| Meta Information | Content |
|---|---|
| Rating | Cautious Buy |
| Target Price | RMB 95–106 (base-case fair value range over 12–18 months; midpoint ~RMB 101) |
| Current Price | RMB 68.82 (2026-10-28 close, completed_close) |
| Margin of Safety | +38.0% (base-case fair value floor of RMB 95 vs current price) |
| Market Cap / Shares | RMB 87.6 billion / 1.273 billion shares |
| Time Horizon | 12–18 months |
| Overall Confidence | 0.58 |
Deye Technology is an asset-light manufacturer with global distribution channels, holding the world's No.1 revenue share in household energy storage inverters (Frost & Sullivan, 2025 basis, 20.6%; this figure comes from a third-party report commissioned by the company, detailed below). Net profit attributable to shareholders of RMB 4.22–4.32 billion for the first three quarters of 2026 (+79.8%–84.1% YoY) has been locked in by the earnings pre-announcement, and H1 2026 operating cash flow of RMB 4.195 billion (OCF/net profit of 1.54x) validates earnings quality; yet the current price of RMB 68.82 implies only ~14.5x 2026E consensus earnings (RMB 6.05 billion) and ~11.6x 2027E (RMB 7.57 billion)—reverse DCF shows the market price embeds an implied 2027–2031 earnings CAGR of only ~4%–9%, significantly below both consensus of +25.3% and the industry household storage shipment CAGR of 27.3%. We judge this to be a mispricing formed by "small downward revisions to expectations and a large price contraction"; the base-case fair value range is RMB 95–106, and the current price carries a 38% margin of safety. The key constraints are equally clear: the pulse-driven nature of markets such as Ukraine/Australia/Pakistan (South Africa precedent: household storage revenue fell from ~RMB 1.31 billion to ~RMB 600 million, estimated basis), battery pack gross margin compression from rising cell prices and the phase-out of export tax rebates (41.3%→31.8%), inventory up +116% from the start of the year combined with the invisibility of channel inventory under FOB revenue recognition, and potential dilution from the H-share offering. Cautious Buy: the risk/reward is asymmetric (upside to base case +38%–54%, downside to bear case −29%–−22%), but the pulse-falsification path is real. We recommend using the gross margin and inventory details in the 2026 Q3 report (disclosed by October 31) as an add-on position verification point.
Most likely catalysts for share price movement over the next 12 months (in order of importance):
Primary share price drivers: monthly production scheduling for overseas household/C&I storage shipments (September: 200,000 units, +30% MoM, sell-side channel check—the company has never confirmed this; a medium-to-low reliability leading indicator) and stabilization and recovery of battery pack segment gross margin.
Verifiable expectation gap: the market is pricing in "growth plunging from +80% in 2026 to below high single digits," while the verifiable reality is that the first-three-quarters pre-announcement has been published and consensus still sees +25.3% in 2027 (Soochow Securities has made small downward revisions to 2026E of −2.3% since mid-October; the revision direction has emerged, but the magnitude is far smaller than the share price drawdown). Verification path: production scheduling and battery pack gross margin in 2026Q4–2027H1.
Verification catalysts and falsification conditions: Q3 report gross margin/inventory details (by 10-31), details of the UK's £15 billion subsidy scheme (Q4), H-share hearing pricing, Malaysia production start (2027-01); reverse falsification signals are production scheduling turning negative MoM for two consecutive quarters, battery pack gross margin falling below 25%, or 2027 consensus revised down below RMB 6.5 billion. Evidence grade B (financial statements and customs data are verifiable, but core leading indicators such as production scheduling/share rely on single-source sell-side notes).
The RMB 4.22–4.32 billion for the first three quarters is locked in via announcement; a full year of RMB 5.5–5.9 billion implies +73%–86% YoY, requiring only a Q4 single quarter of RMB 1.23–1.68 billion (2025Q4 base was only RMB 824 million). It must be emphasized that the "lock-in" ends on September 30: the company's Q4 results have historically been highly volatile (2023Q3 South Africa destocking with single-quarter net profit −36%, 2024Q3 Ukraine pulse +229%, 2025Q4 the year's lowest); consensus of RMB 6.05 billion implies Q4 of ~RMB 1.78 billion (+116%), which is by no means a low bar; and on the investor interaction platform the company consistently responds to production scheduling and shipment guidance questions with "subject to periodic reports," never providing quantitative guidance. On earnings quality: H1 2026 OCF/net profit of 1.54x and >1.1x for four consecutive reporting periods—cash profits are real.
Industry side: global household storage shipments of 38.1GWh in 2025 (Frost & Sullivan), forecast 2026–2030 CAGR of 27.3%; rising European gas prices + the UK's £15 billion subsidy (launched January 2026) + hard-power demand in emerging markets lacking electricity form the demand matrix. Company side: in H1 2026, ten core listed inverter companies saw revenue growth but profit decline overall (net profit −18.2%), with Aro/Hoymiles/Renac turning to losses—Deye is the share consolidator amid industry clearing. But the counter-evidence must also be presented: customs tracking by country (Wanlian Securities basis) shows China's inverter exports to Pakistan fell −65.79% YoY in August 2025 and remained −17.31% in February 2026—Pakistan, the star market with "BESS imports +220%," is already seeing its inverter exports recede, indicating single-market tailwinds have a clear window; Ukraine's ~20% share of production scheduling is essentially a wartime-necessity pulse. Our judgment: the long-term household storage penetration direction is real (Australia penetration <11%, UK just starting), but the company's 2026 growth quality is a hybrid of "structural penetration × multi-market pulse relay," and Deye's core capability lies in completing cross-market transitions (South Africa→Pakistan/Middle East→Ukraine/Australia→UK) via its channel network of 800+ distributors across 150 countries—this capability itself is more sustainable than any single market.
