For Shengyi Technology (600183), was the drawdown from the October high to November 3 in the same direction as the margin financing balance over the same period, or in the opposite direction? On the after-hours Dragon-Tiger List on November 4, were institutional seats buying or selling?
- The November 3 close was 73.87 yuan, down 47.24% from the high
- Over the same period Shanghai copper rose from 10.33 wan to 10.97 wan, up 6.19%
- Margin funds can explain a small slice of the selling pressure, not that 44%.
For Shengyi Technology (600183), was the drawdown from the October high to November 3 in the same direction as the margin financing balance over the same period, or in the opposite direction? On the after-hours Dragon-Tiger List on November 4, were institutional seats buying or selling?
生益科技(600183)从10月高点到11月3日的回撤，和同期融资余额是同向还是反向？11月4日盘后的龙虎榜上，机构席位是买还是卖？
The desk answersverifiedWalnut, from public filings
Half of the premise doesn't hold: the after-hours dragon-tiger list for November 4 doesn't exist right now—it's just past 2 a.m. on November 4, the A-share market hasn't opened for the 4th yet, and the dragon-tiger list only comes out after the close that evening.
What's available is the November 3 list—the day when the three-day cumulative decline exceeded 20% and the company issued an abnormal-fluctuation announcement early on the 4th. Also, the margin balance for November 3 hasn't been posted yet; the latest margin data only goes to November 2, so the balance comparison only goes to the 2nd.
For the part that can be answered, the conclusion is: the drawdown and the margin balance move in the same direction, both falling; on the dragon-tiger list the institutional seats are selling, with a single-day net sell of 8.55 yi on November 3, and the buyer taking the other side is Shanghai Stock Connect.
Margin funds follow price, they don't lead price
The October high was 140.00 yuan intraday on October 8; the November 3 close was 73.87 yuan, down 47.24% from the high; on a closing-price basis, from 131.86 yuan on October 8 to 73.87 yuan on November 3, down 43.98%. Over the same period the ChiNext Index fell 21.24%, the STAR 50 fell 20.95%, and the Shanghai Composite fell only 5.35%—half of this is the market's discount on growth stocks, not something specific to Shengyi.
The margin balance took the same path: 50.84 yi on September 24, 44.35 yi on October 8, 35.52 yi on November 2, down 19.9% since October 8. Same direction, yes, but the timing is staggered.
The real top wasn't in October, it was 157.50 yuan on September 16. From that day to September 24, the stock fell 11.1%, while the margin balance instead rose from 44.45 yi to 50.84 yi, up 14.4%. Leveraged positions entered while the stock was falling, and this segment later became the fuel for the crash.
The retreat only began after October 8, and accelerated after October 22: from October 22 to November 2, the balance fell 17.6% and the stock fell 31.0%—dragged by price, not dragging price.
The scale doesn't match either. On November 3 full-day turnover was 77.10 yi, and the margin balance was 35.52 yi; one day's turnover could swap out nearly half the balance; on October 30 margin net selling was 2.49 yi, with turnover that day of 59.39 yi, accounting for 4.2%. Margin funds can explain a small slice of the selling pressure, not that 44%.
The sellers are institutions, the buyers are northbound
On November 3 it made two lists. On the single-day list (decline deviation of 7%), institutional dedicated seats bought 1.29 yi and sold 9.84 yi, net selling 8.55 yi; Shanghai Stock Connect bought 12.58 yi and sold 8.13 yi, net buying 4.45 yi. On the three-day cumulative list (October 30 to November 3), the institutional dedicated seats had only four sellers and not a single buyer, net selling 25.78 yi; Shanghai Stock Connect net bought 9.85 yi.
The remaining top-five buyers are brokerage seats: Goldman Sachs Shanghai Century Avenue 4.03 yi, CITIC Shanghai Branch 3.73 yi, Guotai Haitong headquarters 3.16 yi, UBS Garden Stone Bridge Road 2.15 yi.
