Ningbo Deye: H1 Operating Cash Flow Was 1.54x Net Income, but CNY 4.071bn of It Came from Expanding Supplier Credit

2025 unit data are especially useful because they stop us mistaking revenue growth for pricing power. Storage-inverter unit sales rose 42.72% to 771,500 while segment revenue rose only 18.92%; implied revenue per unit fell about 16.7%. Battery-pack units rose 67.16% to 777,000 while revenue rose 56.34%, implying roughly a 6.5% fall in revenue per unit. PV-inverter units grew 3.76% but revenue fell 7.21%, an implied price/mix decline of about 10.6%. These are my calculations from the company's reported units and revenue, and they show that 2025 growth was overwhelmingly volume-led rather than price-led.

Battery-pack unit data in the public H1 filing are too thin to repeat that volume-price decomposition cleanly for 2026. Public disclosures cannot establish how much of the 244% battery-pack revenue growth was unit volume versus pack capacity, product mix or price. What can be established is that FX did not create the reported growth. Management explicitly attributed the swing in financial expense from a CNY 173m gain-equivalent in H1 2025 to CNY 174m of expense in H1 2026 mainly to declines in the USD and EUR against the renminbi, while the cash-flow statement records a CNY 160m negative effect from exchange-rate movements. Currency was a headwind to earnings, not a hidden tailwind.

Whether shipment growth equals end-market demand is less certain. At June 30 inventory had more than doubled from year-end, from CNY 1.79bn to CNY 3.85bn. Finished goods rose from CNY 0.99bn to CNY 1.97bn and raw materials from CNY 0.44bn to CNY 1.20bn. Receivables increased 48% to CNY 2.52bn. Management says the inventory increase reflects preparation for strong demand and the receivables increase reflects higher sales. Those explanations are plausible, but they establish Deye's own production build, not distributor sell-through.

There is evidence against the simplest "pure channel stuffing" interpretation. Customer concentration has been falling: the top five represented 53.2% of revenue in 2023, 41.8% in 2024, 31.9% in 2025 and 23.8% during the first four months of 2026; the largest customer's share declined from 19.4% to 10.0% over the same sequence. The company also says its sales teams inspect distributors two or three times a year and monitor inventory turnover and secondary distribution. That broadening of the customer base is healthier than a boom concentrated in one distributor. Yet the prospectus does not disclose aggregate distributor inventory or monthly sell-through, leaving precisely the variable that mattered during the European and South African destocking episodes opaque.

The key unresolved question is whether 2026 is a new run-rate or another distributor-driven crest. The answer matters much more than whether H1 growth was 80%, 90% or 100%. A residential storage inverter is a durable piece of hardware sold through distributors and installers into markets where economics can change quickly with electricity tariffs, subsidies, blackouts, interest rates and battery prices. Orders can therefore run well ahead of installation for several quarters and then disappear while channel inventories clear. SolarPower Europe's data show exactly this kind of divergence: Europe installed a record 36 GWh of batteries in 2025, up 48% overall after the 2024 slowdown, but EU residential battery installations still fell 6% to 9.8 GWh, a second consecutive annual decline. Utility-scale growth masked continued softness in the household segment.

Deye's present wave appears broader geographically than the 2022 European energy crisis or isolated South African/Pakistani booms. The H1 report cites resilient European demand, hard backup-power demand in Ukraine and expansion across emerging markets where grid reliability is poor. The prospectus identifies China, Germany, Pakistan, India and South Africa as jurisdictions that exceeded 5% of revenue during at least one track-record year. The absence of the United States from that list suggests direct U.S. revenue was below the 5% materiality threshold in each relevant period, limiting direct U.S.-tariff exposure relative to suppliers that depend heavily on American residential solar.

The company's history explains why it found those markets. Deye grew out of Ningbo manufacturing rather than Silicon Valley-style solar technology. Its early capabilities were injection-moulded and sheet-metal parts, appliance controls, heat exchangers and dehumidification equipment. The Midea relationship was formative: by the pre-IPO era Deye had become a significant heat-exchanger and controls supplier to Midea, learning manufacturing discipline, cost control, power electronics and mass production before storage became a meaningful business.

The corporate predecessor dates to the early 2000s in Ningbo, while founder Zhang Hejun's manufacturing activity predates the listed entity. The decisive strategic turn came much later. Deye entered storage inverters around 2018, initially developing products suitable for split-phase North American applications and then low-voltage systems for South Africa in 2019. Rather than challenge Huawei and Sungrow head-on in large utility plants, it concentrated on household and small-commercial applications, particularly weak-grid and price-sensitive markets. The HKEX prospectus itself describes this as an "asymmetric competition" strategy: South Africa, Brazil and Pakistan were early priorities while mature U.S. and European categories were more crowded.

That choice left a permanent imprint on the product. Deye emphasised low-voltage, high-current storage architecture, generator compatibility, rapid on/off-grid switching and the ability to parallel multiple inverters. Those attributes matter disproportionately in countries where outages are frequent and backup power is not an occasional luxury. The same engineering later transferred into European and C&I applications. Deye came into storage from the practical problem of making power electronics work cheaply and reliably under imperfect grid conditions rather than from a premium residential-solar ecosystem.

The listing was straightforward. Deye issued 42.667m A shares at CNY 32.74 and listed on the Shanghai main board on April 20, 2021. Gross proceeds were CNY 1.397bn and net proceeds CNY 1.331bn; post-IPO shares were 170.667m, implying a then-current listing market capitalisation of about CNY 5.59bn. The CNY 32.74 IPO price is the transaction price on the then share count and should not be compared mechanically with today's CNY 85.50 because subsequent bonus issues, placements and employee-equity issuance altered the denominator. All per-share valuation work in this report uses the current roughly 1.273bn-share basis.

The first post-listing phase, 2021–22, validated storage as the growth engine. Revenue rose from about CNY 4.17bn in 2021 to CNY 5.96bn in 2022, while attributable profit rose from roughly CNY 0.58bn to CNY 1.52bn. Energy-security concerns, soaring European electricity prices and weak-grid demand turned what had been a strategic option into Deye's dominant source of incremental profit. The market stopped valuing Deye as an appliance-component company and began treating it as a storage-growth equity.

The second phase, 2023–24, proved that the growth model was cyclical even while the franchise strengthened. Revenue reached CNY 7.48bn in 2023 and CNY 11.21bn in 2024; attributable profit reached CNY 1.79bn and CNY 2.96bn. At the same time, residential solar and storage channels in Europe went through an inventory correction after the 2022 rush, while South African and other weak-grid markets experienced their own order swings. Enphase's later disclosure that European Q4 2025 revenue fell another 29% sequentially due to softer demand illustrates how violent residential-channel corrections can remain even for established global brands.

