A-Share Pre-Market Brief | 2026-07-28 Tuesday

• Related news: filing 2026-043 on 2026-07-13, H1 pre-announced net profit attributable to parent RMB 3.099–3.298 billion (+117~131%) and ex-non-recurring RMB 2.719–2.918 billion (+97~112%) (CNR²⁷). The interim report is scheduled for 2026-08-15.
• Positive logic (direct): the pre-announcement's own text has twin engines — "the CCL segment … CCL product revenue and gross profit increased" plus "subsidiary Shengyi Electronics benefited from AI computing and high-speed communications". Note that the company's wording is "higher volumes plus mix optimisation", with no mention of "price increases"; price pass-through is only implied within "gross profit increased".
• Fundamentals (primary-verified): 2026Q1 revenue RMB 8.141 billion (+45.09%), net profit attributable to parent RMB 1.158 billion (+105.47%). Single-quarter gross margin went from 21.27% (24Q1) to 28.10% (26Q1), rising for five straight quarters; net margin of 14.23% is a record high, with the expense ratio down 1.6pct over the same period. Working capital has not deteriorated: receivables +28.5% and inventories +18.7%, both slower than revenue +45.1%; inventory turnover went from 114 to 95.6 days. Full-year 2025 operating cash flow / net profit attributable to parent was 1.58x. Gearing is 44.03%, with net interest-bearing debt of about RMB 1.3 billion.
• Structural fact: the 2025 profit surge was driven mainly by PCB (Shengyi Electronics), not CCL — of the segment gross-profit increase, PCB was 58% and CCL only 35%. The structure had already reversed by 2026Q1, with about 75% of the increase in net profit attributable to parent coming from CCL and others.
• Industry position: on Prismark's basis it ranked second globally in rigid CCL sales for 12 consecutive years from 2013 to 2024, with a 13.7% share in 2024. That figure was not obtained from a primary Prismark ranking; it is the company's disclosed basis as relayed by media. There is no reliable data on its share in the ultra-low-loss M7/M8/M9 niches.
• Technical sentiment: +6.13% at RMB 125.99, no limit-up, turnover ratio 2.39%, volume ratio only 0.74. Down 29.4% over the past month and −32.7% from the 6/25 high of RMB 187.20; −6.3% since the strong pre-announcement on 7/13.
• Valuation: PE (TTM) 77.9x; replacing 2025H1 (about RMB 1.428 billion) with the 2026H1 midpoint (about RMB 3.199 billion) gives TTM ≈ RMB 5.7 billion, implying PE ≈ 54x (the first draft wrote "about 60x", an arithmetic error, now corrected).
• Final judgement: priority deep-dive. Its financial quality stands up best to scrutiny. But a strong pre-announcement failed to lift the stock, which shows expectations are already priced ahead of the accounts; and the pre-announcement itself mentions "rising raw-material prices" — what matters is the spread, not the price.

• Related news: filing 2026-055 on 2026-07-15, H1 pre-announced net profit attributable to parent RMB 2.7836–3.1210 billion (+65~85%). Backing out single-quarter Q2 net profit attributable to parent of RMB 1.517–1.854 billion (Q1 was RMB 1.267 billion), accelerating QoQ.
• Positive logic (direct): the segment data are the cleanest evidence of this price cycle — the electronic yarn and electronic cloth segment's gross margin went from 26.9% in 2025Q1 to 52.9% in 2026Q1 (+26pct), while roving only went from 30.2% to 33.9%. That segment contributed 33.1% of gross profit on 24.3% of revenue, delivering RMB 453 million (70%) of the RMB 645 million YoY gross-profit increase.
• Fundamentals (primary-verified): 2026Q1 revenue RMB 5.282 billion (+17.93%), net profit attributable to parent RMB 1.267 billion (+73.48%); single-quarter gross margin 39.64%, net margin 24.88%. Revenue +17.9% produced net profit attributable to parent +73.5%, operating leverage of 4.1x. Gearing 38.93%, inventory turnover from 148 to 98.6 days.
• Two risks that must be flagged:
  1. 2026Q1 operating cash flow / net profit attributable to parent was only 0.17x (1.28x for full-year 2025), with receivables going from RMB 1.891 billion at end-2025 to RMB 3.143 billion (+66%). Whether this is channel stuffing or normal volume growth can only be judged from the interim report (August).
  2. Basis gap: the original text of the company's 2026-05-29 investor-survey minutes reads: "the company currently has no direct business dealings with AI server or optical-module companies"; new capacity is "mainly thin-cloth series products".
• Capacity and cycle: new electronic-cloth plans of 3.2 + 2.5 = 570 million metres (+54% relative to 2025 sales volume), each with a 1.5-year construction period, so output only arrives in 2027H2–2028. The company's own annual report warns: "industry investment in specialty electronic fiberglass projects under construction or planned has now reached a cumulative RMB 13.38 billion, with capacity expected to be released en masse in 2027".
• Technical sentiment: limit-up +9.99% at RMB 41.07 with a RMB 416 million bid queue (cross-verified against the level-one bid), turnover ratio 5.81%. But the intraday high on 6/26 was RMB 77.20 (a market cap of RMB 309.0 billion) and it closed at RMB 37.34 on 7/24 (RMB 149.5 billion), a maximum drawdown of 51.6% — yesterday's limit-up was a bounce off a deep decline.
• Final judgement: priority deep-dive. Understand it as a "fiberglass cyclical", not an "AI computing stock" — the elasticity window is most likely 2026, with a supply shock in 2027.