The company itself is also buying, but an order of magnitude smaller. As of October 30 it had cumulatively repurchased 83.36 wan shares for 7992 wan yuan, at prices of 86.77 to 102.74 yuan. The November 3 close was 73.87 yuan, so every one of those repurchased shares is at a floating loss.
There's another item buried in the announcement. The November 4 abnormal-fluctuation announcement states: the largest shareholder Guangxin Holdings issued 10 yi yuan of exchangeable bonds (26GXKEB1) with the company's stock as the underlying, with the conversion period starting October 8, and "during this period of abnormal stock fluctuation, there were conversion situations."
The October 10 announcement gave the specific numbers—on October 8 alone, 980.59 wan shares were converted, the collateral account fell from 2000 wan shares to 1019.41 wan shares, and Guangxin's combined holding fell from 23.38% to 22.98%. The 980.59 wan shares were 13.4% of that day's turnover of 7291 wan shares.
Conversion doesn't equal selling that day; the announcement didn't say whether they were sold. But this is an extra piece of supply after October 8, and October 8 was the day of the October high.
The bill lands on fourth-quarter gross margin
The chain connects: on October 8 the five stocks in the chain all hit their October highs on the same day—that's external pricing; through the two money channels of institutional dedicated seats and margin funds, the endpoint is the single-quarter gross margin for Q4 2026—not revenue scale, and even less inventory.
The third-quarter report showed single-quarter revenue of 141.82 yi, net profit attributable to parent of 34.40 yi, and gross margin of 41.63%. The calculation is first-three-quarter gross profit of 117.43 yi minus half-year report gross profit of 58.39 yi, giving 59.04 yi, divided by single-quarter revenue of 141.82 yi. Second-quarter single-quarter gross margin was 32.63%, up 9.00 percentage points in one quarter.
Over the same period Shanghai copper rose from 10.33 wan to 10.97 wan, up 6.19%; the company's October 17 pre-announcement said verbatim "facing rapid increases in raw material prices." Costs are rising and gross margin still expanded 9 points—that can only mean selling prices rose faster than raw materials—volume increases would raise both revenue and cost and cannot change gross margin.
| Company | Segment | October high to 11/3 (%) | Q2 gross margin | Q3 gross margin |
|---|---|---|---|---|
| Shengyi Technology | Copper-clad laminate | (47.24) | 32.63% | 41.63% |
| Huazheng New Material | Copper-clad laminate | (43.34) | 18.72% | 36.06% |
| Nanya New Material | Copper-clad laminate | (46.98) | 21.68% | 38.08% |
| Shengyi Electronics | Circuit board | (45.88) | 33.67% | 36.65% |
Shengyi raised the least because its starting point was the highest; Huazheng and Nanya raised 16 to 17 points, a sharp bounce off a low base. Downstream Shengyi Electronics raised only 2.98 points—the price-increase money basically stayed in the copper-clad laminate segment and didn't move to circuit boards.
The third-quarter report's accounts were laid out on October 17: the preliminary forecast for net profit attributable to the parent was 64.6 to 68.7 yi, and the actual figure on October 30 was 67.27 yi, falling in the middle of the range, not exceeding expectations. So what the 44% after October 8 was betting on was not the third quarter but the fourth quarter: the third quarter used prices negotiated from July to September, while the fourth quarter uses the new prices from October to December. Copper prices fell 3.78% from October 8 to November 3, and the cost side actually loosened a little, leaving selling price as the only variable in the fourth quarter.
Under what circumstances would I be wrong: if the fourth-quarter single-quarter gross margin in the 2026 annual report is still above 41%, then this round of decline is only about valuation, and I would be wrong to read "institutions are selling" as "pricing in advance that the fourth-quarter price spread has peaked"; conversely, if the fourth-quarter single-quarter gross margin falls below 35%, then the third quarter's 41.63% is the peak of this cycle.