Reverse-engineering the implied expectations at the current price: within the parameter domain of WACC 9%–12% and terminal 12–18x, a market cap of RMB 87.6 billion corresponds to an implied 2027–2031 earnings CAGR of ~4%–9% (the range is wide, but all scenarios are significantly below consensus of +25.3%). In other words, the current price requires this company to drop its growth below high single digits in perpetuity after +80% in 2026, while the verifiable trajectory is that the first-three-quarters +80% has been announced, European winter peak-season production scheduling continues, and C&I storage is scaling from 0 to 1. The composition of this round's decline needs honest decomposition: the solar-storage sector broadly fell ~25% in September–October (Sungrow H1 2026 attributable net profit −32%, PB at a one-year low percentile), including a real sector earnings deterioration factor; Deye's own consensus began showing downward revisions from mid-October (Soochow: 2026E RMB 6.03→5.89 billion, −2.3%; target price RMB 142→123), and the summary statistic of "expectations flat over the past 90 days" lags this inflection point. But between the −2.3% revision and the −44% drawdown (from the 52-week high), the remainder is valuation contraction rather than earnings repricing. Horizontally, Deye's premium of 14.5x 2026E (vs Sungrow ~11x, Ginlong ~12x, GoodWe ~13–15x) is supported by the sector's highest ROE (2025 weighted 32.1%), highest net margin (25.5% vs peers' 3–15%), and the world's No.1 niche share; PEG ~0.41. This thesis's confidence is limited by one circularity risk: the denominator of the horizontal multiples is itself consensus that may continue to be revised down; if the 2027 industry deceleration materializes, the "cheapness" will be digested by earnings downgrades rather than repaired by a rising share price.
Four pillars of the bear path: ①single-market pulse receding—South Africa household storage revenue fell from a peak of ~RMB 1.31 billion to ~RMB 600 million in 2025 (~−54.5%, estimated basis of the annual report country table), with a tailwind period of about 2–3 years; Ukraine's current share of production scheduling is ~20%; if post-ceasefire grid reconstruction reduces backup power demand, the estimated drag on net profit could reach the RMB 600 million scale on 2026 revenue; ②battery pack gross margin compression—314Ah cells rose from RMB 0.31/Wh at end-2025 to RMB 0.37–0.40/Wh (+20–30%); the export tax rebate for battery products drops from 9% to 6% from April 2026 and is fully eliminated from January 2027 (inverters not yet on the elimination list); if price pass-through is insufficient, annual net profit erosion starts at RMB 250–300 million—offsetting factors: Sungrow's across-the-board price increases of 5–15% from September 2026 indicate the industry price pass-through channel has opened, and lithium carbonate has fallen back to RMB 105,000–120,000/ton; ③inventory and channel inventory invisibility—inventory of RMB 3.855 billion (+115.91% from the start of the year, ~31.5% of 2025 revenue), and FOB shipment-based revenue recognition makes channel stuffing immediately reflected in the statements; ④FX exposure—H1 FX loss of RMB 230 million (~2.2% of revenue), with derivative instruments balance of only ~RMB 57 million and low hedge coverage. If growth falls back to single digits in 2027, 12–15x × ~RMB 5.0 billion net profit implies RMB 47–59; adding back net cash of RMB 7.25/share gives RMB 54–67, i.e., ~3%–21% further downside from the current price—an undervaluation judgment does not mean no downside.
Net cash of ~RMB 9.23 billion (11.236−2.009, independently verifiable) (~RMB 7.25/share), TTM dividend yield ~4.2%; under our 2027E attributable net profit of RMB 6.6–7.2 billion and 75% payout assumption, the implied dividend yield is ~5.7%–6.2% (this calculation depends on earnings delivery and is not rigid protection). A RMB 100–200 million buyback has executed RMB 40.45 million (0.11%–0.23% of market cap; signal value exceeds support magnitude); National Social Security Fund Portfolio 103 has increased its holdings for four consecutive quarters to 12 million shares (~0.94% of total share capital). The hedges are equally real: if the H-share offering is priced at a discount exceeding 20% with 8–12% dilution, it will simultaneously erode per-share metrics and set a new downward valuation anchor—the capital allocation tension between an 86% payout ratio and dilutive financing needs explaining; the storage subsidiary's High-New Technology Enterprise qualification expired at end-2025 with no renewal record found; if taxed at 25% with retroactive adjustments, the annualized impact is ~RMB 140 million with a one-time top-up of ~RMB 100 million scale.
