the spot, politely, and ask for the plant address and which licensed entity holds it. If they offer to take you to a "partner factory", go if you want to — but record it as a partner factory, not as this supplier's plant. Sections 4 and 6 tell you how common this shape is across the 214 — that is a base rate, and it predicted nothing about the company you are standing in. What you are standing in is the fact; the base rate only told you which question to ask first.
• The address you were given is only a zone name. Ten of the 214 look like that, and in two more the address field simply repeats the company name. Ask for the door number and the floor in writing before you book anything else around it; if it does not arrive, you are booking a day against an address that does not exist yet.
• You booked into the supplier's own phase or changeover days. That afternoon is usually unrecoverable — the plant may be running; what is missing is the people. Re-date it off section 8: Phase 1 suppliers from 23 October; Phase 2 on 15–19 October or from 31 October; Phase 3 anywhere in 15–27 October, or after the close on 4 November. Confirm with the supplier that the plant is working that day. If the supplier will not say which phase it is in, the one channel that does not run through it is the organiser: 4000-888-999 inside China, (0086) 20-28888999 from abroad, info@cantonfair.org.cn. Whether it answers you is the organiser's decision.
• You cannot find the company in any directory. Ours is a list of one session's exhibitors, not a register of companies, and section 3 showed that our own Yiwu set is a floor rather than a census — so absence from it means very little. What to do instead is ask for the business licence and run the unified social credit code on it through the National Enterprise Credit Information Publicity System (gsxt.gov.cn) yourself, as in section 3. Whether to carry this article's priors over to a company that is not in the directory is your call, not ours; they were measured on the 214.
• It is tomorrow and you have nobody. Message us on WhatsApp or WeChat: +86 139 2416 6640 (same number on WeChat), with the town, the date and the language. We will tell you whether we can take it; if the honest answer is "let me check availability" — which is most likely what the thin languages in section 10 will get — we will say that instead of "it should be fine". If you would rather not use a messaging app, the contact page reaches the same desk. And we are not the only route: if an audit firm is running this trip, ask whether they bring their own interpreter; if the supplier offers to provide one, section 10 explains why that is not neutral — it does not say you must refuse.

12. Sources and dates

• Announcement on Booth Applications for the Export Exhibition of the 140th China Import and Export Fair, Business Office of the China Import and Export Fair, 5 May 2026 — the phase dates, the "54 zones" figure, and Annex 1 assigning zones to phases: Canton Fair website. We did not find an official English version of it; the wording quoted from it in this article is in the original Chinese.
• Announcement on Exhibitor Recruitment for the Trade Services Zone of the 140th Canton Fair, 4 September 2026 — opening hours and the two changeover windows: Canton Fair website. We did not find an official English version of this one either; same treatment.
• Notice on Badge Application for Personnel Accompanying Overseas Buyers for the 139th Canton Fair, 2 April 2026 — the organiser's own customer service channel, quoted above from its official English: Chinese and official English. The 139th session's rules do not automatically apply to the 140th.
• Pre-registration for Purchasing Representatives of the 140th Canton Fair Is Open, 3 September 2026 — the current state of accompanying-personnel arrangements: Canton Fair website. We did not find an official English version of this one either.
• All of the above were retrieved from cantonfair.org.cn on 20 September 2026. Where we say we did not find an official English version, this is the check we ran: the site's own content interface returns an English twin under the same id with an EN suffix, and that probe returns the official English of the 139th accompanying-personnel notice quoted above while returning nothing for the three 140th-session notices. 104 of the 1,200 most recent items in the site's content stream carry that EN suffix, so English twins do show up there when they exist.
• Exhibitor counts, company types, addresses, headcount bands, keywords and product lists: our own Canton Fair exhibitor directory, built from the 139th session's full exhibitor list (16 April 2026), 26,270 companies. These are our numbers, not the organiser's. The interpreter figures come from our own pool of 812 profiles.
• Our published interpreting rates are the ones listed on our own FAQ page.

Related reading: the Guangzhou article has the district-level data for Guangzhou, the Foshan article covers a city whose registered types point the other way, and the Shenzhen article is where the phase-mapping method was tested and withdrawn. If you do want an interpreter inside the halls rather than at a supplier's door, the badge rules are in that article.