| Item | Content |
|---|---|
| Rating | Neutral (Maintained) |
| Target Price | RMB 21–23 (fair value range under base case) |
| Current Price | RMB 21.2 (closing price on 2026-08-31) |
| Margin of Safety | Approx. -1% (base-case fair value floor of RMB 21 vs current price of RMB 21.2) |
| Time Horizon | 6–12 months |
| Valuation Assessment | Reasonable to slightly low |
| Item | Prior (2026-06-30) | Current (2026-08-31) | Reasons for Change |
|---|---|---|---|
| stance | Neutral | Neutral (Maintained) | Although the price has fallen 26% from RMB 28.83 to near the base-case fair value floor, the early-September AD/CVD final ruling is a binary risk with symmetric odds (bull case +39% vs bear case -36%), insufficient to turn positive |
| confidence | 0.62 | 0.60 | Counter-evidence confirms earnings quality deserves a discount (overseas gross margin below domestic, minority shareholders absorbing losses, tax benefits); the normalization anchor faces downward revision risk |
| Target price range | RMB 24–28 | RMB 21–23 | PV gross margin at its lowest since 2021 (11.18%) + certainty of tax erosion + normalization anchor switched from 2026E forward to 2027E adjusted non-recurring of approx. RMB 2.0bn × 15-16x |
| judgment | Reasonable to slightly overvalued | Reasonable to slightly low | Price has already digested H1 adjusted non-recurring decline of -24.3% and FX impact; implied expectations are roughly in line with consensus, not stretched |
Old catalysts settled: ① "US CVD final ruling (expected 2026-07)" — postponed, final ruling delayed to around 2026-09-03; ② "share price pullback below RMB 24" — fulfilled (current price RMB 21.2); ③ "interim report to verify PV gross margin and FX convergence" — partially fulfilled (PV gross margin of 11.18% below the 13%+ watch line; FX loss of RMB 349mn not converging).
Hengdian DMEGC is a scarce, scale-profitable name in the PV downcycle (in 2025 the four major module leaders collectively lost approx. RMB 25bn, while the company posted net profit attributable to parent of +RMB 1.851bn), with its global magnetics leadership and small cylindrical lithium batteries providing an earnings cushion. The core of the -24.3% H1 2026 adjusted non-recurring decline is the RMB 349mn FX loss and PV gross margin falling to the cycle bottom, not demand collapse; at the current price of RMB 21.2, the stock has retreated to the lower edge of the base-case fair value range, implying normalized earnings of approx. RMB 1.95bn, roughly in line with sell-side consensus. Three constraints on turning positive: the AD/CVD final ruling (early September) carries retroactive tax tail risk with no specific provision seen in the statements; earnings quality deserves a discount due to tax and electricity-price benefits; and an 84.76% payout ratio is unsustainable with near-zero H1 free cash flow. Maintain Neutral; the AD/CVD final ruling outcome and FX convergence in the Q3 report are the right-side signals.
Upcoming changes: Three verifiable trajectories — ① PV gross margin recovering from 11.18% (lowest in the range since 2021) toward 13-15% (anti-involution production curbs + module prices raised 30-50% from the RMB 0.693/W low; price-sustainability verification window 2026-09/10); ② FX drag converging (H1 loss of RMB 349mn; if the RMB turns to two-way fluctuation, H2 loss converging to RMB 100-200mn); ③ devices revenue ramping from RMB 806mn in H1 (+86.5%) toward RMB 1.6-1.8bn for the full year (AI inductors in overseas customer validation/qualification).
Primary share price driver: The US final AD/CVD duty rates on Indonesia (scheduled around 2026-09-03) — determining the viability of direct-to-US and transshipment pathways for 60%-owned PT NUSA (3.8GW cells + 2GW modules), a two-way catalyst: if rates are significantly below the preliminary combined level (121%-178%), the negative is fully priced out; if high rates are maintained, the annualized profit impact is approx. RMB 100-700mn (midpoint approx. RMB 300mn), triggering impairment at the Indonesia base and retroactive tax clawback (emergency situations can be retroactive to 2025-11/2026-01; the RMB 451mn provision for liabilities in the interim report is only warranty-related, with no AD/CVD-specific provision seen).
Verifiable expectation gap: Limited. The current price implies normalized net profit attributable to parent of approx. RMB 1.95bn, ≈ consensus 2026E (RMB 1.94bn), about 7% above the 2023-25 adjusted non-recurring average (RMB 1.82bn) — market expectations are broadly reasonable, slightly optimistic, with no significant mispricing to arbitrage; excess returns depend on whether "anti-involution implementation pace + AI devices volume ramp" can push earnings onto the consensus 2027E path of RMB 2.26bn. Caveat: consensus is based on only 7 covering analysts, 3 target prices, 100% Buy/Overweight ratings, and a 5.9% upward revision over the past 90 days — in a crowded trade, any downward revision triggered by the final ruling or flattening Q3 PV gross margins would be non-linearly damaging.