The third phase, 2025, was the transition from an inverter-led story to a broader storage-system model. Consolidated growth slowed to 9%, yet batteries expanded 56%, C&I storage inverters roughly 75%, and overseas main-business revenue rose 22.65% to CNY 9.75bn. Domestic main-business revenue fell 24%. The consolidated slowdown concealed a large redistribution inside the company: legacy appliance revenue collapsed, while batteries and higher-power storage took a larger share.

The fourth phase is the current one. In 2026, Deye is attempting to turn an unusually strong inverter channel into a wider integrated-storage platform. It is adding battery systems, higher-power PCS, liquid-cooled C&I batteries and software/energy-management functions while building a 7 GWh C&I storage production line. The company has also redirected some earlier offering proceeds away from a planned 25.5 GW string/storage-inverter capacity programme and cancelled the planned 3 GW microinverter and dedicated inverter-R&D projects, reallocating capital toward the 7 GWh C&I line. That is evidence of pragmatic capital reallocation, but also evidence that management's earlier capacity map was not a perfect forecast of where demand would materialise.

Malaysia is part of the same shift. At June 2026, construction in progress included CNY 205m for a Malaysian plant designed for 3 GW of string/storage inverters and 200,000 battery systems. The Hong Kong application shows Deye New Energy Technology (Malaysia) was incorporated in January 2025 as a wholly owned manufacturing subsidiary. This overseas footprint diversifies production away from a China-only base and should make the trade route more flexible, although a Malaysian factory does not by itself eliminate future rules-of-origin or tariff risk.

The proposed H listing should be interpreted in that context rather than as a near-term earnings event. Deye first filed on January 27, 2026; the application lapsed after six months and was re-filed July 28 with CICC, CITIC Securities and CMB International as joint sponsors. As of this report date the application proof still contains redacted offer size and price fields, so there is no reliable H-share discount, dilution percentage or proceeds quantum to model. Any valuation that treats H-share proceeds as already raised is premature.

Financial history shows how completely profitability changed with the business mix.

Profit rather than revenue is the important line. Deye moved from an appliance manufacturer earning mid-single-digit hundreds of millions to a business earning more than CNY 3bn annually because storage inverters have radically better economics than heat exchangers. In 2025 storage-inverter gross margin was 51.1%; heat-exchanger gross margin was only 7.7%. The transformation improved both scale and the quality of each revenue yuan.

Price history carries an equally important warning. The latest Shanghai close was CNY 85.50 on September 11. Market-data services show a 52-week high of CNY 126.90, reached in late May 2026, well before the August H1 release. Deye closed at CNY 96.50 on August 26, immediately before the H1 report became public, and has since fallen another 11.4%. From the CNY 126.90 high it is down about 32.6%. A spectacular earnings print did not produce a fresh high.

The de-rating is larger than the price decline suggests. Using reported 2025 profit and Q1/H1 2026 data, I estimate trailing attributable earnings around the May high at roughly CNY 3.65bn and current TTM earnings around CNY 4.37bn. On the current share count, that implies trailing EPS rose about 20% while the price fell 33%; the rough trailing P/E compressed from about 44 times at the May high to about 25 times now. Most of the move since May has been multiple compression, not an earnings deterioration.

That distinction matters. The current market narrative is no longer "92% growth forever." At CNY 85.50, Deye trades around 24.9 times reported TTM earnings. The stock still carries a growth premium, but the multiple has already absorbed a substantial normalisation discount. The live debate is whether roughly 20–25 times earnings is enough of a discount for a hardware company exposed to distributor inventories and storage subsidies, given that its inverter margins and current growth rate are far above most listed peers.

The qualitative portrait is "company in transition." The transition from appliances to storage inverters is largely complete, but the transition from selling standalone inverters to selling integrated inverter-plus-battery systems is occurring now. That second transition can enlarge revenue per installation and reinforce channel attachment. It also shifts mix toward batteries, where margins, differentiation and long-term pricing power are weaker.

Business model, moat, industry, and competitors

Deye's 2025 segment economics show where the profit pool resides.

The storage inverter is the economic engine. A CNY of inverter revenue generated far more gross profit than a CNY of battery or legacy heat-exchanger revenue in 2025. Batteries are valuable because they increase wallet share, improve system integration and make the inverter relationship harder to displace, but their 31.8% gross margin was almost nineteen percentage points below storage inverters. Battery-pack margin also fell 9.49 percentage points in 2025, versus only a 0.65-point decline in storage-inverter margin.

That makes H1 2026's mix shift economically ambiguous. Battery revenue grew from roughly CNY 1.42bn in H1 2025 to CNY 4.89bn, becoming almost as large as the entire inverter business. Total H1 gross margin nevertheless held around 37.6%, only slightly higher year on year, suggesting manufacturing scale and inverter economics were good enough to offset much of the low-margin battery mix dilution. But Q2 gross margin slipped to about 34.9%, roughly 2.8 percentage points below the prior-year quarter. The later-quarter deterioration is a better normalisation warning than the six-month average.

The battery business is particularly important because it tests what Deye's moat actually is. Deye does not appear to own a differentiated battery-cell chemistry franchise analogous to CATL. It buys cells and components, designs the pack, BMS, communications and safety architecture, and sells the battery alongside its inverter. The 2025 HKEX supplier disclosures show the two largest battery suppliers accounted for roughly 9.1% and 8.4% of total procurement. The prospectus anonymises them: one is described as part of a Huizhou-headquartered, Shenzhen-listed battery group, the other as part of a Guangzhou-headquartered listed battery group. Those descriptions are consistent with EVE Energy and Great Power respectively, but because Deye does not name them I treat those identities as unconfirmed rather than fact.

Separately, the H1 report shows a CNY 40m deposit with Xi'an FinDreams Battery. That confirms a commercial relationship with the BYD battery ecosystem but does not establish how much cell volume Deye sources from FinDreams. No public filing I reviewed gives enough contract detail to determine take-or-pay clauses, annual pricing formulas or minimum cell-purchase commitments. The battery-sourcing conclusion is concentration without single-source dependence, not vertical integration.

A storage or hybrid inverter differs materially from a conventional grid-tied string inverter. A normal string inverter's principal task is converting PV-panel DC power into grid-compatible AC. A hybrid/storage inverter must additionally move power bidirectionally between battery, household or C&I load and grid; operate safely off-grid; coordinate battery state of charge; switch through outages; manage time-of-use tariffs; and communicate with BMS and energy-management systems. Those extra control functions explain why Deye's storage-inverter gross margin is much higher than its conventional PV inverter margin.

Deye's technical identity comes from low-voltage architecture. It says it was early with high-power single-phase and three-phase low-voltage storage units, generator compatibility, multi-inverter parallel operation and millisecond-scale grid/off-grid switching. Low voltage is attractive in weak-grid residential markets because of installation simplicity and safety, but large current creates difficult thermal, magnetic and electromagnetic-interference problems. Deye's manufacturing and variable-frequency electronics heritage appears genuinely relevant to solving those practical issues.

