Rating: Neutral | Target Price: RMB 225–252 (base-case fair value range) | Current Price: RMB 248.96 (close, 2026-10-23) | Margin of Safety: -9.6% (none) | Time Horizon: 12 months
Kingsoft Office is the undisputed domestic leader in office software, with an excellent subscription-based core: 2026H1 revenue of RMB 3.313 billion (+24.69%), accelerating from FY2025 (+15.78%); adjusted net profit of RMB 1.087 billion (+28.29%); gross margin of 84.9%; net cash of RMB 15.68 billion (13.6% of market cap); and operating cash flow consistently above net profit. However, starting in 2026, reported profit has been severely polluted by external investment funds — of the RMB 2.518 billion H1 net profit attributable to shareholders, investment income accounted for RMB 1.914 billion (69.2% of total profit), and Q2 investment income alone turned negative. The controlling shareholder Kingsoft Corp's profit warning on October 12 pointed to large Q3 losses from fund profit-sharing, and the Q3 report due 10-28 may show a single-quarter reported loss. At RMB 248.96, the current price implies adjusted-basis PE(TTM) of ~64.5x and EV/2027E adjusted of ~39.5x, already pricing in "base case delivered + about half of the AI upside," sitting at the upper end of the base-case fair range (RMB 225–252) with no margin of safety. High quality, but not cheap — we assign a Neutral rating: wait for the Q3 report's true-basis cleanup or a pullback below the base-case midpoint (~RMB 225) before reassessing positioning.
Key evidence:
Within the "high growth" of +236.94% attributable net profit and +241.99% non-GAAP (recurring-excluded), equity-method fund income is the dominant driver — even the "recurring-excluded" basis is polluted (H1 recurring-excluded RMB 2.487 billion is nearly identical to attributable net profit; non-recurring items were only RMB 30 million). Two definitional disciplines must be emphasized: first, Kingsoft Corp's RMB 1.0–1.2 billion warning is at the group level and pre-tax; Kingsoft Office holds the same batch of funds via Shenzhen Shunying (58.82% stake) and Shunjin-Shunying (99.9986%), requiring pro-rata mapping — based on Q3 core-business single-quarter adjusted net profit of ~RMB 510–550 million, fund profit-sharing losses (after-tax, attributable) of RMB 600–700 million would trigger a single-quarter reported loss, with the exact magnitude to be confirmed by the "investment income" line in the 10-28 Q3 report; second, the company's "adjusted net profit" basis excludes fund gains/losses but also adds back share-based compensation (RMB 148 million added back in 2026H1), overstating "true core earnings including SBC" by ~14% — both bases should be shown side by side in valuation. Reported PE(TTM) of 32x, at the 8th percentile of the past 5 years, is an "undervaluation illusion" under investment-income pollution and cannot serve as a basis for buying.
Key evidence:
Structurally, subscription revenue (personal subscription + WPS 365) accounts for ~73% of revenue, with a prepaid model that is self-funding. Two reservations must be disclosed alongside: first, 2026H1 operating cash flow was RMB 643 million, -12.96% YoY (company explanation: drag from prepaid income tax on fund unrealized gains; positive YoY growth excluding this), so cash flow lagged the income statement in the period; second, current contract liabilities of RMB 2.764 billion were up only +6.4% from the start of the year, below revenue growth — prepayment momentum lagged revenue recognition (another ~RMB 1.24 billion of long-term contract liabilities sit in the other non-current liabilities note; the combined ~RMB 4.0 billion figure could not be directly verified this time — flagged as pending verification). WPS 365's verified growth trajectory is FY2025 +64.93%, 2026Q1 +60.79%, 2026H1 +60.84% — sustained above 60% but slowly decelerating, not accelerating.
Key evidence:
On Xinchuang: Document No. 79 is an existing policy issued in 2022; 2026–2027 is the assessment delivery window, not new policy, and the pace is subject to fiscal budgets (institutional licensing fell -8.56% in 2024). Still, WPS software's quarterly growth trajectory of 2025Q3 +50.52% → 2026H1 +33.36% supports the "deadline approaching, volume ramping quarter by quarter" thesis. On AI: the global precedent is Microsoft Copilot's paid-seat penetration of ~4.4% (conservative basis; ~5% per partner figures; 35% for qualified seats per management — wide divergence), and it took Microsoft about three years to reach 4–5%. Our AI upside estimate (48.25 million paying users × 10–20% penetration × incremental ARPU of RMB 50–100 = incremental annual revenue of RMB 240–970 million) is optimistic at the upper end versus global precedent; the base case takes only the low end. WPS AI MAU of 80.13 million is end-2025 data; the company stopped disclosing key AI operating metrics from 2026 and dodges monetization questions on its IR platform — evidence of the "inflection" at the company's P&L level remains blank, and the degree of delivery is the fork between the base case and the bull/bear cases.