• Related news: the 2026-07-14 pre-announcement, H1 net profit attributable to parent RMB 2.830–3.000 billion (+68.17~78.28%) and ex-non-recurring RMB 2.730–2.880 billion. Backing out single-quarter Q2 net profit attributable to parent of RMB 1.588–1.758 billion, +7391% YoY and +2842% QoQ, accelerating further from Q1 (+62.9%). The Thai subsidiary turned profitable in single-quarter 2026Q2.
• [Core correction] the main engine is AI networking, not AI servers. From the 2025 annual report by application area (the original text primary-verified):
  • Data communications total RMB 14.656 billion, 77.4% of revenue, +45.21%
    • High-speed network switches and associated routers RMB 8.169 billion, 43.1%, +109.89% ← the fastest-growing and largest
    • AI servers and HPC only RMB 3.006 billion, 15.9%
    • General servers RMB 2.540 billion (13.4%), wireless communications and other RMB 941 million (5.0%)
  • Smart vehicles RMB 3.045 billion (16.1%), +26.41%, but with gross margin −1.61pct YoY → The elasticity from "a large rise in AI-server PCB value content" maps onto less than 16% of WUS's revenue base. (Note: the widely circulated "US$35,100 → US$116,700 per unit / +233%" was verified to be on a Rubin-architecture full rack basis rather than per unit, and no verifiable primary broker report could be found, so this report does not use it as evidence.)
• [Second correction] "price increases" still have zero evidence in WUS's accounts: 2026Q1 gross margin 35.63% vs 2025Q4 35.66%, flat QoQ (the +2.88pct YoY came from product mix and scale effects). The company has issued no price-increase announcement, while its largest single supplier accounts for 22.28% of purchases and the top five for 42.16% — an asymmetric bargaining position.
• Cash flow and balance sheet (needs caution): 2026Q1 operating cash flow of RMB 511 million vs net profit attributable to parent of RMB 1.242 billion, so net-profit-to-cash conversion plunged to 0.41 (1.86 a year earlier); receivables +17.6% QoQ and inventories +15.7% QoQ, both faster than revenue +14.4% QoQ. Capex of RMB 1.467 billion > operating cash flow of RMB 511 million, with the gap covered by net financing inflows of RMB 1.991 billion; single-quarter interest-bearing debt rose about RMB 1.8 billion and gearing went from 46.46% to 48.64%.
• Capacity cadence: the RMB 3.6 billion "high-end PCB expansion for AI chips" project was only 1.51% invested as at end-2025, with trial production expected in the second half of 2026. In other words, essentially all of the 2026H1 high growth came from mix upgrades and utilisation on existing capacity.
• Valuation and position: PE (TTM) 52.6x, in the 96th percentile of the past 5 years and the 90th percentile of the past 10 years; PB 13.5x. −24.2% from the 2026-06-25 high of RMB 154.99 (−25.7% measured against the 120-day high). Note: the first draft's claim that it had "the smallest drawdown among the mainline names" was wrong — Kinwong at −8.7% and Orient Securities at −17.7% are both smaller; that statement has been deleted.
• Technical sentiment: +4.71% at RMB 117.50, no limit-up, turnover ratio 2.40%, volume ratio 0.64.
• Final judgement: watch closely (the first draft listed it as a "priority deep-dive"; downgraded after primary verification). Earnings growth is real and accelerating, and the Thai turnaround is a genuine marginal improvement; but the main engine is misidentified by the market, price pass-through has no financial evidence, cash flow is weakening markedly during the expansion phase, and the valuation is in the 96th percentile of the past 5 years — with all four stacked, the risk-reward is worse than Shengyi Technology and China Jushi.