The durable moat is the combination of low-voltage power-electronics engineering, cost-conscious manufacturing and a geographically broad distributor network; battery packs themselves are not the moat. The company says roughly 90% of product components are locally manufactured and that it produces key structural components such as enclosures internally. It also adopted domestic IGBT, analogue and digital components relatively early. These are company claims rather than independently audited cost-benchmark data, but the 51% 2025 storage-inverter gross margin and 25%-plus recent net margin are strong economic evidence that its cost structure is unusually favourable.

The second moat is product-market fit in difficult grids. Deye did not build its early franchise around the highest-income household in California. It built around markets such as South Africa and Pakistan where an inverter may need to work with generators, unstable grids, large household loads and customers highly sensitive to capital cost. That heritage broadened the addressable market and produced products that could later be sold into more conventional storage installations. The prospectus says the company now serves more than 150 countries and regions.

The third is channel breadth. Deye can enter a country through distributors without building a large owned retail or installation workforce. That holds selling costs down and lets the company localise rapidly. Customer concentration falling to 23.8% for the top five in the first four months of 2026 is evidence that the channel is broadening. The trade-off is information loss: distributors own local stock, local demand generation and often installer relationships, so Deye sees wholesale orders sooner than it sees the eventual homeowner's demand.

There are meaningful switching costs, but they should not be exaggerated. Once an installer knows the commissioning process, has spare parts, training and technical support, and has matched a Deye inverter with a compatible battery and monitoring platform, staying with that ecosystem reduces labour and warranty risk. Yet residential installers commonly carry several brands. There is no true network effect, no recurring software revenue disclosed at material scale, and no evidence that Deye Cloud has pricing power comparable with enterprise software. The ecosystem is a moderate hardware switching cost, not a captive network.

The battery attachment strategy can deepen those switching costs. Deye says inverter and battery are jointly adapted in communications, energy-management and safety logic and can offer "plug-and-play" installation. This can improve customer experience and give the distributor one party to call when a system fails. But battery cells themselves are a highly competitive, increasingly commoditised input that Deye buys rather than makes. Global storage-cell shipments nearly doubled to about 612 GWh in 2025, and that demand surge has since pulled the cell market out of oversupply and into shortage, with the large Chinese cell makers raising quotes through 2026. Cheap cells flattered pack economics on the way in and are a cost headwind now, and pack gross margin is unlikely to inherit the inverter's 50% economics either way.

Deye's distributor model explains both its operating leverage and its cyclicality. Hardware manufacturing has meaningful fixed costs in engineering, certification, factories, sales offices and after-sales networks, but components remain variable. When volumes rise, procurement economics and factory utilisation improve rapidly. When revenue falls, R&D, certification and overseas service cannot be cut without damaging the franchise. The company therefore enjoys strong positive operating leverage in booms but cannot flex its cost base as quickly as a pure distributor in a destocking cycle.

R&D is heavy but not software-like. Deye spent CNY 562m on R&D in 2025, 4.6% of revenue, with the full amount expensed rather than capitalised; 868 R&D employees represented 14.4% of staff. H1 2026 R&D rose 41% to CNY 372m, though rapid revenue growth reduced R&D intensity. That accounting treatment is conservative and means current earnings already bear the recurring product-development cost.

The company is now working on sixth-generation integrated storage products, seventh-generation miniaturisation, 250 kW PCS, solid-state-transformer research and liquid-cooled C&I batteries. Some of these projects may become differentiators; none deserves material standalone valuation today because sales and unit economics have not yet been disclosed.

Legacy appliances now look more like a manufacturing inheritance than a strategic growth engine. Heat-exchanger revenue fell to CNY 938m in 2025 at only 7.7% gross margin. The Midea group was Deye's second-largest customer, accounting for about 7.7% of 2025 revenue, which is remarkably close to the total heat-exchanger segment's size. Midea was also effectively Deye's largest supplier group for copper tube and aluminium foil, accounting for CNY 759m of procurement, or 10.5% of purchases. Deye says those transactions are conducted on ordinary commercial terms, and its top-five related-party disclosure did not identify Midea as an accounting related party. Economically, however, the remaining heat-exchanger activity is still highly intertwined with Midea.

Margin tells us how management is treating that business. At less than 8% gross margin, heat exchangers do not merit growth capital on financial grounds when storage inverters generate more than 50%. I view the segment as being run primarily for customer relationship, factory utilisation and cash recovery while its consolidated importance fades. Dehumidifiers retain better economics at about 30% gross margin, but H1 2026 revenue fell another 13% to CNY 355m.

Industry structure reinforces Deye's strategic direction. Wood Mackenzie says global PV inverter shipments increased 10% in 2024 to 589 GWac, with Huawei and Sungrow together holding 55% of the broader inverter market. Its H1 2025 global manufacturer ranking again put Huawei and Sungrow first and second, while the top ten accounted for 71%. That dataset covers solar inverters broadly and cannot validate Deye's residential-storage-specific ranking, but it confirms that Chinese companies dominate global inverter manufacturing and that scale is substantial.

The HKEX prospectus, using Frost & Sullivan commissioned research, claims Deye held 20.6% of the global residential-storage-inverter market by 2025 revenue, ranking first with about CNY 3.7bn of residential storage-inverter sales. It also claims a 20.4% share of the C&I storage-inverter market, ranking second with around CNY 1.5bn in a market of roughly CNY 7.2bn. Those figures are internally consistent with Deye's annual report, but I could not replicate the exact 20.6% and 20.4% shares from an independent public dataset. They should be read as company-cited third-party estimates, not independently established market-share facts.

The broader independent data support the direction without proving the precise share. China-based Huawei and Sungrow dominate conventional PV inverters; Growatt, Ginlong/Solis and GoodWe are established global distributed-inverter names; and European residential storage has become increasingly price competitive. Deye's unusual result is that it has apparently built a much stronger position in residential storage than its ranking in broad PV-inverter shipments would suggest.

Residential storage itself sits at the intersection of several cycles. Subsidies, VAT treatment and export tariffs change payback periods, which makes it a policy cycle. It is an electricity-price cycle too, since the value of self-consumption rises when retail power prices rise. Distributors pre-buy hardware, so it is also an inventory cycle. Cell prices drive both system economics and hardware revenue per kWh, a technology and battery-cost cycle that has now turned from years of decline to renewed increases. And in weak-grid countries it is effectively an outage cycle: unreliable electricity can suddenly turn backup storage from discretionary to essential.

Europe captures the tension. SolarPower Europe says total European battery installations accelerated sharply in 2025, yet residential batteries continued to contract in the EU while large-scale systems drove growth. Separately, European rooftop solar suffered from lower subsidies and weaker household demand. A company can report strong "European storage" growth without the whole European residential category booming if it is gaining share, adding countries, selling larger systems or leaning into C&I. That nuance matters when interpreting Deye's current numbers.

