Gold & Metals

Three Chinese counters rise 12 to 14 yuan, Lao Miao flat

Lao Feng Xiang read 1,262 yuan on 30 September, Chow Tai Fook 1,259 and Chow Sang Sang 1,257. Lao Miao held at 1,242, and the three with a published base sit 10.71% to 11.73% below the 25 August high.

By Carat Capital

What changed

The counters turned back up. Three of four Chinese brand gold counters rose on 30 September. Lao Feng Xiang gained 14 yuan a gram to 1,262. Lao Miao held at 1,242 and did not move.

What it means

  • If you buy gold in China: Three counters rose into the holiday week; the cheapest did not.
  • If you price against a counter: The 20-yuan gap is brand pricing, not metal.
  • What this is not: A 1 October board. The figures close on 30 September.

Key figures

Lao Feng Xiang, yuan a gram, 30 Septemberup 14 yuan on the prior day, the dearest of the four1,262
Chow Tai Fook, up 13 yuan1,259
Chow Sang Sang, up 12 yuan1,257
Lao Miao, unchanged on the day1,242

Source: Huxiu, 30 September 2026 board, page read direct, 70,953 bytes · Sina Finance / National Business Daily, 29 September 2026, 32,318 bytes · 25 August base from Carat Capital, 27 August 2026

Three counters up, one flat

Fourteen yuan dearest. Lao Feng Xiang quoted 1,262 yuan a gram on 30 September, up 14 on the prior day. It is the dearest of the four.

Two moved less. Chow Tai Fook read 1,259, up 13 yuan, and Chow Sang Sang read 1,257, up 12. The two are two yuan apart.

Lao Miao held. Lao Miao read 1,242 and the source records no change on the day. It is 20 yuan under Lao Feng Xiang.

Four brand counters, 29 and 30 September30 September 2026 board and its own stated day-on-day moves · 25 August base published 27 August 2026
Brand30 SepMove29 Sep implied25 Aug high% of the high
Lao Feng Xiang 老凤祥1,262+141,248not published-
Chow Tai Fook 周大福1,259+131,2461,41089.29%
Chow Sang Sang 周生生1,257+121,2451,40889.28%
Lao Miao 老庙黄金1,242unch.1,2421,40788.27%
Notes on this table

All figures are yuan a gram for pure-gold jewellery. The 30 September levels and the four moves are the source's own; the 29 September column is Carat Capital's subtraction of each stated move and is shown because it reproduces three of the four figures Carat Capital published for 29 September to the yuan. The last column is Carat Capital's division of the 30 September level by the 25 August high, printed so the share and the gap can be compared: Chow Sang Sang 1,257/1,408 = 89.28% of the high, so 10.72% below it; Lao Miao 1,242/1,407 = 88.27%, so 11.73% below. No 25 August figure was published for Lao Feng Xiang, so no comparison is given for it.

The moves confirm the prior session

Subtract each move. Taking each stated change off its 30 September level gives 1,248, 1,246, 1,245 and 1,242 for the four brands on 29 September.

Three match exactly. Carat Capital published 1,246 for Chow Tai Fook, 1,245 for Chow Sang Sang and 1,242 for Lao Miao on 29 September. All three agree to the yuan.

Two pages, one session. Each page states its own date in its own body text. Neither was dated by inference from a feed wrapper or a URL.

One series dates it differently

A day apart. 证券之星 published a 29 September item with Lao Miao at 1,275 and a 30 September item with it at 1,242. That places the 33-yuan fall a day later.

The snapshot is timed. Those items carry pubDate stamps near 02:01 GMT, about 10:00 Beijing time. A counter revised after that hour would not reach that day's item.

The high holds. Against the 25 August high the three counters with a published base remain between 10.71% and 11.73% below, on Carat Capital's divisions. Either dating gives the same result.

What to watch

  • The 1 October boardWhether the counters hold the gain through the National Day holiday week, when wedding buying clusters.
  • Lao Miao's next revisionWhether the brand that stayed flat follows the other three up or the other three come back to it.
  • Any Chinese brand's third-quarter filingWhether same-store figures show the lower counter price converting into volume.