| Metric | 2023 | 2024 | 2025 | 2026Q1 | 2026H1 (latest periodic report) |
|---|---|---|---|---|---|
| Operating revenue (RMB 100M) | 74.80 | 112.06 | 122.24 | 44.59 | 106.41 |
| Revenue YoY | — | +49.82% | +9.08% | +73.77% | +92.23% |
| Net profit attributable to parent (RMB 100M) | 17.91 | 29.60 | 31.71 | 11.88 | 27.17 |
| Attributable net profit YoY | — | +65.29% | +7.11% | +68.37% | +78.53% |
| Ex-non-recurring net profit (RMB 100M) | 18.53 | 28.05 | 30.35 | 11.47 | 25.70 |
| Gross margin | 39.83% | — | 38.13% | 41.45% | 37.65% |
| Net margin | 23.95% | 26.42% | 25.94% | 26.65% | 25.54% |
| Operating cash flow (RMB 100M) | 20.81 | 33.67 | 39.86 | 18.31 | 41.95 |
| Free cash flow (RMB 100M) | 12.10* | 28.98 | 26.99 | — | 37.03 |
| Cash + cash-like assets (RMB 100M) | — | — | 100.41 | — | 112.36 (2026-06-30) |
| Interest-bearing debt (RMB 100M) | — | — | 34.07 | — | 20.09 (2026-06-30) |
| Debt-to-asset ratio | — | — | 48.11% | 47.13% | 53.93% |
| Net cash (RMB 100M) | — | — | 66.34 | — | +92.27 |
*2023 free cash flow estimated as OCF 20.81 − CapEx 7.49. Balance sheet stocks are per the 2026-06-30 semi-annual report. Additionally: the 2026 first-three-quarters earnings pre-announcement (2026-10-15): attributable net profit RMB 4.22–4.32 billion (+79.81%–+84.07%), ex-non-recurring RMB 3.95–4.06 billion (+73.33%–+78.16%), implying Q3 single-quarter attributable net profit of RMB 1.503–1.603 billion (+82%–+94% YoY, −2%–+5% QoQ).
Reasons for metric changes (accounts with YoY ≥ ±20%, all with company explanations):
2026H1 was the company's best interim report since listing: revenue of RMB 10.641 billion (+92.23%) was already close to 87% of full-year 2025, and net profit attributable to shareholders of RMB 2.717 billion (+78.5%) reached 86% of full-year 2025. Structurally, battery packs (+244%) overtook inverters (+94%) as the largest revenue source, validating the company's "inverter + battery pack bundled sales" playbook. Two divergences in quality metrics warrant tracking: ① OCF of RMB 4.195 billion (+176.7%) far exceeded net profit, but this includes a contribution from expansion of payables financing; ② the combination of inventory +115.91% and contract liabilities +29.59% (RMB 622 million) shows advance receipt growth significantly lagging revenue growth — if channel destocking follows, revenue elasticity will be amplified. The Q3 positive preannouncement (net profit attributable of RMB 4.22–4.32 billion) confirms continued growth but flat sequential momentum (Q3 standalone QoQ of −2% to +5%); Soochow Securities attributes this to "FX losses + delayed revenue recognition." On market pricing: on the day of the preannouncement (10-15), the stock fell 4.2% to RMB 73.08, indicating high growth has been partially priced in and the market is beginning to lose faith in the "sequential growth every quarter" slope — excess returns must now come from verification of third-quarter report details (whether Q3 gross margin ≥36%, battery pack segment gross margin, whether inventory destocks alongside Q4 shipments), rather than from another beat on totals.
Business model: an asset-light manufacturer + branded energy-storage equipment vendor with global distribution channels. Fixed assets are only about 13.6% of total assets, with no goodwill; revenue is one-off equipment sales (no subscription/recurring revenue; the Deye cloud platform is an embryonic value-added service), but the household storage category has quasi-repeat-purchase attributes — "penetration improvement × battery-pack attach-rate improvement" — where the same household first buys an inverter and then expands with battery packs. Pricing power is cost-leadership rather than premium pricing: energy-storage inverter gross margin of 51.1% is the industry's highest, but the company states on its investor platform that "pricing follows market principles"; the estimated annual ASP decline for storage inverters is about 17% (from ~RMB 8,117/unit in 2024 to ~RMB 6,763/unit in 2025), offset by cost reduction.
Cash content of earnings: OCF/net profit over four consecutive periods: 1.16 (2023) → 1.14 (2024) → 1.26 (2025) → 1.54 (2026H1); FCF/net profit: 0.74 → 0.98 → 0.85 → 1.36 — the vast majority of book profit converts into freely disposable cash (including contributions from bill discounting and supply-chain financing, but the directional conclusion is solid). Recurring-earnings test: 2025 non-recurring-adjusted profit of RMB 3.035 billion vs. net profit attributable of RMB 3.171 billion; non-recurring gains/losses were only RMB 136 million (mainly government grants and wealth-management income), a 4.3% gap, with no one-off gains dressing up results; the projected gap for 2026Q1–3 is about RMB 260 million, roughly 6% — equally clean. Under A-share standards, the company does not disclose aggressive Non-GAAP adjustments; there is no "treating SBC as nonexistent" issue.