Verification catalysts and falsification conditions: See the tracking section at the end of the report.
Key evidence
The high growth in the devices sub-segment includes non-AI categories, and the AI inductor share is undisclosed; large overseas customers remain in validation/small-batch stages — the "highest certainty" characterization should be downgraded to "relatively high certainty." Magnetic materials gross margin fell 1.82pp YoY to 25.89% (revenue +12.44%, cost +15.26%); "volume and profit both up" does not hold on the materials side. R&D expense +59.9% (RMB 364mn) is the current cost of AI investment, with returns unproven. Management explicitly stated for SST that "a substantive scale is difficult to achieve in 2026" and declined to give AI magnetics revenue guidance — this non-bandwagoning stance actually enhances credibility, but it also means that of the current price, approx. RMB 5.6/share of growth premium (approx. 26% of the current price) is partly built on this option and should be viewed at a discount.
Key evidence
Triple discount on earnings quality: ① overseas gross margin is 6.5pp below domestic, contradicting the differentiation premium narrative; ② earnings include Yibin low-cost hydropower and the 15% Western Development tax rate benefit (Sichuan DMEGC H1 net profit of RMB 290mn, approx. 30% of attributable profit), with the effective tax rate falling from 17.8% to 11.1%; ③ attributable resilience partly relies on NUSA's 40% minority absorbing losses (P&L swing of RMB 367mn) — if the final ruling is unfavorable, impairments/clawbacks will flow back to attributable profit per majority interest. On anti-involution, management's language is conditional ("if industry supply-demand dynamics and competitive order further improve, it may… have a positive impact"), and domestic new installations fell 66% YoY in 2026H1 — demand collapse could offset curtailment-driven price increases. The US AD/CVD final ruling (early September) + retroactive duties is the biggest single near-term risk.
Key evidence
The current price implies normalized attributable profit of approx. RMB 1.95bn (15x normalized PE + net cash back-solved), in line with consensus 2026E. But the normalization anchor (2023-25 adjusted non-recurring average of RMB 1.82bn) includes 2023's elevated PV gross margin of 20.8%, and the current 11.1% effective tax rate is below the historical 17.8% — if normalized at historical tax rates, the adjusted non-recurring midpoint should be below RMB 1.82bn, making the conclusion that "current-price implied expectations are broadly reasonable" somewhat loose. PB of 3.25x sits at the 52nd percentile over 5 years; the "96th percentile" ROE of 17.9% is weak evidence relative to an industry-wide loss-making cohort (company ROE has declined for four consecutive years from 23.1% in 2022).
Key evidence
Excluding the RMB 130mn hedging gain, the pre-tax economic impact of FX in H1 was approx. RMB 450mn (rather than the nominal swing of RMB 600mn). Of the triple tax hit, the export rebate abolition is already reflected in the H1 PV gross margin -5.52pp; the battery consumption tax and small-power rebate abolition represent certain erosion for 2026H2-2027, totaling approx. RMB 200-300mn annualized (self-estimated in this report based on lithium battery revenue and export share; no quantified company guidance), equivalent to 11-16% of 2025 attributable profit. FX is the single largest uncontrollable variable: if the RMB appreciates another 2-3% in H2, full-year FX losses could reach the RMB 700mn scale, and a single quarter's profit could be flipped from positive to negative by FX alone.
Key evidence
Unrestricted cash of RMB 2.253bn is below short-term borrowings of RMB 2.788bn — the "no liquidity risk" characterization needs softening: short-term borrowings are mainly bill discounting financing matched to notes payable of RMB 4.789bn in the settlement system, an operating financing structure rather than debt-servicing pressure; but with the trough lasting over a year and the current dividend intensity maintained (H1 dividend-related outflows approx. RMB 1.33bn, 23x FCF), free cash flow turns negative and the payout ratio must be cut. OCF/attributable profit plunged from 2.15 in FY2023 to 0.77 (company explanation: mismatch between deposit terms of guarantees and business validity periods + VAT credit refund base — a reversible item); Q3 guarantee-deposit inflows and receivables drawdown (receivables +20.7% far outpacing revenue +3.4%) are the verification points for cash flow authenticity. Inventory of RMB 6.417bn (unverified) (finished goods provision ratio only approx. 4%; PV plants of RMB 1.455bn with no impairment provision) would require approx. RMB 400mn of additional write-downs in a scenario where module prices fall another 10%.