Distributor economics explain why Chinese suppliers have taken so much share from Western residential names. A distributor wants acceptable reliability, local certification, rapid technical support, enough product variants to solve different grid configurations and a landed price that leaves margin for installer and wholesaler. Chinese manufacturers combine large component supply chains with shorter product cycles and aggressive pricing. Deye adds weak-grid functionality and low-voltage systems. Western vendors such as Enphase can still justify premium prices where module-level electronics, software, warranty reputation and installer familiarity are valuable, but the premium narrows when customers prioritise storage capacity per yuan. This conclusion is an inference from the cost, product and channel evidence rather than a single published market-share study.

The channel itself retains meaningful pricing power. An inverter maker may quote a wholesale FOB price, but the homeowner buys an installed system that includes distributor margin, installer labour, permits, switchgear, battery, commissioning and after-sales support. That dilutes the manufacturer's ability to translate end-user willingness to pay directly into inverter ASP. Deye's 2025 unit data, with storage-inverter ASP falling roughly 17% despite strong volume growth, are direct evidence that market-share gains and scale are being exchanged for lower unit monetisation.

The closest listed Chinese peers are GoodWe and Ginlong/Solis; Sungrow is an essential reference but is much more utility-scale and system-integration heavy.

Market-price data are as of approximately September 11, 2026; different vendors can calculate trailing earnings differently. Operating figures are company results or contemporaneous industry reporting.

Sungrow has become the large-scale system company Deye deliberately avoided becoming in its early years. Its advantage is bankability, utility-scale power electronics, large storage systems and project execution. In H1 2026, storage-system shipments reached about 25 GWh even as group revenue fell 29%. That mix makes Sungrow more exposed to utility project timing and system-level price competition, while Deye remains much more concentrated in residential and smaller C&I channels. Investors should not interpret Sungrow's lower P/E mechanically as evidence that Deye is expensive; the earnings cycles and capital intensity differ.

GoodWe is the cleaner residential/C&I comparator. It sells grid-tied and storage inverters through many of the same global installation channels. Its H1 2026 revenue rose 53% to CNY 6.25bn and it returned to a CNY 286m profit after the prior-year loss. It sold about 181,700 storage inverters in H1. Deye sold more than three times as many residential plus C&I storage units and generated a net margin almost six times GoodWe's roughly 4.6%. Deye's premium economic quality is real; the question is how much survives a normalised pricing environment.

Ginlong, operating as Solis, became a globally distributed string-inverter specialist and is now pushing hard into storage. Its H1 storage-inverter revenue more than doubled to CNY 1.64bn, already 42.2% of revenue, while conventional grid-tied inverter revenue fell 42%. That is arguably the clearest listed evidence that Deye's profitable niche is attracting direct competition from established distributed-PV channels. Ginlong's storage acceleration is strategically more relevant to Deye than Sungrow's headline revenue decline.

Enphase occupies a different premium niche. Its microinverter architecture places conversion electronics at each panel, and the company has spent years building installer relationships, batteries and home-energy software around that architecture. Customers pay for module-level control, safety, monitoring and a tightly integrated ecosystem. Its weakness is cost and regional concentration. Enphase disclosed a 29% sequential decline in European revenue in Q4 2025 as demand softened, after several years in which inventory corrections repeatedly distorted reported shipments. That history is a direct warning against treating Deye's distributor shipments as equivalent to final demand.

SolarEdge is the stronger cautionary analogue. Its optimizer-plus-string-inverter model once carried premium residential-solar economics and large European exposure; the subsequent collapse in demand and channel inventories produced heavy losses, and the stock currently trades on negative trailing earnings. Its experience shows that installed base, certifications and installer familiarity do not prevent a hardware franchise from suffering severe earnings compression when distributors destock at the same time competitors cut price.

Huawei, Growatt and Sigenergy matter despite being unlisted. Huawei brings scale, power electronics and integrated smart energy. Growatt has long competed aggressively in mass-market residential channels. Sigenergy is pushing premium all-in-one storage architecture and could pressure Deye from the opposite direction: a Chinese cost base married to a more integrated premium product. These private-company competitive threats are harder to quantify because audited segment disclosures are limited; they belong in the moat analysis, not in the listed-peer valuation table.

Deye's ecological niche is clear. It is a high-share challenger in residential and small-C&I hybrid power electronics, particularly strong where price sensitivity, backup power, low-voltage storage and flexible grid/generator operation matter. It takes profit pool from premium Western residential electronics and from generic Chinese inverter vendors whose products are less adapted to weak grids. Its own profit pool is most vulnerable to Chinese competitors that can combine comparable low-voltage functionality with batteries, local support and lower prices.

Governance deserves a separate discount. Zhang Hejun and his family control roughly 60.3% of voting rights through direct and controlled holdings. Zhang remains chairman and effective controller, while his son Zhang Dongye is vice-chairman, general manager and legal representative; another son, Zhang Dongbin, serves as a director. Press reporting describes an ongoing generational transition, with operating authority increasingly concentrated in Zhang Dongye.

Family control is economically aligned but institutionally concentrated. The family has enormous financial exposure to the equity and therefore benefits from dividends and long-run value creation. The same structure weakens outside shareholders' practical influence over succession, board composition and capital actions. IPO-era reporting also highlighted pre-restructuring fund-management irregularities involving founder-family accounts in 2017. I found no comparable material issue in the recent filings, and the 2025 annual financial statements received an unqualified audit opinion from Lixin, but that history supports a governance discount rather than a governance premium.

Recent capital allocation is shareholder-friendly in cash terms. The 2025 distribution package included a cash dividend of CNY 1.80 per pre-capitalisation share plus a 10-for-4 capitalisation. On the enlarged share count, the CNY 1.80 declaration is economically equivalent to about CNY 1.29 per current share. Total 2025 distributions including interim dividends and buybacks reached about CNY 2.74bn, 86.4% of attributable profit. H1 2026 then added a proposed CNY 1.60 current-share dividend.

On September 8 Zhang Hejun proposed a new CNY 100m–200m A-share buyback, disclosed September 9. The shares would be used for employee ownership or equity incentives rather than immediately cancelled, so this should be viewed primarily as compensation/incentive funding rather than pure share-count reduction. The size is also modest at roughly 0.1%–0.2% of the current market capitalisation.

Current fundamentals, valuation, and risks

H1 2026's most striking feature is that the income statement, cash-flow statement and balance sheet tell three different parts of the same story. The income statement says demand is surging. Gross profit reached about CNY 4.01bn on CNY 10.64bn revenue, a 37.6% gross margin, while selling, administration and R&D expenses grew much more slowly than revenue. The cash-flow statement says collections were strong. The balance sheet says the company and its supply chain simultaneously built inventory very aggressively.

H1 cash flow was excellent in headline terms but partly supplier-financed, so 1.54x cash conversion should not be annualised. Operating cash flow of CNY 4.195bn was 154% of net income, but the reconciliation shows inventory consumed CNY 2.078bn and operating receivables consumed CNY 902m while operating payables contributed CNY 4.071bn. Bills payable had risen to roughly CNY 4.66bn and accounts payable to roughly CNY 5.00bn by period end. The operating cash surplus depended materially on supplier credit expanding with the production ramp.