Return on capital: 2025 ROIC was approximately 23.4% (NOPAT RMB 3.218 billion / invested capital RMB 13.736 billion), significantly above WACC (approximately 9–11%); excluding roughly RMB 10 billion of cash-like assets, the return on operating capital is even higher. 2025 weighted ROE was 32.1%, in the industry's top tier.
CapEx quality: 2025 CapEx of RMB 1.287 billion was 5.2x depreciation and amortization (2.5x in 2024), primarily growth investment — the Malaysia base (3GW inverters + 200,000 battery packs, production starting January 2027) and the Cixi C&I storage production lines; the company does not disclose a maintenance/growth split, so maintenance investment cannot be precisely isolated, but capacity utilization (inverters 79.6%, battery packs 78.1%) is still in a ramp-up phase, and current investment matches demand expansion.
Moat and quality red flags: moat = low-voltage off-grid/hybrid inverter technology adapted to weak-grid scenarios (4ms grid-tie/off-grid switchover) + a channel of 800 distributors across 150 countries + the lowest cost ratio among peers (~72%, about 11pp lower than Sungrow). Red flags: top-five customer concentration of 31.94% (41.84% in 2024 — improving but above the 30% threshold); ~80% overseas revenue with large USD/EUR exposure; demand highly dependent on policy and unexpected events (Australian subsidies, the Ukraine war, European electricity prices); the annual report states "in H2 the Ukrainian energy system collapsed due to the war… Q3 erupted again, and European warehouse spot inventory sold out" — this is a genuine source of high-quality growth, but also a source of volatility.
Words-vs-deeds consistency: partially delivered. ① The 2024 annual report operating plan — "consolidate existing advantage markets, explore emerging markets" (qualitative) → in 2025, Europe recovered, the Middle East broke through, India and Vietnam doubled, Australia ramped up — qualitative delivery; ② major change in execution of the 2023 private placement fundraising (RMB 2 billion): the original "annual capacity of 25.5GW string/storage inverter production lines" was terminated after completion in 2025 and redirected to 7GWh of C&I storage; funding for the 3GW microinverter and R&D center projects was cut to zero; cumulative repurposed fundraising of RMB 651 million, or 32.69% of total raised funds — in hindsight the strategic pivot was right (C&I storage boom exceeded string inverters), but the original feasibility analysis was overly optimistic; twice adjusting with demand suggests "agile but not rigorous"; ③ the company consistently gives no quantitative guidance (responding to production-scheduling/shipment questions on the investor platform with "refer to periodic reports"), indicating conservative expectation management.
Shareholder friendliness: friendly. 2024/2025 payout ratios of 82.9%/86.4%; an interim 2026 dividend of RMB 2.037 billion (payout 75%); in 2025, RMB 100 million of buybacks for employee stock ownership; in September 2026, another RMB 100–200 million buyback proposed by controlling shareholder Zhang Hejun. Deduction: while pushing an H-share equity financing alongside cash-like assets exceeding RMB 11 billion and an 86% payout ratio, the capital-allocation logic of "high dividends + new equity raising" requires an explanation from management.
Risk signals: concentrated controlling-family ownership (Aisirui 35.20% + Zhang Hejun 22.26%); the controlling shareholder's affiliated vehicle Hengli Venture Capital sold 1.58 million shares (0.17%) in January 2026 — small in amount but the direction warrants tracking; participants in the 2024 private placement have a cost basis of RMB 55.56 with limited paper gains at the current price; the storage subsidiary's high-tech enterprise qualification renewal has not been publicly disclosed (see E4 for details).
2025 revenue mix (share of main-business revenue):
| Segment | Revenue share | Gross margin | YoY | One-line business logic |
|---|---|---|---|---|
| Energy-storage inverters | 42.8% | 51.10% (−0.65pp) | +18.92% | Core household-storage product; #1 globally by revenue; main profit driver |
| Energy-storage battery packs | 31.4% | 31.81% (−9.49pp) | +56.34% | Bundled with inverter channel synergy; 2026H1 revenue +244%, rising to second pillar |
| PV inverters (string + micro) | 8.6% | 32.11% (−1.40pp) | −7.21% | Grid-tie complementary category; intense competition, stagnant growth |
| Heat exchangers | 7.7% | 7.71% (−4.41pp) | −52.03% | Air-conditioner component OEM; shrinking with customer orders |
| Dehumidifiers | 6.6% | 29.72% (−4.72pp) | −16.91% | #1 online in domestic sales; weather-driven, persistently shrinking |
| Solar air conditioners and others | ~2.6% | 31.82% | +15.57% | Off-grid PV air conditioners sold to Middle East/Africa; small but beautiful |
The profit engine is energy-storage inverters: 42.8% revenue share × 51.1% gross margin contributes ~58% of main-business gross profit; battery packs, the second-largest by revenue, have a gross margin about 19pp lower — high volume, thin margin. The gross-margin structure spans 43pp (51.1% vs 7.7%), essentially two different businesses: branded power-electronics products (off-grid technology fit + channel brand premium + cost leadership) vs. home-appliance OEM work (customer-specified, processing-fee logic). The 2026 structural shift: in H1, inverters (RMB 5.132 billion, +94%) + battery packs (RMB 4.894 billion, +244%) together made up 94.2% of revenue, with traditional appliance components (heat exchangers + dehumidifiers) falling to single digits — the company's "storage content" underwent a qualitative change within one year, which also means profit margins are now more sensitive to energy-storage supply-chain prices (cells, competitive landscape).