| Metric | FY2023 | FY2024 | FY2025 | 2026H1 |
|---|---|---|---|---|
| Revenue (RMB 100mn) | — | 185.6 | 225.86 | 123.42 (+3.41%) |
| Net profit attributable to parent (RMB 100mn) | 18.18 | 18.27 | 18.51 | 9.48 (-7.03%) |
| Adjusted non-recurring net profit (RMB 100mn) | 20.25 | 16.72 | 17.61 | 8.11 (-24.28%) |
| Gross margin | — | — | — | 15.7% (-2.42pp) |
| Net margin attributable to parent | — | 9.8% | 8.2% | 7.7% |
| Operating cash flow (RMB 100mn) | 39.00 | 35.22 | 32.58 | 7.29 (-57.12%) |
| Free cash flow (RMB 100mn) | — | — | — | 0.57 |
| Cash (period-end, RMB 100mn) | — | — | — | 88.89 (restricted 35.72) |
| Interest-bearing debt (period-end, RMB 100mn) | — | — | — | 37.12 (short-term 27.88 + leases 9.24) |
| Debt-to-asset ratio (period-end) | — | — | — | 61.43% |
| Net cash (period-end, RMB 100mn) | — | — | — | 51.77 |
Note: FY2024 revenue back-solved from the +21.70% YoY disclosed in FY2025; FY2023 revenue and some annual gross margins were not within this data extraction scope and are marked "—". Balance-sheet items are period-end figures from the 2026-06-30 interim report.
Reasons for metric changes (items with YoY ≥±20%, all cited from the interim report):
H1 revenue RMB 12.342bn (+3.41%), attributable profit RMB 948mn (-7.03%), adjusted non-recurring RMB 811mn (-24.28%). Q2 standalone revenue RMB 6.863bn (+2.23% YoY, +25.26% QoQ), attributable profit RMB 562mn (+0.09% YoY, +45.76% QoQ), adjusted non-recurring RMB 494mn (-19.58% YoY, +55.52% QoQ), standalone quarterly gross margin 14.56% (-2.57pp QoQ). H1 attributable profit equals 49% of consensus 2026E (RMB 1.94bn) — a neutral pace; but the adjusted non-recurring figure is clearly weaker than attributable, with the RMB 138mn gap mainly being non-recurring FX hedging gains. Three factors in the profit decline: FX loss of RMB 349mn (largest single item), PV gross margin -5.52pp, export rebate abolition. The large QoQ increase in Q2 attributable profit suggests early signs of an earnings bottom, but gross margin is still falling sequentially — the reversal is unconfirmed; the "earnings bottom is behind us" claim requires dual verification from Q3 financial expense convergence and sequentially flat PV gross margin.
Business model: Asset-heavy manufacturing across three main businesses — magnetics & devices (global ferrite leader, integrated from presintered powder to materials to devices, with automotive/AIDC customer certification barriers), integrated PV cells and modules (23GW cells + 21GW modules, differentiated-scenario modules focused on European distributed), small cylindrical lithium batteries (global No. 2 in the 18650 format, focused on small power). Revenue is non-recurring in nature; export share is 47.11%. Pricing power is bifurcated: high-end automotive magnetics products and differentiated PV products command premiums, but PV industry overcapacity constrains overall pricing power.
Cash content of earnings test: OCF/attributable profit for FY2023-2025 was 2.15→1.93→1.76, consistently above 1 with excellent historical cash conversion; but it plunged to 0.77 in 2026H1 and FCF/attributable profit was only 0.06 — net working capital increase of RMB 793mn (receivables +RMB 726mn, inventory +RMB 544mn) exceeded the revenue increase (RMB 407mn), indicating heavier capital tied up. The company attributes this to timing mismatch of deposits and the tax-refund base (reversible items); Q3 is the verification point; if no inflow by Q3, it should be re-assessed as structural deterioration.
Recurring earnings test: The H1 gap between adjusted non-recurring of RMB 811mn and attributable profit of RMB 948mn is RMB 138mn (-14.5%), mainly the RMB 129mn FX hedging gain classified as non-recurring — unusually, in direction this is "non-recurring gains inflating attributable profit," i.e., true operating profit is weaker than headline attributable profit. The company's self-reported Non-GAAP basis has no aggressive SBC add-back issue (all R&D expensed).
Return on capital: Annualized 2026H1 ROIC of approx. 6.0% (net profit RMB 875mn / (net assets RMB 10.949bn + interest-bearing debt RMB 3.712bn)), below a 9% WACC — asset-heavy returns are low during the industry downturn; in normal historical years, ROE of 17-23% was significantly above the cost of capital, and the current level reflects a cycle bottom rather than structural deterioration; however, vigilance is needed that continued PV capacity investment during the trough further dilutes returns.