The story so far

  1. Chow Sang Sang read 1,245 yuan on 29 September, 11.6% off the 25 August high
  2. Ten Chinese gold counters priced within seven yuan of each other
  3. Three Shanghai counters priced the same night three different ways

Go deeper

What would change this call

A counter page from the brands themselves. Everything above comes from Chinese financial media reporting what the brand counters showed, not from the four companies' own published boards, and the brands revise those boards on their own schedule without announcing the time. A brand-published daily board, dated and timestamped, would settle the one open question in this article, which is not the levels but the hour: whether a revision after about 10:00 Beijing time explains why one daily series dates the same four numbers to the following session. If the brands' own pages show intraday revision, the dating used here is right and the other series is simply sampling once a day. If they show one price set per calendar day with no revision, then one of the two datings is wrong and it is not yet established which.

Why a counter price is not a spot price

A Chinese brand gold counter quotes a finished-jewellery price per gram, and it is built from three layers. The first is the metal, referenced to the Shanghai Gold Exchange rather than to London, which means a Chinese counter can move on a day when the London screen does not and vice versa. The second is workmanship, the cost of turning a bar into a chain or a bangle, which varies by piece and is where a brand's manufacturing scale shows up. The third is brand margin, set centrally and revised when head office decides rather than when the metal moves. That third layer is why four counters in the same country on the same morning can sit twenty yuan apart on a product that is chemically identical, and why one of them can stay flat on a day the other three move. It is also why a counter price is a poor proxy for the gold price and a good proxy for what a Chinese consumer actually pays. The practical consequence for a trade reader is that the gap between the dearest and the cheapest counter, rather than the level of either, is the number that tells you about pricing power, and on 30 September that gap was 1.6% of the cheapest board, on Carat Capital's division of twenty yuan by 1,242.

Method

Three doors were worked and all three are cited. Huxiu's 30 September item was fetched direct with a browser user agent, 200, 70,953 bytes, and gives the four levels and the four day-on-day moves in one sentence of its own body, dated 9月30日 in that body and not only in the wrapper. Grepped in that saved file before anything was written: 1262 five hits, 1259 five, 1257 four, 1242 four, 9月30日 thirteen, 上涨14元 five, 上涨13元 four, 上涨12元 four and 持平 eight. The 29 September side was re-fetched for this article rather than carried from yesterday's copy: Sina Finance's page carrying National Business Daily's report, 200, 32,318 bytes, timestamped 09.29 13:56 in its own byline and opening 9月29日消息, with 1242 five hits, 1245 seven and 1246 one, and zero hits for 1275. The 25 August base figures of 1,408, 1,410 and 1,407 are this paper's own board published on 27 August and were not re-fetched today. The third door is the 证券之星 daily series, read through its Google News RSS mirror, 200, 101,260 bytes, where the 9月29日 Lao Miao item carries 1275 with a pubDate of Tue, 29 Sep 2026 02:01:51 GMT and the 9月30日 item carries 1242 with a pubDate of Wed, 30 Sep 2026 02:01:37 GMT. That series dates the same numbers one session later than the other two, the divergence is disclosed in the article's own text as a divergence between outside sources, and no attempt is made here to declare one of them wrong. What can be said is narrower and is said instead: the two pages that state their date inside their own body agree with each other and with this paper's 29 September figures, and the series that disagrees samples once a day at about 10:00 Beijing time. All four percentages of the 25 August high, the 20-yuan spread, the 1.6% gap and the four implied 29 September levels are Carat Capital's arithmetic and are labelled where they appear. Archive check, read rather than counted: a case-insensitive search of the built site for Lao Miao, Chow Sang Sang and Lao Feng Xiang returns eighteen files and every article among them was opened. The nearest is 2026-09-30, which covers the 29 SEPTEMBER board and its fall from the 25 August high; this item covers the 30 SEPTEMBER board and a rise. Same brands, different session and the opposite direction, so the verdict is ADVANCED, not repeat. The 27 August piece covers a 25 AUGUST board and is cited here as the base rather than re-reported.

Sources3