Accounting red flags (pattern/severity/evidence):
Cross-period consistency: ① AR growth vs revenue growth (2023→2025: RMB 560 million → 1.70 billion → 1.70 billion, a plateau after a surge) — company explanation and data are inconsistent (2024 explanation vague); ② inverter capacity utilization 87.73%→79.57% — the company's explanation of "a dynamic indicator within a reasonable range" is consistent with the expansion pace, but continued declines warrant vigilance; ③ battery pack gross margin 41.30%→31.81% — the company attributes this to "industry cycle, competition, raw materials, product mix," consistent with cell price-increase data; ④ FX swings in finance costs (−RMB 214 million → −131 million → −129 million → +174 million) — consistent with FX direction. Overall judgment: no systematic earnings-management traces, but the revenue recognition structure (FOB + distribution) naturally amplifies reported volatility; the tracking focus should be marginal changes in inventory/contract liabilities rather than profit itself.
Current price RMB 68.82 (2026-10-28 close), market cap RMB 87.6 billion. PE(TTM) 20.1x (tool methodology, TTM earnings through 2026H1); about 17.2x including the Q3 preannouncement midpoint. PE percentile over 1/3/5 years: 1%/30%/22% — must be noted: the company is in a boom-expansion year, with TTM earnings containing +80% growth; mechanical percentiles being low are for reference only and do not support an "all-time-low valuation" conclusion. PB 7.63x (55th percentile over 3 years, against 2025 weighted ROE of 32.1% — the PB-ROE combination is not extreme). PS(TTM) about 5.1x. Liquidity normal (10-28 single-day turnover about RMB 520 million, ~0.6% of market cap).
| Company | Market cap (RMB bn) | PE(TTM) | 2026E forward PE | 2026E growth | Gross margin (2025) | Notes |
|---|---|---|---|---|---|---|
| Deye | 87.6 | 20.1x (17.2x incl. preannouncement) | 14.5x | +73%~+86% (own estimate RMB 5.5–5.9bn) | 38.1% | #1 household-storage inverter revenue, ROE 32.1% |
| Sungrow | 141.5 | 12.9x | ~11.0x | ~−5% (consensus already revised down) | 31.8% | Utility-scale storage giant; 2026H1 net profit −32% |
| Ginlong Solis | 18.9 | 33.5x | 12.0–12.6x | ~+100% (low-base recovery) | ~34% | Grid-tie origins; H1 net profit −29.7% |
| GoodWe | 11.5 | 26.4x | 12.8–15.2x | ~+460% (loss-to-profit recovery) | 21.5% | Weak earnings quality; adjusted/attributable only 26% |
(Market caps and PEs as of the 2026-10-28 close; earnings forecasts from Aug–Sep 2026 broker reports; Sungrow's denominator is still in a downward-revision channel, so cross-sectional multiples carry distortion.)
RMB 87.6 billion market cap ÷ 2026E consensus earnings of RMB 6.05 billion = 14.5x; ÷ 2027E consensus of RMB 7.57 billion = 11.6x. Reverse DCF (WACC 9–12%, terminal multiple 12–18x parameter ranges) solves for an implied 2027–2031 earnings CAGR of approximately 4%–9% — the current price requires the company's growth to plunge below high single digits after 2026's +80% and stay there long term; whereas the actual trajectory is: first three quarters +79.8%~+84.1% already announced, consensus 2027 +25.3%, industry household-storage CAGR 27.3%, and the company's household-storage shipments already +109% in 2026H1. Two-way test: if the implied low growth is right, then European restocking ebbing + Ukraine front-loading + battery-pack margin erosion must all occur simultaneously; if consensus is right, the same DCF framework values the stock at about RMB 161/share. Our ruling: the mispricing direction is bullish, but the magnitude is constrained by the downward-revision risk in consensus itself (Soochow already cut 2026–28E by −2.3%/−4.3%/−8.7% in mid-October), hence medium confidence.
| Scenario | Probability | Fair value range | Core assumptions | vs. current price |
|---|---|---|---|---|
| Bear | 32% | RMB 49–54 | 2027 net profit attributable ~RMB 5.4bn (−1%): European restocking ebbing + Ukraine front-loading + battery-pack gross margin falling below 25%; 12x exit (mature-hardware valuation) | −29%~−22% |
| Base | 48% | RMB 95–106 | 2026 net profit attributable RMB 5.5–6.0bn, 2027 +15%~+20% to RMB 6.6–6.9bn; battery-pack gross margin stabilizing at 28–30%; 18–20x exit | +38%~+54% |
| Bull | 20% | RMB 151–157 | 2027 net profit attributable RMB 7.8–8.0bn (~+30%): C&I storage doubling + Malaysia ramp + UK subsidies beating expectations; 25x exit | +119%~+128% |
The base-case 18–20x exit is anchored on: the upper-middle of comparable companies' 2026E forward PE range of 11–15x (Deye's ROE/net margin/share are all sector-best, warranting a quality premium) + sell-side target prices generally at 20x 2027 earnings (Soochow RMB 123, Huachuang RMB 120.68) + a discount to the company's own 3-year median PE of ~23.9x — the anchor is independent of the company's current multiple; the current price is not used to justify the current price. The bear case at 12x references the midpoint of mature hardware companies. External bearish reference: Soochow's post-revision target of RMB 123 (2026-10-15; 2027E RMB 7.86bn × 20x) is still a sell-side midpoint rather than a stress scenario; the disagreement with our bear case lies fundamentally in earnings-path assumptions (RMB 7.86bn vs ~5.4bn), so it is not adopted as the bear-case boundary. Risk note: if the H-share issuance occurs at a >20% discount during the base-case realization, the transition between base and bear cases may be continuous rather than a discrete jump — a pacing risk, not a terminal-value risk.