That is not inherently low-quality cash flow. A manufacturer with bargaining power should use supplier terms. The concern would arise if inventory later needs to be discounted while those payables fall due. Finished goods almost doubled in six months and raw materials nearly tripled. The inventory write-down recorded in H1 was only about CNY 8.3m, so management currently sees very little obsolescence. The investor's job is to watch whether future sales validate that judgement.

Receivables are less alarming than the inventory build. Gross sales nearly doubled while accounts receivable rose 48%, meaning receivable intensity actually improved on an annualised-revenue basis. Most foreign-currency accounts receivable were denominated in USD: the H1 note shows about CNY 2.37bn equivalent of USD receivables, plus much smaller EUR and AUD balances. That is why a stronger renminbi can hit reported financial income even when collection quality is sound.

The balance sheet remains financially strong despite the working-capital expansion. H1 attributable equity was CNY 11.49bn. Cash and cash equivalents were about CNY 4.97bn, and the company also held more than CNY 4bn of trading financial assets, while short-term borrowing had declined from roughly CNY 3.30bn at year-end 2025 to about CNY 1.88bn. Liquidity risk is low in the conventional sense. The more meaningful balance-sheet risk is inventory economics, not solvency.

The five-year cash record is broadly supportive. Using the annual-report cash-flow statements on a rounded basis, cumulative 2021–2025 operating cash flow was roughly 1.1 times cumulative attributable earnings. The exact annual conversion rate moved sharply with distributor receipts, inventory and supplier credit, but there is no long-run pattern of reported profit systematically failing to become cash. The unusually high H1 2026 conversion should nevertheless be normalised because of the CNY 4.07bn payable contribution.

Maintenance capital expenditure is small relative to the current earnings base. H1 purchases of property, plant, equipment and intangibles were CNY 492m, while depreciation and amortisation in the cash-flow reconciliation were about CNY 153m. With C&I, Malaysia and logistics capacity actively under construction, I estimate maintenance capex at roughly CNY 150m–200m for the half and growth capex at roughly CNY 290m–340m. This split is an analytical estimate, not management disclosure.

On that basis, H1 owner earnings were close to accounting profit: CNY 2.716bn net income plus roughly CNY 153m D&A, less perhaps CNY 150m–200m of maintenance capex, gives approximately CNY 2.67bn–2.72bn. The gap from reported earnings is well below the framework's 30% threshold for abandoning P/E in favour of a purely owner-earnings valuation. Normalised P/E and owner-earnings yield can be used together.

The harder adjustment is cyclical rather than accounting. Annualising H1 would give more than CNY 5.4bn of attributable earnings, but that would assume the strongest revenue-growth period in Deye's listed history persists without channel digestion or price pressure. I do not use that assumption in the valuation. My scenarios instead ask what Deye could earn in a normalised 2027 after the current order wave has either consolidated or corrected.

Reported trailing valuation is about 24.9 times earnings at CNY 85.50. That is above Sungrow's roughly 16 times, below the current trailing multiples of recovering GoodWe and Ginlong, and dramatically below Deye's own rough 44 times trailing earnings at the May 2026 price high. The market has already normalised a meaningful amount of the exuberance without pricing the company like a no-growth industrial.

I do not assign a spurious numerical historical P/E percentile. Multiple free market-data series handle Deye's bonus issues and earnings restatements differently, and the company's business mix changed so much after listing that a 2021 P/E is not economically comparable with today's storage-heavy company. The defensible historical observation is that the current multiple is far below the 2026 peak re-rating while the current earnings base is much larger.

Peer comparison also argues against simple relative valuation. Sungrow is cheaper but more utility/system exposed. GoodWe and Ginlong are more expensive on trailing earnings because their current earnings are depressed. Enphase remains profitable but has a very different U.S.-weighted microinverter model, while SolarEdge's trailing earnings are negative. Deye deserves a premium to a commodity inverter manufacturer if its 50%-plus storage-inverter gross margin and global niche share persist; it does not deserve a software multiple because distributor switching costs and hardware ASPs remain competitive.

The absolute valuation below uses normalised owner earnings and P/E rather than capitalising H1 growth. Revenue assumptions are deliberately below a straight-line extrapolation of the first half in the conservative case.

†Annualised-return estimates include illustrative cumulative dividends of roughly CNY 6, CNY 8 and CNY 10 respectively over three years. They are scenario calculations, not forecasts or investment advice.

The conservative scenario assumes the current wave proves partly cyclical. Revenue in 2027 would settle materially below a full annualisation of H1 2026, battery mix would remain high, and storage-inverter margins would normalise downward. A 20–22 times multiple still gives Deye credit for its balance sheet, channel, technology and long-term storage market, rather than treating it like a commodity appliance manufacturer.

The base case assumes that 2026 contains some restocking but that geographic diversification, C&I growth and battery attachment leave Deye structurally larger afterward. Revenue around CNY 19bn and owner earnings around CNY 4.6bn would mean the company retains much of its 2026 scale without requiring another year of near-doubling. A 24–26 times normalised earnings range is demanding for hardware but defensible if inverter gross margin remains well above peers and cash conversion remains healthy.

The optimistic case assumes Deye converts its distributor footprint into a genuine integrated-storage platform. Residential demand stays healthy across several regions, the 7 GWh C&I line ramps successfully, battery attachment rises without destroying margin and Deye preserves a storage-inverter margin advantage despite competition. A 27–29 times multiple then rewards above-industry growth and high returns on equity.

At CNY 85.50, the market is paying for moderate continued growth, not for another 92% half, but it still offers no conservative margin of safety. Current price is roughly 38%–50% above the conservative CNY 57–62 fair-value range and sits inside the broader base hold band. The stock is no longer priced as though H1 growth continues indefinitely; it is still priced above what I would pay for the business if 2026 proves to be another channel peak.

Another way to express the expectation embedded in the price is to work backward. Assuming roughly CNY 5.0bn of 2026 normalised earnings, a 20 times exit P/E in 2029, cumulative dividends of about CNY 7 per share and a required 10% annual equity return, Deye would need about CNY 6.8bn of 2029 earnings. That is an earnings CAGR of roughly 11% from CNY 5bn. With an 18 times exit multiple, required earnings growth rises to roughly 15%; with 22 times it falls to roughly 7%. The current price requires healthy medium-term growth, but nothing resembling H1's 79% profit growth. These are model outputs rather than market consensus estimates.

The most fragile base-case assumption is the roughly 25% normalised net margin. Reduce that assumption to 70%, or about 17.5%, while leaving revenue and the valuation multiple unchanged, and indicative value falls toward roughly CNY 61–66 per share. That exercise captures the real danger of a hardware business: a few percentage points of gross-margin compression can destroy more equity value than modest unit-volume growth creates.