Maintenance capex: H1 CapEx/D&A = 6.73/4.41 = 1.53, at the edge of the ">1.5 warning line" — PV plants and magnetics bases are still in investment phases, leaving limited distributable cash for shareholders; if PV capex contracts after anti-involution clears excess capacity, this ratio should fall.
Red flags: ① Accounts receivable +20.7% far outpacing revenue +3.4% (no specific company explanation); ② Of inventory of approx. RMB 6.4bn (unverified), PV plants of RMB 1.455bn carry no impairment provision and finished goods provision ratio is only approx. 4%; ③ R5-rated trust wealth-management products of RMB 286mn overdue and unrecovered (fully loss-provisioned in 2023, reflecting weaknesses in treasury risk controls); ④ Controlling shareholder has pledged 210mn shares (25.8% of its holding).
Say-Do Consistency: Partially delivered, pragmatic messaging. ① The 2024 target of the third ESOP (net profit attributable to parent +20% vs 2022) was actually completed at 93.12%, and the 2025 target (+30%) at 89.31%, both unlocking proportionally — two consecutive years of missing full targets, but by small margins; ② Cautious messaging on AI/SST themes: explicitly stated that "SST is unlikely to form material scale in 2026" and "that segment's revenue share is not high," declining to give AI magnetic materials revenue guidance — no hype during the AI thematic frenzy (after consecutive limit-ups in late May, the stock pulled back); a positive; ③ The stance on anti-involution is conditional rather than a commitment, consistent with the company's一贯 cautious style. Overall judgment: pragmatic, prefers conservative disclosure.
Shareholder Friendliness: High, but current intensity is unsustainable. FY2025 cumulative dividend of RMB 1.569 billion, payout ratio of 84.76% (significantly above the 2023-25 average of ~53%); an additional RMB 368 million interim dividend in 2026; cumulative dividends since listing of RMB 5.328 billion, 39.37% of cumulative net profit attributable to parent; no dilution via share issuance/rights issues/convertible bonds; buyback account holds 25,375,600 shares (sourced from ESOP transfers and incentive buybacks, not cancellation). H1 dividend-related cash outflow of ~RMB 1.33 billion vs FCF of RMB 57 million — the 84.76% payout ratio is built on existing cash; if the trough persists, the ratio needs to revert to a ~50% center.
Risk signals: ① Controlling shareholder Hengdian Group sold 9.644 million shares on 2026-02-11 (0.59% of total share capital) and pledged 210 million shares (25.8% of its holdings, 12.91% of total share capital) — the major shareholder is a net seller; ② R5 trust of RMB 286 million overdue (loss recognized); ③ Financial assistance provided to overseas controlled subsidiaries (interest rate 3%); related-party transactions are small in amount and subject to quota controls, with no abnormal fund occupation.
| Segment | Revenue share | Gross margin | YoY | Business logic |
|---|---|---|---|---|
| PV products | 59.71% | 11.18% (-5.52pp) | -8.50% | Integrated cells & modules, differentiated-scenario modules + overseas high-value markets, shipments over 12GW |
| Magnetic materials | 17.66% | 25.89% (-1.82pp) | +12.44% | Global leader in hard & soft ferrites, driven by automotive-grade high-end products |
| Lithium batteries | 14.16% | 20.09% (+7.19pp) | +35.91% | Small cylindrical focused on small-power applications, shipments over 300 million units, full-tab pilot line in production |
| Devices | 6.53% | 20.68% (estimated) | +86.50% | Chip inductors/EMC filters etc., main growth engine for AI-server-related volume |
| Others | 1.94% | 13.15% (estimated) | +6.48% | Miscellaneous business |
Main profit drivers: Estimating gross profit contribution by "revenue share × gross margin," PV contributes ~RMB 668 million (42.5% share), still the largest source of gross profit; magnetic materials + devices combined at ~RMB 592 million (37.7%) follow closely; lithium battery RMB 284 million (18.1%). Gross margin structure shows a gap of 14.7pp between the highest (magnetic materials 25.89%) and the lowest (PV 11.18%) — essentially the pricing-power gap between a certification-barrier business (magnetic materials: long automotive-grade/AIDC certification cycles, integrated cost advantage) and an overcapacity business (PV: industry-wide price war). Marginal changes are notable: while PV gross margin fell 5.52pp YoY, lithium battery rose 7.19pp and devices held above 20% — the profit structure is migrating from "PV dominance" toward "balanced three engines," a structural gift of this cycle to the company.
Accounting red flags:
Cross-period consistency:
No signs of profit-manipulation-style financial engineering found; the main warnings lie in asset quality (inventory/receivables) and cash flow timing, not revenue recognition or expense capitalization.