Sources: Guangdong Shengyi Technology Co., Ltd. "2026 Third Quarter Report" 2026-10-30 (Cninfo 1225644504); "Voluntary Disclosure Announcement of Forecast Increase in Performance for the First Three Quarters of 2026" 2026-10-17 (Announcement No. 2026-057); "Announcement on Changes in Shareholder Equity of Shareholders Holding More Than 5% Reaching the 1% Threshold" 2026-10-10 (Announcement No. 2026-056); "Announcement on Progress of Share Repurchase" 2026-11-03 (Announcement No. 2026-070); "Announcement on Abnormal Fluctuations in Stock Trading" 2026-11-04 (Announcement No. 2026-071); Zhejiang Huazheng New Material Co., Ltd. "2026 Third Quarter Report" 2026-10-23 (Cninfo 1225615477); Nanya New Material "2026 Third Quarter Report" 2026-10-28 (Cninfo 1225630419); Shengyi Electronics "2026 Third Quarter Report" 2026-10-30 (Cninfo 1225645469); Shanghai Stock Exchange Dragon and Tiger List 2026-11-03 (Eastmoney Dragon and Tiger List details verified on the same day); Eastmoney margin financing and securities lending individual stock details through 2026-11-02; Sina Shanghai Copper Continuous (CU0) daily chart through 2026-11-03.
这条的前提不成立了一半：11月4日的盘后龙虎榜现在不存在——现在是11月4日凌晨两点多，A股4日还没开盘，龙虎榜要等当天收盘后晚间才发。
能拿到的是11月3日那一份，也就是三天累计跌幅超20%、公司4日一早发异动公告的那一天。另外11月3日的融资余额也还没挂出来，两融数据最新只到11月2日，余额的对比只到2日。
能答的那部分，结论是：回撤和融资余额同向，都在往下掉；龙虎榜上机构席位是卖的，11月3日单日净卖8.55亿，接盘的是沪股通。
融资盘跟着价格走，不是带着价格走
10月最高价出在10月8日盘中140.00元，11月3日收73.87元，从高点算跌47.24%；按收盘价算，10月8日131.86元到11月3日73.87元，跌43.98%。同期创业板指跌21.24%、科创50跌20.95%，上证只跌5.35%——这里面有一半是大盘给成长股的折价，不是生益一家的事。
融资余额走的是同一条路：9月24日50.84亿，10月8日44.35亿，11月2日35.52亿，自10月8日降19.9%。同向没错，但节奏是错开的。
真正的顶不在10月，在9月16日的157.50元。从那天到9月24日，股价跌了11.1%，融资余额反而从44.45亿加到50.84亿、加了14.4%。杠杆盘是在股价往下走的时候进场的，这一段后来变成了暴跌的燃料。
10月8日之后才开始撤，10月22日之后加速：10月22日到11月2日，余额降17.6%，股价跌31.0%，是被价格拽着走，不是它拽着价格。
规模也对不上。11月3日全天成交77.10亿，融资余额35.52亿，一天成交额换得掉将近一半余额；10月30日融资净卖出2.49亿，当天成交59.39亿，占4.2%。融资盘能解释一小块抛压，解释不了那44%。
卖的是机构，接的是北向