The flat-earnings test is similarly sobering. If earnings stop growing for three years and the P/E is unchanged, capital appreciation is zero and the return becomes essentially the cash dividend yield. With a 50%–70% payout on an earnings yield of roughly 4%, that is only about 2%–3% a year before any multiple movement. I did not obtain a same-date official Chinese ten-year government-bond quote within the research source set, so I do not make a false-precision comparison with the risk-free rate. The important point is that a flat-earnings outcome produces a thin equity risk premium.

The margin-of-safety verdict is: none. Current price is above conservative value, the base case relies on unusually high hardware margins persisting, and the strongest recent cash-flow figure contains a large supplier-financing component. Deye can still generate a good return from here if the base or optimistic case is right; that is expected-growth return, not margin-of-safety return.

The bull-bear disagreement concentrates in a few observable variables rather than broad arguments about the energy transition.

Bulls have real evidence. Deye's residential and C&I storage franchise is producing extraordinary volume growth; customer concentration is falling; battery attachment is converting installed inverter channels into additional revenue; C&I storage inverter revenue rose about 75% in 2025; and H1 cash generation exceeded accounting earnings. A broader geographic footprint means the company is less dependent on any one South African or Pakistani cycle than it was several years ago.

Bears also have real evidence. Storage-inverter ASP fell about 17% in 2025 even as volume surged; battery gross margin fell almost ten percentage points; company inventory more than doubled in six months; Q2 gross margin softened to about 35%; and competitors such as Ginlong are now seeing their own storage-inverter revenue more than double. Those are exactly the conditions under which a category can keep growing while shareholder economics normalise.

The first permanent-loss risk is distributor and country-cycle reversal. I assign medium-high probability and high impact over a three-year horizon. If H2 shipments leave distributors with excess stock, 2027 orders can fall much faster than installations because customers work down inventory. Deye's own finished-goods build would then collide with lower distributor orders. Revenue falls first, price concessions follow, gross margin contracts through factory under-utilisation, and the market reclassifies the company from secular storage winner to cyclical hardware vendor. The observable indicators are inventory/revenue, receivables growth, quarterly inverter revenue, distributor commentary and customer concentration.

The second is price and mix compression, with high probability and high earnings impact. The evidence is already present: 2025 storage-inverter revenue per unit fell about 17%, battery-pack revenue per unit about 6.5%, and battery-pack gross margin was 31.8% versus 51.1% for storage inverters. If battery attachment keeps rising while competitors compress hybrid-inverter prices, Deye can report strong revenue while consolidated gross margin falls. The most important indicator is gross profit dollars and segment margin rather than shipment growth alone.

The third is battery commoditisation and warranty exposure. Probability is medium and impact could be high. Battery packs are becoming almost half of sales while Deye depends on external cell suppliers. Cell prices have turned upward in 2026 after years of decline, so the live exposure is margin compression from rising input costs rather than a write-down of inventory bought at high cost, and any systemic battery-safety or BMS defect creates warranty, recall and brand risk far beyond the pack's current margin contribution. The two largest disclosed battery suppliers already represent roughly 17.5% of 2025 procurement together.

The fourth is FX, trade and geopolitical exposure. Probability is medium and impact medium to high. Overseas sales represented about 80% of 2025 main-business revenue, and H1 2026 already showed the earnings sensitivity of USD/EUR weakness against the renminbi. Direct U.S. sales appear relatively small, but new origin rules, European trade measures or sanctions compliance can affect distributors, components and logistics even when the final customer is elsewhere. Malaysia reduces manufacturing concentration without eliminating those risks.

The fifth is governance and succession. I assign medium probability and medium impact, with tail risk higher. The founder family controls about 60% of votes and two sons sit in management or the board. The succession path appears increasingly centred on Zhang Dongye, which reduces uncertainty over who operates the company but does not create minority-shareholder checks. A poor H-share capital raise, aggressive related-party transaction or succession dispute would affect the governance multiple before it necessarily affects operating profit.

The sixth is valuation compression. Probability is medium-high because it does not require an operating crisis. A 25 times earnings stock can fall materially if storage hardware is rerated toward Sungrow-like high-teens multiples. If normalised earnings were CNY 4bn and the market assigned 18 times, equity value would be about CNY 57 per share even without balance-sheet distress. The transmission is simple: slower growth reduces both the earnings numerator and the multiple investors will pay for it.

Catalysts, tracking, and cross-synthesis

The next twelve months are primarily an evidence-gathering period. The positive catalyst with the greatest information content would be two or three quarters in which distributor sell-through remains strong while Deye's inventory growth slows. That combination would show that H1's build was preparation for real installations rather than simply inventory migration from Deye to wholesalers.

A second positive catalyst would be C&I storage scaling without residential-margin dilution. In 2025 C&I storage-inverter revenue reached about CNY 1.47bn and rose roughly 75%, while the new 7 GWh C&I line is due to absorb capital that had previously been allocated to other inverter projects. C&I can widen the addressable market and reduce dependence on household subsidy cycles, but the economics need to be demonstrated after the new capacity begins contributing.

A third is successful Malaysia ramp. Producing 3 GW of inverter capacity and 200,000 battery systems outside China could improve delivery times and give Deye another manufacturing origin for certain markets. A fourth is the proposed CNY 100m–200m buyback becoming an executed board programme. The signalling value would be stronger if shares were ultimately cancelled; the current proposal envisages employee incentives, so its direct per-share accretion is limited.

A sensibly priced H-share offering could also be constructive because it would diversify the shareholder base and finance international capacity. The opposite is equally possible: a large offering at a steep discount would crystallise dilution and signal management's willingness to raise equity despite large internal cash generation. Because no terms have been published, I assign no value to the event today.

The principal negative catalysts are easier to identify: a Q3/Q4 gross-margin fall into the low 30s, inventory continuing to outgrow revenue, a sharp slowdown in battery-pack orders, a major distributor destocking announcement, new European or emerging-market subsidy reductions, or another period of renminbi appreciation that turns overseas revenue growth into FX losses.

The tracking dashboard should focus on economic sell-through rather than headline shipment growth.

The next earnings date is currently shown by market-data calendars as October 26, 2026; the exchange filing should be treated as definitive once the company publishes the formal reporting schedule.

Inventory/revenue and gross margin are the two most useful indicators. Rising inventory can be benign ahead of a demand surge, but rising inventory combined with falling margin tells a much darker story: the company is producing into weakening price. Receivables are the second cross-check. If revenue slows while receivable days lengthen, distributors may be accepting product only because credit terms are becoming easier.

Storage-inverter ASP is valuable even though Deye publishes unit and segment data only periodically. A further 15%-plus price decline would imply that volume growth must remain extremely strong merely to preserve gross profit. Customer concentration is the channel-health check: rising concentration would suggest the recent broadening is reversing.