Current market data and multiple percentiles: Market cap RMB 34.486 billion (~1.627 billion shares); PE(TTM) 19.38x (74th percentile over 1 year / 83rd over 3 years / 61st over 5 years); PB 3.25x (52nd percentile over 5 years); dividend yield ~2.8% (based on FY2025 final dividend). Percentiles for reference only: PV is at a cyclical bottom, and H1 included RMB 349 million of FX losses, so earnings are unrepresentative and the PE percentile is mechanically elevated; for cyclicals, PB + normalized earnings should be the primary anchor. Peer comparison: SW PV cells & modules median PE 34.0x (peers broadly loss-making, median distorted), PB 2.6x, median ROE -23.9% — the company's "lowest PE + highest ROE" combination reflects a scarcity premium for profitability amid industry-wide earnings distortion.
| Company | 2025 revenue | Revenue YoY | PV gross margin | ROE | Module shipments/rank |
|---|---|---|---|---|---|
| Hengdian DMEGC | RMB 22.586bn | +21.70% | 15.25% (2025) / 11.18% (2026H1) | 17.42% | 24.9GW / Global No. 8 (~4.6%) |
| JinkoSolar | RMB 65.492bn | -29.18% | -0.82% | Negative (loss of RMB 6.786bn) | 86.81GW / No. 1 |
| LONGi Green Energy | RMB 70.347bn | -14.82% | N/A | -11.15% | 86.58GW / No. 1 |
| Trina Solar | RMB 66.975bn | -16.61% | N/A | Negative (loss of RMB 7.031bn) | Over 67GW / No. 3 |
| JA Solar | RMB 49.129bn | ~-30% | N/A | Negative (cumulative loss over RMB 9bn) | 66.53GW / No. 3 |
Market-implied expectations: Current price RMB 21.2 = 19.0x normalized EPS (RMB 1.12, net-cash-inclusive basis; 16.1x after excluding net cash of RMB 3.18 per share). Back-solving at 15x normalized PE + net cash, the current price implies the market believes normalized net profit attributable is ~RMB 1.95 billion — about 7% above the 2023-25 adjusted average of RMB 1.82 billion, ≈ consensus 2026E (RMB 1.94 billion), and implies a path of growth toward RMB 2.2-2.3 billion in 2027. The reality is that 2026 adjusted profit is likely flat or slightly down (H1 -24.3%, mainly due to one-off FX losses); the implied expectation holds only if "anti-involution price repair + magnetic materials/lithium battery offset" materializes in 2026H2-2027 — the expectation is broadly reasonable, slightly optimistic, not yet overextended, but no bargain either.
Three-layer value (EPV): Asset value (liquidation floor) ~RMB 6.52/share; EPV with zero growth ~RMB 15.6/share (normalized EPS RMB 1.12 ÷ WACC 9% + net cash per share RMB 3.18; under sensitivity ranges of WACC 8-10% and normalized EPS 1.05-1.12, the range is RMB 14.5-16.8); current price exceeds EPV by ~RMB 5.6/share (~26% of current price), a premium paid for anti-involution repair + AI devices + lithium battery growth + long-term SST option. The current price is mainly supported by the EPV layer (~74%), with the growth option at ~26% — a fairly robust structure for a company that remains consistently profitable amid massive industry-wide losses and holds RMB 5.18 billion net cash; however, note that normalized EPS of RMB 1.12 includes the high-margin year of 2023 and higher-tax years, so the zero-growth floor carries downside risk (at RMB 1.05, EPV would be ~RMB 14.9).
Three scenarios and odds:
| Scenario | Probability | Fair range | Key drivers |
|---|---|---|---|
| Bear | 30% | RMB 12-15 (-43% to -29% vs current) | Anti-involution falsified + high final AD/CVD tariffs (including retroactive tax and Indonesia base impairment) + export tax rebate cancellation; PV gross margin stuck at 9-11%; 2027 adjusted profit falls to RMB 1.3-1.45bn; trough EPS RMB 0.85-0.90 × 10-12x + net cash |
| Base | 50% | RMB 21-23 (-1% to +8% vs current) | PV gross margin repairs to 14-15% (back to 2025 levels, not the 2023 peak) + magnetic materials & devices/lithium battery growth offset; 2027 adjusted profit ~RMB 2.0bn × 15-16x + net cash RMB 3.18/share |
| Bull | 20% | RMB 28-31 (+32% to +46% vs current) | Anti-involution substantively effective (supply clearing + higher price center); PV gross margin back to 17-18%, Indonesia capacity exemptions ramping; 2027 adjusted profit ~RMB 2.5bn × 17-18x + net cash |
Probability-weighted expected value ~RMB 20.95, essentially flat with the current price (~-1%); the current price sits between the bear-case upper bound and the base range — odds roughly symmetric, slightly skewed down. Anchoring rationale: the base scenario's 15-16x exit multiple is below the company's 5-year median PE of 17.3x (a cyclical-center discount), cross-validated with PB-ROE (normalized ROE 16-18% × PB 2.6-3.2x); the bull case's 17-18x is anchored on the company's own trading range during its 2024 earnings delivery period and the upward revision path of the consensus target of RMB 26; the bear case's 10-12x is an extreme-trough floor (trough EPS RMB 0.89 × 10x + net cash ≈ RMB 12.1). The bear case already covers the market's most pessimistic public narrative (Gantan Hao, 2026-08-20: end of the Indonesia arbitrage model + NUSA profit at zero + retroactive taxes + impairments — that article gave no target price; its implied profit-impact magnitude corresponds to this bear case's RMB 12-15 range); the base range sits between the sell-side target (RMB 26) and the most pessimistic narrative, not extending beyond either end of the market distribution.