• Related news: the 2026-07-13 pre-announcement of H1 net profit attributable to parent RMB 2.100–2.300 billion (+54.41~69.12%). Attributed to "rapid AI development, with the company seizing opportunities in the memory market and computing upgrades, plus capacity release at the Guangzhou plant".
• Key note: the company itself lists the "memory market" as its first driver, while the overnight US memory chain is selling off collectively (SanDisk −8%, SK Hynix −5%, Micron −3%), and yesterday the A-share memory chain already "could not be carried along" (GigaDevice −5.03%, Puya Semiconductor −6.23%). This is the only name in the branch whose attribution collides directly with the overnight negative.
• Fundamentals: PE (TTM) 66.24x, PB 14.63x, market cap RMB 240.8 billion. A △ name: segment structure, cash flow and working capital all unverified.
• Technical sentiment: +6.20% at RMB 353.44, no limit-up, turnover ratio only 1.51%; −18.6% over 20 days, −27.1% from the high.
• Final judgement: watch closely.

• Related news: the 2026-07-13 pre-announcement of H1 net profit attributable to parent of RMB 730–820 million, +935.75%~1063.45%; Q1 was RMB 202 million, so single-quarter Q2 backs out to RMB 528–618 million, +161~206% QoQ. The company attributes this to "CCL and prepreg demand exceeding supply, with sales volumes up YoY and selling prices continuing to rise, lifting gross margin" (CLS²⁸, Securities Times²⁹).
• Positive logic (most direct): its attribution wording is the purest of the four — "volumes up + prices up → gross margin up", with no other business diluting it.
• The scoring problem corrected in the rework: the first draft judged its "expectation gap" as "slightly positive" (about 8 pts) on the grounds of a −44.6% oversold move. But the Section 4 model of this report explicitly states "share price falls after a pre-announcement → 0–4 pts", and Shengyi Technology in the same situation was judged "low expectation gap". That was a double standard; it has been rejudged at ≤4 pts, cutting the total from 70 to 64, dropping it below Shennan in the ordering and downgrading it from the Top 5 recommendations to "watch closely (not a priority deep-dive)".
• Fundamentals: PE (TTM) 96.75x, PB 12.45x, market cap RMB 46.4 billion. A △ name, so this item is capped at 12 pts — its cash flow, working capital and segment structure are all unverified, carrying overestimation risk.
• Technical sentiment: limit-up +10.00% at RMB 63.69, turnover ratio 5.53%; −44.6% over 20 days, −48.8% from the high.
• Risks: the growth rate contains an obvious low-base component; as a second-tier maker, whether it can break through on high-end grades such as M9 is undecided; PB 12.45x is no longer cheap.
• Final judgement: watch closely. The highest earnings elasticity and the lowest position in the branch, but also the thinnest verification tier among the top names — until primary verification is complete it should not be treated on a par with Shengyi and Jushi.