Looking vertically across the whole corporate history, Deye has proven one capability more convincingly than any other: it can transfer a manufacturing and power-electronics cost culture into a new adjacent product category and commercialise that category globally. It moved from injection-moulded parts and appliance controls into heat exchangers, dehumidification, inverters and then battery-integrated storage. The leap from Midea supplier to a company earning more than CNY 2.7bn in six months was too large to attribute to a single subsidy cycle.

Its past success was nevertheless a combination of skill and unusually favourable era tailwinds. Management deserves credit for avoiding the most crowded utility inverter battlefield, choosing low-voltage storage, entering weak-grid countries early and building a distributor network before residential storage became mainstream. It also benefited enormously from circumstances no management team created: Europe's 2022 energy shock, persistent outages in South Africa and Pakistan, falling lithium-battery costs and government support for distributed storage. Both statements can be true at once.

Those tailwinds are still present in altered form. Electricity security remains politically important, batteries are cheaper, storage systems are more capable, and C&I customers increasingly have a real economic use for peak shaving and backup. The geographic driver is broader than the original South African/Pakistani playbook. Yet the historical lesson remains intact: end-market economics can change faster than installed manufacturing capacity.

Horizontally, Deye's advantage over GoodWe and Ginlong is visible today in margin and scale. Its H1 net profit was nearly ten times GoodWe's despite revenue less than twice as large, and its storage-inverter volume exceeded the closest listed distributed-inverter peers. Against Enphase and SolarEdge the edge is more about cost and product flexibility. Sungrow and Huawei beat it on enterprise-scale bankability and breadth in large utility systems. And against the new all-in-one challengers, its weakness is that the battery pack is still more an attached product than a proven high-moat franchise.

Current valuation rewards both proven capability and future execution. At roughly 25 times trailing earnings, investors are no longer paying the roughly 40-plus-times multiple embedded in the May high, yet the stock still assumes that Deye's normalised earnings settle far above the 2025 CNY 3.17bn level. A return to CNY 3bn–4bn earnings would make the current price expensive; sustained CNY 5bn–6bn earnings would make it reasonable.

The market may currently be underestimating two things in opposite directions. It may be underestimating how much C&I and battery attachment can raise the revenue generated from an existing distributor relationship. A distributor that already sells Deye inverters can add batteries and larger commercial systems without rebuilding the channel from zero. That is a genuine economic advantage.

It may simultaneously be underestimating how quickly battery mix and inverter price competition can dilute margins. Battery packs went from a secondary line to almost half of H1 main-business sales in little more than a year. A company can double sales and still disappoint shareholders if the incremental CNY of revenue earns much less than the historical inverter CNY. The Q2 margin decline is the first metric that needs to stabilise.

For the next year, the most important variable is sell-through: whether Q3 and Q4 revenue, inventory and margin collectively confirm H1 demand. For three years, it is whether Deye can establish a profitable C&I and integrated-battery franchise without sacrificing its inverter economics. For five years, it is whether Deye remains a differentiated power-electronics platform or converges into a low-margin hardware assembler as inverter functions standardise and cell makers or larger energy-system companies integrate downstream.

The company becomes a much better investment under either of two conditions. The first is price: a fall into the high-CNY-40s would give a 20%-plus discount to my conservative normalised value. The second is evidence: if the company sustains more than 20% organic revenue growth through the channel reset, keeps gross margin above roughly 35%, brings inventory/revenue back below 20% and grows C&I without increasing receivable intensity, conservative value itself should rise.

The original judgement should be overturned negatively if Deye's storage-inverter advantage proves temporary: two quarters of sub-33% group gross margin, ongoing double-digit ASP declines, inventory still growing materially faster than sales and storage-inverter revenue falling would show that 2026 was largely a cyclical crest. It should be overturned positively if integrated-storage growth persists after distributor inventory normalises while margins remain in the mid-to-high 30s.

Bull reasons, each tied to earlier evidence:

• H1 2026 revenue and attributable profit rose 92% and 79%, while customer concentration continued to fall, showing the current growth wave is broader than a single large distributor.
• Storage-inverter gross margin was 51.1% in 2025, providing an unusually large profit cushion relative to conventional inverter hardware.
• Battery attachment has transformed an existing inverter channel into a second revenue stream: H1 battery revenue rose 244% to CNY 4.89bn.
• C&I storage inverter revenue grew about 75% in 2025 and new 7 GWh capacity is being built, reducing dependence on residential subsidy cycles if execution is successful.

Bear reasons:

• Storage-inverter revenue per unit fell about 17% in 2025, showing that Deye's rapid volume growth comes with heavy price/mix deflation.
• Battery packs carry roughly 32% gross margin versus 51% for storage inverters, so the fastest-growing line mechanically lowers business quality unless scale or integration closes the gap.
• Inventory more than doubled to CNY 3.85bn in six months and finished goods nearly doubled, while public filings do not disclose aggregate distributor inventories.
• Ginlong's H1 storage-inverter revenue rose 107% and GoodWe returned to strong overseas growth, proving competitors are entering the same profit pool rather than conceding it to Deye.

A three-year pre-mortem gives the risk concrete form. Script one: during 2027, GoodWe, Ginlong, Growatt, Huawei and Sigenergy expand low-voltage and all-in-one storage offerings across Europe, Australia and emerging markets. Wholesale hybrid-inverter prices fall another 20%–25%. Deye follows to defend channel share, storage-inverter gross margin falls from the 2025 level of 51% toward 38%–40%, battery packs remain around the mid-20s, and consolidated net profit normalises around CNY 3bn. If the market simultaneously cuts Deye from about 25 times earnings to 15 times, equity value is roughly CNY 45bn, or about CNY 35 per current share: almost 60% below CNY 85.50.

Script two: distributors over-order into the 2026 shortage and subsidy wave. Installations continue growing in 2027, but wholesalers spend three quarters reducing inventory. Deye's annual revenue falls roughly 25% from the boom run-rate, factory utilisation falls, inventory write-downs rise and receivable terms stretch. Profit drops 40%, while investors recall SolarEdge's and Enphase's channel corrections and apply an 18 times multiple. A CNY 3.5bn earnings base at 18 times is only about CNY 49 per share, before any severe governance or trade shock. The company survives comfortably because the balance sheet is strong; the shareholder still suffers a large permanent-looking loss if the entry price assumed boom earnings were durable.

My final synthesis is that Deye has earned the right to be treated as a serious global power-electronics company. The inverter franchise is not a marketing invention: the margin, volume, geographic reach and repeated product adaptation support it. The founder's manufacturing culture created a cost base that translated unusually well into low-voltage storage. The battery expansion is economically sensible because it monetises the same customer relationship twice.

The weak point is the distance between a good business and a predictable one. Deye operates in one of the most reflexive hardware markets in clean energy. High power prices or blackouts create urgent orders; distributors extrapolate demand; Chinese factories scale quickly; prices fall; inventories accumulate; then shipments can undershoot end installations while the channel clears. Deye has already lived through versions of that cycle. H1 2026 may prove structurally different because geography and C&I are broader, but the public data do not yet prove it.