Own earnings forecast (reconciliation anchor):
| Period | Revenue | Net profit attributable | Key assumptions |
|---|---|---|---|
| 2026E | RMB 24.5–25.8bn (+100%~+111%) | RMB 5.5–5.9bn (+73%~+86%) | Household-storage inverter shipments >1.3M units for the year, C&I storage 150k–200k units, battery-pack revenue >RMB 11bn; inverter gross margin 47–48%, battery packs 27–29% (cell price increases + export-tax-rebate cuts); FX drag of RMB 200–300 million |
| 2027E | RMB 30.0–33.5bn (+22%~+30%) | RMB 6.6–7.2bn (+15%~+22%) | Household-storage growth falling to ~+20% on a high base, C&I storage doubling to 250k+ units, Malaysia ramping up; battery-pack gross margin recovering to ~30% to offset the export tax rebate cancellation; below consensus of RMB 7.57bn — we assume a high base, neutral FX, and neutral margins |
Undervalued. Target price range RMB 95–106 (base-case fair value), midpoint ~RMB 101; margin of safety +38.0% (on the base-case fair-value floor). Quality and price assessed separately: quality is the top tier of the PV-storage sector (ROE 32.1%, net margin 25.5%, net cash RMB 9.2bn, #1 global niche share, OCF/net profit consistently >1.1); price is set on a "rapid growth decay" basis. Probability-weighted fair value ~RMB 95.5; the current price sits below the midpoint between the bear and base cases. The odds are asymmetrically bullish, but "undervalued" does not provide unlimited downside protection — under the bear case, the current price still carries theoretical downside of 22%–29% (narrowing to 3%–21% after adding back net cash).
Deye's true competitive arena is global distributed residential/C&I solar-plus-storage equipment, not front-of-the-meter utility-scale storage or centralized power plant equipment. Market size by source reliability (graded):
Demand-side quantified chain (penetration ramp-up math): Driving variables = ① unmet power demand × electricity prices in emerging markets (Pakistan imported 16–17GW of modules in 2024, cumulative over 51GW; Ukraine added 2.9GWh of storage in 2025, jumping to No.4 in Europe) × ② system cost declines (LFP cells from USD 460/kWh in 2015 to ~USD 110/kWh in 2024, −75%; China's average module export price per watt to Pakistan −54% over 2022–2024) → at equal budgets, per-household installable capacity doubles → ③ penetration rate (Australia's residential storage penetration <11%; Pakistan's cumulative BESS imports only 7.6GWh, residential battery attachment rate still low; UK's GBP 1.5 billion subsidy launching January 2026) → ④ Deye's share (No.1 with 20.6% of residential storage inverter revenue). But the counter-evidence must be presented in parallel: customs-basis inverter exports to Pakistan turned sharply negative YoY from August 2025 (−65.79%) — the "structural penetration" narrative for that market is being falsified by export data. Its BESS import +220% statistic (battery vs. inverter) coexists with export declines, suggesting penetration proceeds in layers by product category, and the dividend window for the inverter segment is shorter.
Upstream: IGBT/power semiconductors (supply loosened after the 2021–22 chip shortage plus domestic substitution; inverter makers have strong bargaining power) and storage cells (70–80% of battery pack cost) — 314Ah cells rose from RMB 0.30–0.34/Wh in late 2025 to RMB 0.37–0.40/Wh by June 2026 (+20–30%); lithium carbonate rose from a 2025 low of RMB 58,000 to a peak of ~RMB 200,000 in May 2026, falling back to RMB 105,000–120,000/ton in October; cell lead times extended from 30 days to 75 days in 2025Q3 (CNESA), with profits migrating upstream for a period. Midstream: inverter/battery pack manufacturing; Chinese vendors hold 9 of the global top 10 slots (WoodMac), with pronounced profit stratification — inverter power electronics gross margins of 21–51%, far above modules (0–3%); Deye, leveraging its home-appliance supply chain DNA, achieved the lowest cost ratio among peers (~72%). Downstream: channels/installers; residential storage is a classic channel business (regional distributors → installers → households). Deye has 800+ distributors across 150 countries, exclusive UK agent Sunsynk (~30% UK share), and started in South Africa via CNBM's channels. Deye sits in the midstream branded manufacturing link (the fattest profit pool in the chain); its bargaining power vs. upstream cells is relatively weak (forced to give up margin on battery packs during the 2026 price-hike cycle), while downstream it builds stickiness via integration + low-voltage easy installation + cost-performance, with buyout-based channels ensuring good cash collection.