Own earnings forecast (presented alongside management guidance and sell-side consensus):
| Period | Revenue | Net profit attributable | Key assumptions |
|---|---|---|---|
| FY2026E | RMB 22.8-23.8bn (+1%~+5%) | RMB 1.75-1.95bn (adjusted RMB 1.6-1.8bn) | Module shipments ~25GW, flat to slightly up; full-year PV gross margin 11-13%; FX assumption: no repeat of a RMB 349mn-scale loss in H2; consensus of RMB 1.94bn sits at the top of our range |
| FY2027E | RMB 24.0-25.5bn (+5%~+8%) | RMB 1.95-2.25bn (adjusted RMB 1.85-2.15bn) | If anti-involution delivers, PV gross margin repairs to 14-15%; magnetic materials & devices +10-15%, lithium battery +25-35%; drags: small-power tax rebate cancelled in 2027 + PV cell consumption tax of 2% from April 2027 |
Conclusion: Reasonably valued, slightly cheap on fundamentals — excellent quality (the only player in the industry with scale profitability + net cash + high dividends) but the price has only just returned to the lower edge of fair value, with no margin of safety. Target price RMB 21-23 (base scenario). Trading implication: no need to chase at current levels, no need to panic either; the final AD/CVD ruling landing (clearing the negative or confirming it) and Q3 FX loss convergence are right-side add-on signals; if the stock rallies above RMB 23 without confirmation signals, the margin of safety is exhausted.
Industry size: ① PV — global module shipments of ~536GW in 2025 (InfoLink, 12 ranked companies); China module output ~514GW (+13.5%); but 2026 is an inflection year: InfoLink expects module demand to be revised down to 529-624GWdc — the first negative growth in a decade; BNEF expects 649GW of global new installations in 2026 (first decline in twenty years); CPIA expects China's new installations to fall back to 180-240GW (Jan-Feb 2026 YoY -17.71%, first negative since grid parity; 2026H1 domestic new installations -66% YoY). ② Magnetic materials — China's total magnetic materials output of ~1.62 million tonnes in 2025 (hard ferrite 750k tonnes, soft magnetic 600k tonnes, rare-earth permanent magnets 270k tonnes, CECA); global soft magnetic materials market ~USD 34.1 billion in 2024, expected USD 62.6 billion by 2033 (IMARC, third-hand source, to be verified), ~7% CAGR going forward; AI compute is a structural increment — per IDC, China's AI compute market was USD 19 billion in 2024 (+86.9%), expected USD 25.9 billion in 2025. ③ Small cylindrical lithium batteries — global cylindrical battery shipments of ~14.2 billion units in 2025, expected 18.98 billion by 2030 (third-hand source, to be verified); China's electric two-wheeler lithium battery shipments of 20.5 million packs (+26.5%, Qidian Research).
Value chain and value distribution: PV upstream: polysilicon → wafers → cells → modules; in 2025 polysilicon average price rose 38.9% from the start of the year, and the surge in silver pushed non-silicon costs above 30% of total — upstream is clearly squeezing the module segment; the company sits in midstream integration (cells + modules, no wafers or polysilicon), with weak bargaining power over upstream, offset via supply-chain synergy and cost control, and also weak bargaining power over downstream power plants (poor price pass-through, end-market wait-and-see). Magnetic materials upstream is bulk oxide feedstock such as iron red/iron phosphorus (the company self-produces prefired materials, vertically integrated); downstream is EV Tier 1s/AIDC power supply makers/appliance OEMs — value mainly accrues to midstream materials and devices (the company's magnetic materials gross margin of 25.89% is the highest of its three businesses). Lithium battery downstream is two-wheeler OEMs (Yadea/Aima etc.), with relatively weak bargaining power.