11月3日上了两个榜。单日榜（跌幅偏离7%）上，机构专用席位买1.29亿、卖9.84亿，净卖8.55亿；沪股通买12.58亿、卖8.13亿，净买4.45亿。三日累计榜（10月30日至11月3日）上，机构专用席位只有四个卖方、买方一席没有，净卖25.78亿；沪股通净买9.85亿。
买方前五剩下的是券商席位：高盛上海世纪大道4.03亿、中信上海分公司3.73亿、国泰海通总部3.16亿、瑞银花园石桥路2.15亿。
公司自己也在买，量级差十倍。截至10月30日累计回购83.36万股、7992万元，成交价86.77到102.74元。11月3日收73.87元，这笔回购每一股都浮亏。
还有一条藏在公告里。11月4日异动公告写明：第一大股东广新控股以公司股票为标的发了10亿元可交换债（26GXKEB1），换股期从10月8日起，"在本次股票异常波动期间，存在换股的情况"。
10月10日的公告给了具体数——10月8日当天换股980.59万股，担保专户从2000万股降到1019.41万股，广新合并持股从23.38%降到22.98%。980.59万股是当天成交7291万股的13.4%。
换股不等于当天卖出，公告没说卖没卖。但这是10月8日之后供给端多出来的一块，而10月8日就是10月最高价所在的那天。
账落在四季度的毛利率
链条接得起来：10月8日链上五只票同一天出10月最高价，这是外部定价；经过机构专用席位和融资盘两条钱的路，终点是2026年第四季度的单季毛利率——不是收入规模，更不是存货。
三季报单季营收141.82亿、归母34.40亿、毛利率41.63%。算法是前三季毛利117.43亿减半年报毛利58.39亿，得59.04亿，除以单季营收141.82亿。二季度单季毛利率32.63%，一个季度抬了9.00个百分点。
同期沪铜从10.33万涨到10.97万、涨6.19%，公司10月17日预告里的原话是"面对原材料价格急速上涨"。成本在涨、毛利率还扩了9个点，只能是售价涨得比原料快——销量上升会同时抬高收入和成本，改变不了毛利率。
| 公司 | 环节 | 10月最高至11/3（%） | 二季度毛利率 | 三季度毛利率 |
|---|---|---|---|---|
| 生益科技 | 覆铜板 | (47.24) | 32.63% | 41.63% |
| 华正新材 | 覆铜板 | (43.34) | 18.72% | 36.06% |
| 南亚新材 | 覆铜板 | (46.98) | 21.68% | 38.08% |
| 生益电子 | 线路板 | (45.88) | 33.67% | 36.65% |
生益抬得最少，因为它起点最高；华正、南亚抬了16到17个点，是低基数弹得狠。下游的生益电子只抬了2.98个点——涨价的钱基本上留在覆铜板这一环，没往线路板走。
三季报的账10月17日就摆出来了：预告归母64.6到68.7亿，10月30日实出67.27亿，落在区间中段，没有超预期。所以10月8日之后这44%赌的不是三季度，是四季度：三季度用的是7到9月谈定的价，四季度用的是10到12月的新价。铜价10月8日到11月3日回落3.78%，成本端反而松了一点，四季度的变量只剩售价。
什么情况下我错了：如果2026年报里四季度单季毛利率还在41%以上，那这轮跌的就只是估值，我把"机构在卖"读成"提前反映四季度价差见顶"就错了；反过来，四季度单季毛利率掉到35%以下，三季度的41.63%就是这轮周期的顶。
- 01广东生益科技股份有限公司《2026年第三季度报告》2026-10-30（巨潮1225644504）
- 02《2026年前三季度业绩预增的自愿性披露公告》2026-10-17（公告编号2026-057）
- 03《关于持股5%以上股东权益变动触及1%刻度的提示性公告》2026-10-10（公告编号2026-056）
- 04《关于股份回购进展公告》2026-11-03（公告编号2026-070）
- 05《股票交易异常波动公告》2026-11-04（公告编号2026-071）
- 06浙江华正新材料股份有限公司《2026年第三季度报告》2026-10-23（巨潮1225615477）
- 07南亚新材《2026年第三季度报告》2026-10-28（巨潮1225630419）
- 08生益电子《2026年第三季度报告》2026-10-30（巨潮1225645469）
- 09上海证券交易所龙虎榜2026-11-03（东财龙虎榜明细同日校验）
- 10东方财富融资融券个股明细至2026-11-02
- 11新浪沪铜连续（CU0）日线至2026-11-03