• Related news: the interim report was disclosed on 7/28 — revenue RMB 24.7909 billion (+21.35%), net profit attributable to parent RMB 4.7814 billion (+151.06%), ex-non-recurring RMB 4.7853 billion, weighted ROE 18.41%, operating cash flow RMB 7.200 billion (+62.29%).
• Profit quality (the highest on the list): non-recurring items total only −RMB 3.934 million, so ex-non-recurring net profit is higher than net profit attributable to parent; no asset impairment was recognised in H1; OCF / net profit attributable to parent 1.51x, cash-collection ratio 1.09; receivables turnover just 9.4 days. Of the RMB 2.877 billion YoY increase in net profit attributable to parent, 96.5% is recurring operating profit.
• Driver breakdown: the SHFE aluminium 2026H1 average price was RMB 24,325/tonne (+19.96%) and alumina −11.5%. Electrolytic aluminium segment revenue of +20.3% almost exactly matches the aluminium price increase → zero volume growth; 100% is the price scissors gap. Single-quarter gross margin rose to 38.87%.
• Four deduction facts: ① on 7/10 it already pre-announced an "exact figure" of RMB 4.800 billion, and the actual RMB 4.781 billion was 0.4% lower, so this is not a beat, and the stock had already risen 18.7% in the 18 days after the pre-announcement; ② no interim distribution or bonus issue (it did make a Q3 interim distribution in 2024, showing this is not an institutional obstacle); ③ minority interests take 21.3% of profit, and the fastest-growing unit, Yunnan Shenhuo (+214%), happens to be the least owned (58.25%); ④ the SHFE aluminium average price so far in Q3 is RMB 23,035/tonne, already 5.6% below Q2; the 900,000 tonnes in Yunnan (53% of capacity) use grid power, not captive power, and alumina is mainly purchased externally with no hedge.
• Valuation: PE (TTM) 8.28x, PB 2.06x, dividend yield 3.16%. But YTD −5.20% and −19.19% over the past 3 months, so the share price and earnings have fully diverged — the market is pricing a "cycle-peak discount", and a low PE is not evidence of cheapness.
• Final judgement: watch only. Shareholder Shangqiu Putian's reduction window runs to 8/13.

• Related news: the restructuring draft was disclosed at the SSE on 7/28 (not on the evening of 7/27); it plans to issue shares and pay cash to acquire 100% of Shanghai Securities for a consideration of RMB 25.120 billion at an issue price of RMB 10.29/share (the 120-day average, with no discount), issuing 2,288,629,735 shares, equal to 21.22% of the enlarged share capital, with no supporting fundraising.
• [Three popular claims that must be corrected]:
  1. It does not constitute a major asset restructuring — the draft's transaction-nature table explicitly ticks "no".
  2. There is no trading suspension today — the suspension was already executed from 2026-04-20 to 05-06, with trading resuming on 05-07; none of the 33 documents disclosed on 7/28 is a suspension/resumption notice.
  3. There is no earnings-compensation undertaking and no impairment-compensation undertaking (the draft explicitly ticks "no"), while the pro-forma statements show new goodwill of RMB 4,875.4599 million (RMB 4.875 billion) as at 2026-03-31 — that presentation itself indicates the transaction is accounted for as a business combination not under common control (under common control, pooling of interests would generate no goodwill). If Shanghai Securities' future earnings fall short, 100% of the impairment loss is borne by all Orient Securities shareholders.

     The QC question of "whether this might constitute a combination under common control (both parties are Shanghai state-owned)" is reasonable; this report uses the presentation of goodwill in the pro-forma statements as counter-evidence. But this report did not check the draft's common-control tick box word for word, so readers who need to act on it should verify independently.

• Its own earnings (positive): the 2026 interim earnings flash (7/25): revenue RMB 9.560 billion (+19.49%), net profit attributable to parent RMB 4.518 billion (+30.46%), ex-non-recurring RMB 4.418 billion (+30.94%), weighted ROE 5.50%. Backing out single-quarter Q2 net profit attributable to parent of RMB 2.931 billion, +44.6% YoY. The interim report is scheduled for 8/29.
• The true basis of the dilution: BVPS is marginally accretive, 9.98 → 10.05 (+0.70%, because the issue price is above pre-deal BVPS); but EPS is diluted — pro-forma −1.5% for 2025, and pro-forma 2026Q1 goes from 0.19 to 0.16, dilution of 15.8% (the latter is closer to the present); ROE 6.81% → 6.60%; gearing excluding client funds 75.66% → 74.00%.
• Valuation arithmetic: Orient Securities itself is at PB 0.905x and PE (TTM) 11.53x; the target is being bought at 1.2485x PB and 19.0x PE. It is buying a 1.25x PB asset with a 0.91x PB valuation of its own. At market prices the true economic cost is RMB 22.35 billion, implying 1.11x PB and 16.9x PE — still above its own valuation.
• Target quality: Shanghai Securities' 2025 revenue was RMB 3.425 billion with net profit of RMB 1.323 billion (+38.8%); but its 2026Q1 securities trading and investment revenue was −RMB 88 million (a proprietary trading loss), with single-quarter net profit of RMB 151 million and an annualised ROE of only about 3.0%, well below Orient Securities' roughly 7.5% annualised over the same period. The valuation base date of 2026-03-31 is precisely that loss-making proprietary quarter, yet the appraisal premium is 24.85%.
• Deal-failure risk: a whitewash waiver from the Hong Kong SFC plus independent shareholder approval under the Hong Kong Takeovers Code are required; the draft explicitly says "if … cannot be obtained, this transaction will not be implemented". It also needs A/H class shareholder meetings, the Shanghai SASAC, SSE review and CSRC registration.
• Technical sentiment: +2.48% at RMB 9.08, turnover ratio only 0.76% — the coldest liquidity on the list; −17.7% from the high. The share price is 13.3% below the issue price and 2.8% below the pre-suspension RMB 9.34, showing the market has not assigned a sustained positive premium since trading resumed on 5/7.
• Final judgement: watch closely. The event is new, the position is low and it is unpriced, and its own H1 flash of +30.46% is hard earnings; but the certainty of scale expansion far exceeds the certainty of per-share value creation.