Supply-demand and competitive landscape: PV supply is severely oversupplied — in December 2025, ~95% of companies signed production- and price-limit agreements (self-discipline prices: polysilicon RMB 51,000/tonne, modules RMB 0.735/W); the MIIT is promoting orderly exit of outdated capacity in 2026, and mandatory national standards on energy consumption and efficiency effective July 2 accelerated clearing; module prices have risen 30-50% from the December 2025 low of RMB 0.693/W, but this is a "policy-supported, cost-push repair" rather than demand-driven; whether prices can hold at RMB 0.68-0.70/W in Sep-Oct is key. Concentration: module CR4 ~58% (Jinko 86.81GW, LONGi 86.58GW, Trina over 67GW, JA 66.53GW); the company has 24.9GW, ~4.6% share, tied for eighth with Canadian Solar; the magnetic materials industry is fragmented, with the company's domestic hard ferrite share above 20% ranking first; high-end high-frequency materials are dominated by TDK/Proterial with a structural shortage; small cylindrical is top-concentrated (EVE/Hengdian DMEGC/Samsung SDI as top three), and BYD's entry into two-wheeler batteries may reshape the landscape. Entry barriers: PV — scale + technology iteration + overseas channels, with slow existing-capacity clearing; magnetic materials — integrated know-how + automotive-grade/AIDC certification cycles; lithium batteries — safety certification and customer lock-in. Substitution threats: rare-earth permanent magnets replacing ferrites in high-end motors (the company responds with new permanent magnet materials R&D), metal soft magnetic powder cores replacing ferrites at high frequencies (the company is also positioned here), and sodium batteries replacing lithium batteries in two-wheelers.
Regulation: ① Anti-involution enters a critical phase — production/price limits + national standards + capacity exit, favoring leading integrated players; ② PV export tax rebates cancelled from April 2026 + phased introduction of a cell consumption tax — export profitability center systematically shifting down; ③ US trade barriers fully blocking Southeast Asia's channel to the US (April 2025 final AD/CVD rates on four Southeast Asian countries exceeding 3400% at the high end, 2026 AD/CVD proceedings against Indonesia/Laos etc., UFLPA scrutiny), with global supply chains restructuring toward US domestic/India/Middle East — the company's PV business is mainly in Europe and other markets, with direct US exposure smaller than the top four, but its Indonesia base is in the crosshairs; ④ 3C certification/whitelist/mandatory national standards for two-wheeler lithium batteries rolled out, favoring compliant leaders; ⑤ Consumer trade-in programs boosting appliance/two-wheeler demand.
Company's industry position: "One lead, two chases" — magnetic materials: absolute leadership (domestic hard ferrite share above 20%, global No. 1 in scale; 300k-tonne capacity + prefired material integration forms a dual moat of cost + certification; AI/AIDC brings structural increments, chip inductor shipments doubling); PV: the top student among the chasers (global No. 8, ~4.6% share; deep in European distributed via differentiated-scenario modules, 2025 shipments +45%, at full capacity and full sales while maintaining positive gross margin — the four major leaders' module gross margins turned negative over the same period, with combined losses of ~RMB 25 billion; the company's overall ROE of 17.42% is a standout in the industry; share trending up, but via niche markets rather than mainstream ground-mounted plants); lithium batteries: global No. 2 in small cylindrical (18650 basis, behind only EVE), benefiting from two-wheeler lithium adoption and standard-driven clearing, facing downward-pressing competition from BYD. Moat sources: magnetic materials scale and integrated cost, differentiated product positioning, globalized capacity layout (Yibin low-cost hydropower + multiple bases in Indonesia/Vietnam/Thailand).
Cyclical positioning: Using the company's PV gross margin as the cyclical coordinate—12.5% in December 2021 (recovery period post-531) → 14.86% in 2022 → 20.81% in 2023 (peak) → 18.74% in 2024 → 15.25% in 2025 → 11.18% in 2026H1 (the lowest point of the range, 0th percentile), with a range midpoint of approximately 15–16%. This cycle has been rising from 2021–2023 and declining from 2024 to now for roughly two and a half years; the company is currently at a "profit trough of stable volume but falling prices": shipments are at full capacity and full sales (over 12GW in H1) but prices are near cost lines. Leading indicators: ① whether module prices can hold at 0.68–0.70 CNY/W (after mid-August silicon material rose +23.81% in a single day and wafers +40% in a week, end-market uptake remains in doubt); ② production restriction compliance rate and progress of capacity exits under national standards; ③ monthly domestic installations (H1 YoY -66% is the biggest demand-side warning); ④ US AD/CVD final determination.