Key evidence:
The math implied by the current price: solving backward at a 35x justified forward multiple, the market cap requires adjusted net profit of ~RMB 2.85 billion (~22.5% three-year CAGR vs. our FY2025 estimate of RMB 1.55 billion) — i.e., the market has priced in "base case delivery (2027E ~RMB 2.53 billion) + about half of AI upside"; yet reality is that Q2 single-quarter adjusted growth slowed to ~24.2% from Q1's 32.2%, and AI revenue has zero separate disclosure. The three-scenario probability-weighted midpoint of ~RMB 236.9 is ~4.8% below the current price. Two comparisons: sell-side consensus target of RMB 338.8 already exceeds our bull case of ~RMB 325 (partly anchored to reported EPS); the company's July–October 2026 buyback execution range of RMB 205.71–231.88 offers a company-action value reference floor. Base-case fair value is RMB 225–252; the current price sits at the top of the range with a margin of safety of -9.6%.
Key evidence:
Both directions of the data must be presented objectively: total global MAU of 676 million still grew +3.87%, but growth rests entirely on PC (+10.14%) and overseas (overseas PC MAU 47.83 million +38.11%; overseas personal revenue +43.79%) offsetting mobile contraction. ARPU, estimated as domestic personal revenue / period-end paying users, fell from ~RMB 74.3 (2024) to ~RMB 72.5 (2025), then annualized back up to ~RMB 76 in 2026H1 (+4.8%) — the AI membership tier is showing initial traction but only modestly (the H1×2 annualization understates the H2 peak season; directional reference only). The personal business is 61% of revenue; if mobile contraction spreads to PC and AI price increases are insufficient to offset, the consumer engine will degrade from "penetration + price hikes" to a single leg — one of the core transmission paths of the bear case.
Key evidence:
Word-deed consistency is judged "pragmatic": the 2026H1 preannouncement (revenue RMB 3.214–3.413 billion; adjusted RMB 1.000–1.174 billion) landed mid-range, and past preannouncements were broadly delivered. However, 2021–2026 marks six consecutive years of equity incentives (~0.2% annualized dilution), multiple executives reduced platform-level equity (FY2025 annual report basis), buyback shares are used for equity incentives rather than cancellation, and investors have publicly questioned the "limited buyback intensity" on the IR platform. Overall judgment: credible earnings delivery, neutral-to-weak governance.
| Metric | FY2023 | FY2024 | FY2025 | 2026H1 (latest period) |
|---|---|---|---|---|
| Revenue (RMB 100M) | 45.56 | 51.21 | 59.29 | 33.13 |
| Revenue YoY | +17.27% | +12.40% | +15.78% | +24.69% |
| Attributable net profit (RMB 100M) | 13.18 | 16.45 | 18.36 | 25.18 (+236.94%, distorted) |
| Recurring-excluded attributable net profit (RMB 100M) | 12.62 | 15.56 | 18.03 | 24.87 (+241.99%, distorted) |
| Adjusted attributable net profit (company basis, RMB 100M) | — | — | ~15.5 (estimated) | 10.87 (+28.29%) |
| Gross margin | 85.30% | 85.14% | 85.95% | 84.90% |
| Net margin | 28.92% | 32.12% | 30.98% | 75.99% (distorted) |
| Operating cash flow (RMB 100M) | 20.58 | 21.96 | 25.02 | 6.43 (-12.96%) |
| Free cash flow (RMB 100M) | 18.92 | 20.21 | 24.04 | 5.98 |
| Cash + cash-like assets (RMB 100M, period-end) | — | — | 148.70 | 156.81 (as of 2026-06-30) |
| Interest-bearing debt (RMB 100M, period-end) | — | — | ≈0.6 | 0.64 (as of 2026-06-30) |
| Debt-to-asset ratio (period-end) | — | 28.13% | 29.20% | 27.79% (2026-06-30) |
| Contract liabilities (RMB 100M, period-end) | 18.92 (current) | 30.79 | 38.47 | 27.64 current + ~12.4 long-term (pending verification) |
| R&D expense (RMB 100M) / expense ratio | 14.73 / 32.32% | 16.96 / 33.11% | 20.95 / 35.34% | 11.52 / 34.76% (H1) |