The global inverter market is highly concentrated: in 2024, Huawei + Sungrow held ~55% of shipments combined (Huawei 176GW, Sungrow 148GW); top 10 about 71–86% (2022 CR10 = 86%). Segmented stratification: F&S 2025 revenue basis — Deye No.1 in residential storage inverters at 20.6%, No.2 in C&I storage inverters at 20.4%; on GW-basis global overall rankings (S&P/WoodMac), Deye does not make the top five — the two metrics must be understood in parallel: Deye wins on "unit count × low-voltage small power," with lower per-unit value than high-power models. Rumored shares of ~60% in Pakistan/Southeast Asia and 55%+ in South Africa come from non-authoritative sources such as Xueqiu (low reliability, to be verified). What is confirmable: No.1 in South Africa residential storage inverter revenue (F&S), with 50%+ market share in 2023, falling to ~RMB 600 million (estimated) in 2024–25 as competitors flooded in. Competitive intensity: in 2026H1, ten listed inverter companies posted revenue −3.9% / net profit −18.2%; Aro/Hoymiles/Renewsys/Sungrow-turned-loss-makers (Aiko? no — Solplanet? keep as is: 艾罗/禾迈/昱能/首航) — i.e., Aier/Hoymiles/Renewsys/Sungo turned to losses... Let me restate: Aier (Solax), Hoymiles, Renac (Re ")" — correction: 艾罗=Solax Power? Render as Solax, Hoymiles, Renac... Actually standard English names: 艾罗能源 = Solax Power; 禾迈 = Hoymiles; 昱能 = APsystems; 首航 = Sungrow? No, 首航新能 = Senergy/Sofar? 首航新能源 = Sineng? It's "Sofar"? Use: Solax, Hoymiles, APsystems, and Senergy turned to losses; Sungrow, Deye, and Sigfox (思格 = SigenStor/Studer? 思格新能源 = Sigen) — render as Sungrow, Deye, and Sigen took 95% of industry profits — share consolidation amid shakeout is the industry backdrop for Deye's rising share. Entry barriers: certifications (VDE/G99/AS4777, etc.), channel networks, brand bankability (BNEF bankability), cost. Substitution threats: residential all-in-one units (Tesla Powerwall/Huawei) upgrading over split systems; utility-scale storage's low cost squeezing residential storage economics (utility-scale exceeded half of new storage additions in Europe for the first time in 2025).
The industry cycle is stratified: total inverter volume is in the downswing of maturity; residential storage is in mid-growth structural boom (Europe destocking completed in 2024, recovery in 2025–26 plus new subsidies); C&I storage is at 0→1. Leading indicators: cell prices and lead times (30→75 days = strong demand but margin erosion), lithium carbonate, channel inventory, subsidy schedules, monthly production schedules, and country-by-country customs exports. Regulation is the biggest variable this cycle: the US FCC listed Chinese networked inverters on its Covered List effective 2026-07-28 (with grandfathering; Deye's off-grid models are not covered, US revenue share <5%, limited direct impact); the EU bans public funds from supporting projects using Chinese inverters (China-EU consultations in October marginally eased; spread to private markets is a tail risk, probability ~15–20%); Pakistan's proposed module GST hike from 10% to 18%, BESS total tax burden ~48%, net billing reform in February 2026 (negative for pure grid-tie, positive for storage-equipped systems); India's ALMM; China's export tax rebates: battery categories cut from 9% to 6% from April 2026 and fully abolished from January 2027 (legislated); inverters not yet on the abolition list — this is the most certain negative for the 2027 income statement; the hedge depends on pass-through capability (Sungrow raised prices 5–15% in September 2026). Positives: UK GBP 1.5 billion, Poland EUR 230 million, Hungary EUR 260 million subsidies; the Netherlands abolishing net metering; Australia expanding subsidies to AUD 7.2 billion (phasing down from May 2026).
| Company | 2025 Revenue (RMB bn) | 2025 Growth | Gross Margin | Net Profit to Parent (RMB bn) | 2026H1 | Key Differentiator |
|---|---|---|---|---|---|---|
| Deye | 12.224 | +9.1% | 38.1% | 3.171 | Revenue +92% / net profit to parent +79% | Residential storage + emerging markets + low-voltage off-grid; 25.5% net margin, highest in industry |
| Sungrow | 89.148 | ~+14.5% | 31.8% | 13.461 | Revenue −29% / net profit −32% | Utility-scale storage + ground-mount giant; 7x scale but mirrored growth |
| Ginlong Solis | 6.952 | +6.3% | ~34% | 0.743 | Net profit −29.7% | Grid-tie string + European channels; storage second curve just starting |
| GoodWe | 8.889 | +31.9% | 21.5% | 0.135 | Revenue +53% | All-in-one units + Australian battery ramp; weak earnings quality (non-GAAP/GAAP 26%) |
| Hoymiles / APsystems (microinverters) | 1.78 / 0.30 (H1) | — | ~27% | Loss-making | Continued losses | Microinverter route squeezed by residential all-in-one units; being shaken out |
Niche champion: No.1 globally in residential storage inverter revenue (F&S 2025, company-commissioned basis, lacking independent third-party cross-verification — the company does not make the top five on S&P/WoodMac GW-basis overall rankings; both metrics shown side by side above). Share trending upward (2026H1 residential storage shipments +109% amid industry-wide shakeout); moat = low-voltage off-grid technology suited to weak grids (4ms switchover) + channel network (cross-market relay capability: South Africa → Pakistan/Middle East → Ukraine/Australia → UK) + industry-lowest cost ratio. Positioning risk: Deye's growth model is essentially "chasing one market window after another," with single-market dividend periods of roughly 2–3 years (South Africa precedent); sustained growth requires the company to keep finding the next market — this is both a capability and a continuous test of management.