Supply response: Under the triple constraints of industry self-discipline production cuts (95% of companies signed), national standards on energy consumption and efficiency (to take effect 2026-07), and elimination of export tax rebates, 2026 domestic cell/module output will fall YoY -21.9%/-35.1% (CPIA), accelerating exit of tail-end capacity; but demand will decline for the first time in a decade in 2026 (529–624GWdc), meaning a race between "supply contraction vs demand contraction" will limit the height of price recovery—this is the biggest difference from the 2012–2013 shakeout: demand is no longer growing; recovery relies on margin normalization rather than volume growth.
Through-cycle earnings: Normalized (mid-cycle) EPS of 1.12 CNY (2023–25 recurring net profit average basis, corresponding to a PV gross margin midpoint of ~15% + magnetic materials 28% gross margin + lithium battery recovery; note: includes an elevated 2023 and a higher-tax-rate year, with downside risk of revision to around 1.05 CNY); trough EPS of 0.85–0.90 CNY (bear case: PV gross margin stuck at 9–11% + tax erosion). The current P/E is in "trough-high P/E" form—the apparent TTM 19.4x multiple is built on distorted earnings; strictly prohibited to use the current forward P/E as the sole valuation anchor. Normalized P/E of 15x ±5x corresponds to a price range of 14.4–25.6 CNY (including net cash of 3.18 CNY); the current price of 21.2 CNY sits near the upper-middle of that range. Asset value floor: book value per share of 6.52 CNY + net cash of 3.18 CNY, with PB of 3.25x at the 52nd percentile over 5 years.
Downside stress test: ① PV gross margin falls to 8%: gross profit declines by approximately 470 million CNY (based on annualized PV revenue of 14.74 billion CNY), after-tax net profit attributable impact of approximately -380 million CNY, 2026E attributable net profit falling to approximately 1.52 billion CNY (-20%); ② shipments decline 20%: after-tax impact of approximately -270 million CNY (excluding deterioration in fixed cost absorption); ③ trough superposition scenario (8% gross margin + shipments -20% + additional H2 FX losses of 350 million CNY + NUSA impairment of 300 million CNY): attributable net profit approximately 730 million CNY, -61% vs 2025, EPS approximately 0.45 CNY. The balance sheet can withstand this: net cash of 5.18 billion CNY (gross), long-term borrowings ≈ 0, H1 operating cash flow still positive—the real constraint lies in the income statement and dividend sustainability, not debt service; but if the trough lasts more than a year, the 84.76% payout ratio must be cut.
Management's cyclical discipline: Somewhat counter-cyclical, a plus—PV capacity expansion was concentrated in the 2021–2022 boom period (a pro-cyclical move, but already digested and focused on differentiated niches, avoiding the mainstream ground-mounted utility-scale capacity arms race). Actions in the current trough: "prudent expansion" in lithium batteries, increased R&D in AI magnetic materials (counter-cyclical investment), maintaining high dividends + buybacks, no acquisitions at market highs. Verdict: capital allocation discipline is top-tier among peers.
Overall rating: Neutral (stance: neutral, confidence 0.60, time horizon 6–12 months). Hengdian DMEGC's quality is beyond doubt—the combination of global magnetic materials leader + differentiated PV + lithium small cylindrical cells delivered consecutive profitability through an era of massive industry losses, with 5.18 billion CNY in net cash, high dividend payout, and pragmatic management communication; the price also carries no bubble—the current price of 21.2 CNY sits at the lower bound of the baseline fair value of 21–23 CNY, and the 26% decline since the last channel check has already priced in most negatives. But "not expensive" does not equal "worth buying": the AD/CVD final ruling in early September is a binary event with a retroactive tax tail, earnings quality contains tax and electricity price tailwinds, and H1 cash flow and dividend sustainability remain to be verified. With symmetric odds (bull +39% vs bear -36%), left-side entry offers insufficient risk-reward. Strategy: holders can continue holding to collect dividends and an option on recovery; incremental capital should wait for right-side signals—the AD/CVD final ruling being announced (regardless of direction, the elimination of uncertainty itself is a catalyst) or Q3 financial expense convergence + PV gross margin flattening sequentially; if the stock price rallies above 23 CNY without validation, the margin of safety is exhausted.
Risk warnings: High AD/CVD final tariff + retroactive tax payment; continued RMB appreciation leading to expanded FX losses; failure of anti-involution price support, with module price declines triggering inventory impairment; triple tax hit erosion exceeding expectations; controlling shareholder stake reduction and pledges.
Key monitoring points: ① Final tariff rates from MOFCOM and whether NUSA receives a separate rate (around 2026-09-03); ② Q3 financial expenses sequentially, PV segment gross margin, components revenue growth, inventory and receivables rundown (Q3 report in late October 2026); ③ whether InfoLink weekly module quotes can hold at 0.68–0.70 CNY/W; ④ announcements of a new round of buybacks/group share increases; ⑤ whether interim dividend policy returns to the ~50% midpoint.