Reasons for metric changes (items with YoY ≥ ±20%, each with company explanation): ① 2026H1 investment income of RMB 1.914 billion (+1334.29%), attributable net profit +236.94%, recurring-excluded +241.99% — company explanation: valuation re-rating of investees of Shunjin-Shunying/Shunying funds (Zhipu, Fourier, etc. went public), with a self-warning that "single-quarter swings or even losses cannot be ruled out"; ② income tax expense +422.70% — increased deferred tax provision due to taxable temporary differences on investment income; ③ operating cash flow -12.96% — company cites higher employee compensation payments and prepaid taxes, "positive YoY excluding the prepaid income tax impact"; ④ FY2025 WPS 365 revenue +64.93% — "AI-centric upgrades across technologies and functions, accelerating expansion of benchmark clients"; ⑤ FY2025 WPS software licensing +15.24% (2024: -8.56%) — leading Xinchuang share, release of government procurement demand, reflecting tender recognition timing; ⑥ FY2025 accounts receivable +26.11% (low base; AR/revenue only 10.7%); ⑦ FY2025 non-current assets due within one year +4957.61% — long-term fixed deposits maturing in concentration, a cash structure change.
Business model: Asset-light subscription software — personal "free + premium subscription" (approximately 61% of revenue), institutional seat subscription (WPS 365, 12%), and government/enterprise licensing + maintenance (25%); subscription categories total approximately 73% of revenue, with strong recurring income and a prepayment model (total contract liabilities of approximately RMB 4.0 billion, of which RMB 2.764 billion current, verified). Pricing power is neutral-to-weak: blended gross margin rose steadily from 85.1% to 85.9%, but personal ARPU declined through 2024–2025 (RMB 74.3 → 72.5) and only recovered to approximately RMB 76 in 2026H1; price increases depend on delivery of new AI benefits.
Cash content of earnings: OCF/net profit attributable to parent for 2023–2025 was 1.56/1.33/1.36; FCF/net profit for 2024/2025 was 1.23/1.31 — book profits convert fully, even superfully, into free cash; cash received from sales/revenue is approximately 1.20. In 2026H1, OCF/net profit was only 0.24, but this is mainly due to fund mark-to-market gains inflating the net profit denominator and prepaid income tax (versus approximately 0.55 on the adjusted basis) — a definitional distortion rather than operational deterioration.
Recurring earnings test: Both attributable (RMB 2.518 billion) and non-recurring-excluded (RMB 2.487 billion) net profit for 2026H1 include approximately RMB 1.9 billion of equity-method fund gains; the true core-business figure is the adjusted RMB 1.087 billion — the non-GAAP (ex-non-recurring) measure has become invalid for this company. As a reverse test of the company's Non-GAAP looseness: the adjusted measure treats SBC as if it doesn't exist (added back RMB 148 million in 2026H1 and RMB 213 million in 2025); a core net profit including SBC would be approximately 14% lower, so investors should read both measures side by side. The fund exposure (long-term equity investments of RMB 3.468 billion as of end-2026H1) can recur in either direction every quarter until the funds exit: the same funds posted small losses in the prior-year period, large mark-to-market gains in 2026Q1, and turned negative in Q2 — "anchoring on the adjusted measure" does not mean fund risk is cleared.
Return on capital: ROIC approximately 14.9% (2025, calculated as non-recurring-excluded net profit / average attributable equity), above the ~9% reference WACC for A-share software — value-creating; excluding approximately RMB 15 billion of cash-like assets, operating ROIC is significantly higher. R&D of RMB 2.095 billion is fully expensed (zero capitalization) — conservative accounting.
CapEx quality: CapEx RMB 98 million < D&A RMB 107 million (2025), CapEx/depreciation 0.91; the Wuhan campus has entered its harvest period after being placed into service, with virtually no maintenance-capex burden — a classic cash-cow structure. No maintenance/growth capex split disclosed, but in software, maintenance spending consists mainly of R&D expense (already expensed).