• The company's own basis (original text on the Shenzhen exchange interaction platform, 2026-06-23): "the company currently has annualised VC capacity of 3,000 tonnes and annualised VC purification capacity of 12,000 tonnes". The distinction is essential: the 3,000 tonnes is in-house synthesis capacity; the 12,000 tonnes is purification capacity, which requires purchasing crude VC for refining, and crude prices are rising too, so it does not close the raw-material exposure.
• Exposure estimate (this report's estimate, not company disclosure): with 2026 electrolyte shipments of roughly 950,000–1,050,000 tonnes and a VC dosing ratio of 2.5%–3%, annual VC demand is about 24,000–32,000 tonnes, giving a self-supply rate of about 9~12% and net external exposure of about 21,000–29,000 tonnes/year. Every RMB 10,000/tonne rise in VC → +RMB 210–290 million of annualised cost. Per tonne of electrolyte: 3% × RMB 200,000 = RMB 6,000/tonne, which is 21% of the roughly RMB 28,250/tonne selling price of LFP electrolyte (only about 5% on the same basis a year ago).
• The company has already admitted Q2 profit declined (original text of the 2026-07-10 analyst meeting record, No. 2026-003): "in the first half of 2026 the company's electrolyte shipments grew more than 40% YoY, and second-quarter electrolyte shipments rose about 20% QoQ; the company's profit in the second quarter of 2026 declined QoQ, mainly because prices of core raw materials such as lithium carbonate rose in the second quarter while the market average price of lithium hexafluorophosphate fell." → Backing out single-quarter Q2 net profit attributable to parent of RMB 1.046–1.346 billion, down 18.6%~36.8% from Q1's RMB 1.654 billion. Volumes up 20% yet profit down — its profit comes from the spread between the LiPF6 selling price and the lithium carbonate cost, not from shipment volumes, and certainly not from VC.
• The market reaction already confirms it: after the H1 pre-announcement (7/10), the stock fell 7.16% on the next trading day, 7/13. And on 7/27, while Rike Chemical and Sunvim Group hit limit-up, Tinci was only +3.96%.
• Cash flow (the weakest link): 2026Q1 operating cash flow −RMB 21 million vs net profit attributable to parent RMB 1.654 billion, with the roughly RMB 1.68 billion gap sitting in working capital (receivables +RMB 2.7 billion in the quarter). Mitigating: Q1 is seasonal (2025Q1 was also −RMB 122 million), and receivable days improved from 127 to 111.
• What still holds: No. 1 in electrolyte in China (2025 sales above 720,000 tonnes, +44%); self-sufficient in LiPF6 and a net seller of it; gearing 35.56%, close to net cash; PE (TTM) 27.23x, about 19–21x once H1 lands.
• The genuine VC capacity holders (note the boards): 华盛锂电 (Huasheng Lithium Battery, 688353, STAR Market), 泰和科技 (Taihe Technology, 300801, ChiNext), 富祥股份 (Fuxiang Co, 300497, ChiNext), 永太科技 (Yongtai Technology, 002326, Shenzhen main board), 孚日股份 (Sunvim Group, 002083, Shenzhen main board). Only Yongtai Technology and Sunvim Group are main board; but Yongtai is on PE 374.79x and Sunvim's position is extremely hot with a 23.67% turnover ratio and a 3.83 volume ratio yesterday, and this report verified neither company's VC revenue share, so this branch currently has no name that combines right logic + good position + main board + verified.
• Final judgement: watch only (direction corrected). The fundamentals are not bad and the valuation is not expensive, but the logic chain "the VC price increase is positive for Tinci" is wrong, and the QoQ Q2 decline has already been confirmed by the company itself.