Generic financial frameworks tend to miss three defining dimensions of the inverter industry:
① Shipment structure matters more than total volume. Deye's 2025 sales structure: 771,500 storage inverters (+42.7%), 634,000 string inverters, 228,200 microinverters; 2026H1 total 1.153 million units (residential storage 570,300, +109%; C&I storage 60,500 units). The storage/grid-tie mix determines the gross margin anchor (storage inverters 51.1% vs. the "PV inverter" segment containing string inverters at 32.1%); tracking shipment structure reveals earnings changes earlier than tracking revenue. Monthly production schedules are the most sensitive leading indicator (September: 200,000 units, +30% MoM, sell-side survey basis — unconfirmed by the company; must be cross-validated against customs exports and quarterly reports).
② Market structure determines growth sustainability. Overseas revenue is ~80% of total (2025); overseas new-energy product revenue by country: Germany ~RMB 1.88 billion, South Africa ~RMB 600 million, Ukraine ~RMB 560 million (FY2025 country table, estimated basis) — mature European markets provide the floor while pulse markets provide elasticity. Single-country concentration, subsidy calendars (Australia phasing down 2026-05, UK launching 2026-01, Pakistan tax reform) predict next quarter's revenue better than aggregate industry data.
③ Unit price is a long-term downward variable; the volume-price scissors gap is the norm. Storage inverter ASP fell from ~RMB 8,117/unit in 2024 to ~RMB 6,763/unit in 2025 (−17%); battery pack ASP RMB 4,932/unit (−6%) — the industry trades price for penetration, and the company sustains profits by keeping "cost-down speed > price-cut speed." 2026Q1 gross margin of 41.45% proves pass-through capability exists, but Q2 blended gross margin fell back to ~34.9% (cell price hikes + FX) — quarterly gross margin volatility is this company's biggest single-point risk in financial reporting and the number most worth watching in the Q3 report.
Overall verdict: Cautiously bullish, overall confidence 0.58, time horizon 12–18 months. Confidence on the five theses: 0.80/0.65/0.62/0.75/0.65 — most certain is "2026's high growth is already locked and announced" (fact-type); least certain is "the mispricing will be corrected by the market" (direction-type, constrained by sector beta and consensus downgrade risk). Strategy implication: at the current price of RMB 68.82, there is a 38% margin of safety versus the base-case fair value of RMB 95–106; the odds are asymmetrically bullish (upside to base case +38%~+54%, to bull case +119%~+128%; downside to bear case −29%~−22%, or −21%~−3% after adding back net cash); but a genuine pulse-falsification path exists. On position pacing, the Q3 report (due by 2026-10-31) — Q3 gross margin, battery pack segment gross margin, and inventory drawdown — should serve as the core verification points; H-share issuance pricing (a discount >20% is the warning line) determines pacing, not direction.
Risks and falsification signals (written summary): ① monthly production schedules turning negative YoY for two consecutive quarters, or Ukraine's share of production falling below 10% with no UK/emerging-market relay — pulse fade falsifies C2; ② battery pack segment gross margin falling below 25% without price-hike offsets, or Q3 blended gross margin below 36% — margin squeeze falsifies C1's full-year ceiling; ③ 2027 consensus downgraded below RMB 6.5 billion — mispricing thesis (C3) invalidated; ④ storage subsidiary's income tax rate jumping to 25% with retroactive payments — annualized profit impact of ~RMB 140 million plus compliance blemish; ⑤ H-share issuance at a discount over 20% or dilution over 12% — valuation anchor shifts down; ⑥ EU restrictions on Chinese inverters extending to private markets — European revenue under pressure (Germany alone at RMB 1.88 billion in 2025 is the largest single overseas new-energy market; Europe's overall share is higher, estimated basis).
Catalyst calendar: Q3 report official disclosure by 2026-10-31 (core verification point) → 2026Q4 UK subsidy details finalized + H-share hearing window → 2027-01 Malaysia plant starts production → 2027-04 annual report and Q4 gross margin confirmation → 2027H1 C&I storage revenue ramp validation.
Report date: 2026-10-29. Price and valuation anchor: 2026-10-28 closing price of RMB 68.82 (completed_close). This report is based on public information (company announcements/periodic reports/regulatory filings/news and sell-side research); leading indicators (production schedules, regional shares) partly rely on sell-side survey data, with reliability noted throughout the text.