Red flags and concentration: Top five customers are only 10.94% (low concentration); top five suppliers are 57.89% and the largest is a related party (procurement of RMB 290 million, 26.09% — Kingshin ecosystem). If AI compute and models (Xiaomi's MiMo already integrated) concentrate further with related parties, there are potential concerns on both cost bargaining power and governance optics.
FY2025 segments (2026H1 continues in the same direction):
| Segment | Revenue Share | Gross Margin | FY2025 YoY | 2026H1 YoY | One-Line Business Logic |
|---|---|---|---|---|---|
| WPS Personal (domestic subscription + overseas) | 61.2% | 80.96% | +10.42% (domestic +7.87%/overseas +53.67%) | +15.20% | Free acquisition + membership monetization, AI aids conversion |
| WPS 365 (institutional subscription) | 12.1% | 91.49% | +64.93% | +60.84% | Annual per-seat subscription; SOE/private enterprise penetration; fastest growth |
| WPS Software (institutional licensing/Xinchuang) | 24.6% | 95.02% | +15.24% (2024: -8.56%) | +33.36% | Government/enterprise licensing + maintenance; fluctuates with Xinchuang tender cycles |
| Others (advertising etc.) | 2.0% | 93.94% | -8.79% | +37.30% | Deliberately downsized; low share |
Profit engines: Estimated gross profit contribution by "revenue share × gross margin" — personal approximately 57.7%, institutional licensing approximately 27.2%, WPS 365 approximately 12.9%. Personal, the largest by revenue, is also the largest gross profit contributor, but the fastest-growing 365 and the highest-leverage licensing together contribute roughly 40% of gross profit — the main engine of profit growth over the next two years. Gross margin structure differs by 14.06 percentage points (licensing 95.02% vs personal 80.96%): licensing is standardized software licensing with near-zero marginal delivery cost; personal subscription includes channel splits, payment fees and customer acquisition costs (this is structural inference; the company does not disclose cost breakdown separately), with weaker pricing power than the government/enterprise side.
Accounting red flags:
Cross-period consistency:
Overall verdict: No hard traces of fabricated revenue/profit; cash flow and profit match well. The main issue is definitional (ex-non-recurring invalid, adjusted adds back SBC), not manipulation.
Current market data (2026-10-23 close): Share price RMB 248.96, market cap RMB 115.5 billion, total shares 464 million. Multiples and percentiles: apparent PE(TTM) 32.0x (8th percentile over 5 years — percentile for reference only; earnings are contaminated by investment income and mechanical percentiles are distorted); adjusted PE(TTM) approximately 64.5x (estimated); PB 7.66x (8th percentile over 5 years); PS(TTM) 17.5x; EV/2026E adjusted approximately 49.4x, EV/2027E base approximately 39.5x; PEG≈1.6 (EV/2027E ÷ 2027E growth of 24%; versus Microsoft's forward ~26x/13% growth, PEG≈2.0 — not more expensive than Microsoft on a growth-adjusted basis).
| Benchmark | Forward PE | Growth | Notes |
|---|---|---|---|
| Kingsoft Office (2027E adjusted) | 45.7x (39.5x on EV basis) | adjusted approximately 24% | 73%+ subscription, net cash RMB 15.68 billion, A-share scarcity |
| Microsoft (forward) | 25–27x | approximately 13% | Global office SaaS pricing anchor; Copilot precedent |
| Adobe (forward) | approximately 15x | approximately 10–12% | Software valuation floor suppressed by AI self-cannibalization fears |
| A-share software peers (Foxit/Weaver/Trio) | Near-zero profit or N/A | 0–11% | Highlights Kingsoft's scarcity of "growth + profitability" |
Market-implied expectations: Current price RMB 115.5 billion = net cash RMB 15.68 billion + core business EV RMB 99.8 billion. Back-solving at a reasonable 35x forward multiple, the market implies adjusted net profit of approximately RMB 2.85 billion (+84% versus our 2025 estimate of RMB 1.55 billion, a three-year CAGR of approximately 22.5%) — the current price requires the company to deliver RMB 2.85 billion in adjusted net profit by 2027–2028, while reality is 2026H1 of RMB 1.087 billion (+28.29%), Q2 already slowing to approximately +24.2%, and zero separate AI revenue disclosure. Two-way test of the AI demand math: if AI membership penetration truly delivers in the low-to-mid range of the elasticity band (RMB 240–970 million incremental revenue) plus Xinchuang volume, RMB 2.85 billion is achievable (upper-base); if one linearly extrapolates Microsoft's conservative precedent of 4–5% penetration over three years, 10–20% penetration of domestic personal willingness-to-pay looks optimistic, and the implied expectation cannot hold. Current pricing sits at roughly the "60th percentile between base and bull."