My final rating is Hold. At CNY 85.50, the stock sits near the lower half of my CNY 74–108 base hold zone. The valuation no longer assumes H1's growth rate continues indefinitely, but price remains far above the level at which the conservative case offers a genuine margin of safety. For an existing holder willing to accept storage-cycle risk, the balance sheet, earnings quality and growth franchise support holding. For new capital, the asymmetry is weaker: a normalisation shock can take value toward the CNY 50s while the base three-year return is only mid-single-digit annualised.

The stock becomes materially more attractive only when either price falls into the high-40s or end-demand evidence proves that 2026 growth survives a distributor inventory reset. Waiting carries a real opportunity cost: if C&I and battery attachment make the current revenue level durable, the shares may never revisit that price. That is preferable to paying today for durability that has not yet been demonstrated.

【Company-profile scores】

• Fundamental quality: high
• Growth: high
• Moat: medium
• Financial soundness: strong
• Management credibility: medium
• Valuation attractiveness: medium
• Risk level: high
• Suitable investor type: cyclical

【Investment rating】

• Rating: Hold
• One-line thesis: Deye's storage franchise and cash generation are real, but CNY 85.50 already prices sustained post-boom earnings while channel inventory remains unobservable.

【Ideal Buy Price】45–49 CNY

Basis: at least a 20% margin of safety below the CNY 57–62 conservative normalised value derived from roughly CNY 3.6bn owner earnings and a 20–22 times multiple.

• Acceptable hold price: 74–108 CNY
• Clearly overvalued price: 138–147 CNY
• Current-price classification: acceptable hold
• Whether to wait for a better price: yes. The preferred entry is CNY 45–49; alternatively, a higher entry would require at least two post-H1 quarters with gross margin above 35%, inventory/annualised revenue below 20% and organic revenue growth above 20%.
• Target holding horizon: 3–5 years
• Expected annualized return: conservative about -7%; base about 6%; optimistic about 19%, including the scenario dividend assumptions stated above.
• Max-loss risk: approximately 55%–60% in the pre-mortem case where earnings normalise near CNY 3bn and the P/E compresses to roughly 15 times.
• Reassessment-trigger signals: two consecutive quarters below 33% gross margin; inventory/annualised revenue above 25%; storage-inverter revenue declining for two consecutive reporting periods; receivables growing more than 20 percentage points faster than revenue; or an H-share financing whose dilution exceeds 10% at a material discount to the A-share line.

This is valuation-scenario analysis within a research framework, not investment advice.

【Valuation Range】

• current: 85.50 (close as of 2026-09-11)
• bear (conservative · ideal buy zone): [45, 49]
• base (fair · acceptable hold zone): [74, 108]
• bull (optimistic · above the clearly-overvalued line): [138, 147]

Sources and research uncertainties

The most important primary source is Deye's 2026 interim report. It provides the CNY 10.641bn revenue, CNY 2.717bn attributable profit, CNY 4.195bn operating cash flow, balance-sheet working-capital movements, R&D spending, current share count and proposed CNY 1.60 interim dividend. The report also supplies management's current description of the inverter-plus-battery strategy and H1 inverter volumes.

The second is the 2025 annual report, including its audited segment economics, unit shipments, geographic split, customer/supplier concentration, cash-flow data and capital distribution. This is the basis for the 51.1% storage-inverter gross margin, 31.8% battery margin, CNY 9.75bn overseas main-business revenue and 2025 volume-price calculations.

The July 28 HKEX application proof is particularly valuable because it gives a more granular description of Deye's history, distribution model, customers, suppliers, founder control and Frost & Sullivan market-share work. It remains a draft document; all offering terms are redacted and its market-ranking data should be distinguished from independent industry statistics.

The September 9 buyback proposal is a company announcement rather than market speculation. It states that Zhang Hejun proposed CNY 100m–200m of repurchases funded from company cash, with the shares designated for employee ownership or equity incentives and a proposed implementation window of up to twelve months after board approval.

Independent industry cross-checks come primarily from Wood Mackenzie for broad inverter concentration and SolarPower Europe for European battery deployment and residential-storage cyclicality. They support Deye's industry direction but do not independently reproduce Frost & Sullivan's 20.6% residential-storage-inverter or 20.4% C&I-storage-inverter share estimates.

Peer cross-checks use H1 2026 operating disclosures and contemporaneous market data. GoodWe's recovery, Ginlong's rapid storage-inverter expansion and Sungrow's much larger utility-storage system business provide useful boundaries around Deye's earnings quality. Enphase and SolarEdge disclosures provide the most useful Western evidence on how residential channel inventory can overwhelm apparently durable installed-base franchises.

Five blind spots remain material.

First, Deye does not disclose H1 2026 revenue by individual country. The prospectus identifies countries that exceeded 5% in historical periods, but it cannot tell us precisely what percentage of the current boom comes from Germany, Ukraine, Pakistan, India, South Africa, the Middle East or Southeast Asia. That prevents a rigorous country-by-country installed-base demand model.

Second, aggregate distributor inventory is not disclosed. Company inventory is observable and Deye says it monitors distributors, but wholesale shipments cannot be reconciled with end-user installations on a consolidated basis. This is the single largest uncertainty in deciding whether H1 growth represents demand pull or channel restocking.

Third, H1 segment gross margins and battery-pack unit volumes are not disclosed at sufficient granularity to separate price, capacity mix and physical volume in the 244% battery revenue increase. The 2025 decomposition holds; the 2026 decomposition is necessarily incomplete.

Fourth, the HKEX prospectus anonymises Deye's largest battery suppliers and does not disclose detailed long-term cell-price formulas. The likely identities inferred from headquarters and listing descriptions should not be treated as confirmed.

Fifth, H-share offer size, price, A-share discount and dilution remain unknown because the latest application proof is still redacted. There is no legitimate way to add prospective H-share cash to current equity value or calculate dilution at the research base date.

Other tickers mentioned

• 300274.SHE: Sungrow, the larger utility-scale inverter and storage-system reference peer.
• 688390.SHG: GoodWe, one of the closest listed residential and C&I hybrid-inverter competitors.
• 300763.SHE: Ginlong Technologies, whose Solis brand is rapidly expanding from string inverters into storage.
• ENPH.US: Enphase Energy, the premium microinverter and residential-battery benchmark for Western markets.
• SEDG.US: SolarEdge Technologies, a residential inverter and optimizer peer whose inventory downturn illustrates channel-cycle risk.
• 000333.SHE: Midea Group, Deye's historically important heat-exchanger customer and procurement counterparty.
• 300014.SHE: EVE Energy, a listed Huizhou battery producer whose description is consistent with one anonymised supplier, although identification is unconfirmed.
• 300438.SHE: Great Power, a listed Guangzhou battery producer whose description is consistent with another anonymised supplier, although identification is unconfirmed.

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.