It is simultaneously the leader of yesterday's PCB limit-up wave (a RMB 438 million bid queue) and the mainline name with the highest relative strength on the board (only −8.7% from its high and +19.6% over 20 days, while the rest of the branch is generally −30~50%). After primary verification the conclusion is the opposite — but QC flagged two overstatements in the first draft, which are corrected here as well.

• Earnings facts (not in dispute): 2026Q1 revenue RMB 3.892 billion (+16.41%), net profit attributable to parent RMB 233 million (−28.37%), ex-non-recurring RMB 189 million (−18.53%), total profit −36.04%, weighted ROE only 1.76%, operating cash flow RMB 240 million (−51.97%).
• [Correction one] the first draft wrote "five straight quarterly declines in gross margin", contradicting its own data. The actual sequence is 20.78%→21.95%→21.96%→21.56%→18.76%, i.e. the first four quarters oscillated between 21% and 22% and only the latest quarter fell off a cliff, −2.80pct QoQ. The correct statement is "a cliff-edge drop in the latest quarter's gross margin", not "five straight declines". (Net profit, on the other hand, has fallen consecutively: RMB 325 → 325 → 298 → 283 → 233 million.)
• [Correction two] the first draft wrote "the cause of the decline is precisely this price cycle", which imposes causation. Back-calculating from the report's own data: net profit attributable to parent RMB 233 million (−28.37%) → RMB 325.3 million a year earlier; ex-non-recurring RMB 189 million (−18.53%) → RMB 232.0 million a year earlier. Non-recurring items fell from RMB 93.3 million to RMB 44.0 million, a decline of RMB 49.3 million, which is 53.4% of the RMB 92.3 million decline in net profit attributable to parent. That is: more than half of the −28.37% comes from shrinking non-recurring items such as government grants, unrelated to CCL costs; CCL price increases can only explain part of the −RMB 43 million on the ex-non-recurring line, and exchange losses are also mixed into that line. This report cannot isolate the standalone contribution of CCL costs.
• But the core conclusion still holds; only the reasons must change: the genuinely solid negative evidence is not "five straight declines in gross margin" but these three —
  1. ex-non-recurring −18.53% YoY, which is negative growth in a quarter when the whole industry pre-announced positively;
  2. on the evening of 7/13, 15 PCB-chain companies collectively pre-announced positively (an average lower bound of +162.73% net-profit growth), and Kinwong was not on the list; as of today there is still no 2026 interim pre-announcement (the Shanghai main board only mandates one when the YoY change exceeds 50% → implying an H1 change within ±50%), with the interim report scheduled for 2026-08-22;
  3. it is not an AI PCB company: on the Hong Kong prospectus basis (relayed by media), automotive electronics was 45.4% of 2025 revenue, while "communications and data infrastructure" was only 10.4% (2025) / 15.6% (January–April 2026); HDI is only 8.1%, with gross margin down from 20.8% to 12.2%. Its industry position is "No. 1 globally in automotive PCB (10.6% share)", not No. 1 in PCB overall (11th globally, 2.5% share).
• The true boundary of its AI exposure (no fabrication): volume supply of 800G optical-module PCB is genuine, and it has 1.6T mass-production capability; but for AI-server HDI the prospectus wording is "expected to supply a globally leading AI computing infrastructure enterprise, and has obtained that customer's certification" — "obtaining certification" ≠ "volume shipments" ≠ "recognised revenue", and there is no verifiable customer name, order value or supply schedule.
• Valuation: PE (TTM) 71.1x, 82.8x ex-non-recurring, PB 6.07x. Gearing went from 40.25% to 46.77%, long-term borrowings jumped from RMB 284 million to RMB 3.444 billion, free cash flow has been persistently negative, and a Hong Kong IPO is in progress.
• Final judgement: Pass (branch attribution overturned). Yesterday's strength was the capital-vehicle effect of "a low-positioned, large-cap PCB name + a Hong Kong IPO event". The direction (communications revenue +114.6%) is real; the current scale simply cannot support an AI narrative at an RMB 80.9 billion market cap.