Three-layer value (EPV): Asset value approximately RMB 32.5/share (net cash RMB 15.68 billion ≈ RMB 33.8/share, net of lease liabilities etc.) — the floor. Zero-growth EPV approximately RMB 88.2/share (normalized FCFF RMB 2.4 billion / 517 million shares ÷ WACC 9.5% + net cash RMB 33.79; 2025 FCF of RMB 2.404 billion as the base, self-consistent with "adjusted net profit + SBC add-back ≈ OCF"; if SBC dilution cost is strictly excluded, the floor is approximately RMB 69). The growth option accounts for approximately 64.6% of the current price — nearly two-thirds of today's price rests on undelivered growth from "AI usage-based billing + second round of Xinchuang + internationalization," while the EPV floor is only one-third. This is a classic "market cap almost all future" name: growth weighting is not low, so the cheap-judgment rule of "market only pays for zero growth" does not apply; whether it is expensive depends on growth quality and delivery, answered by scenario math.
Three scenarios and odds (12 months):
| Scenario | Probability | Fair Range | vs. Current | Core Assumptions and Exit Multiple Anchor |
|---|---|---|---|---|
| Bear | 25% | RMB 145–160 (midpoint 152.5) | -38.7% | AI monetization falsified (zero AI revenue disclosure still in the FY2026 report, Lingxi targets missed) + adjusted growth falling below 15% (2027E approximately RMB 2.05 billion) × 26.5x (A-share software valuation band for growth <15%, versus Adobe's 15x AI-self-cannibalization floor + A-share liquidity premium) |
| Base | 52% | RMB 225–252 (midpoint 238.5) | -9.6%～+1.2% | Adjusted 2027E approximately RMB 2.53 billion (2025→2027 CAGR approximately 27.8%, subscription ramp + Xinchuang volume, low-end AI elasticity of RMB 200–400 million) × 35–40x (against Microsoft's 26x/13% PEG of 2.0, 24% growth warrants 35–40x, PEG 1.5–1.7; at the lower end of its own historical forward midpoint of 40–60x) |
| Bull | 23% | RMB 305–345 (midpoint 325) | +30.5% | "Subscription + AI usage" billing scaling in 2027 with AI revenue broken out separately; adjusted 2026–2027 CAGR approximately 39% → 2027E approximately RMB 3.0 billion × 45x (global AI application premium + A-share scarcity, anchored to the delivery path of the Copilot seat penetration precedent) |
Probability-weighted midpoint approximately RMB 236.9, about 4.8% below the current price — odds nearly symmetric, slightly skewed downward; the divergence between scenarios is driven mainly by the same set of AI penetration/ARPU assumptions, with limited independence (this limitation is disclosed as is). External coordinates: the sell-side consensus median target of RMB 338.8 already exceeds our bull case (partly anchored on apparent EPS; downside revision risk after the Q3 report); the lowest target of RMB 201 is still above our bear case, since sell-side floors typically only cut multiples without layering in AI-falsification earnings downgrades — not fully comparable on basis. The company's buyback execution range of RMB 205.71–231.88 serves as a company-action value reference. Fund profit-sharing losses are mark-to-market-linked volatility; the wild swings in apparent 2026 EPS (consensus 2026E 7.90 → 2027E 5.74) are precisely a reflection of this distortion.
Our own earnings forecast (reconciliation anchor): FY2026 revenue RMB 7.24–7.37 billion (+22~24%), adjusted net profit attributable RMB 1.97–2.09 billion (+27~35%; core net profit including SBC approximately RMB 1.0–1.15 billion); FY2027 revenue RMB 8.69–9.07 billion (+20~23%), adjusted net profit attributable RMB 2.44–2.57 billion (+21~27%). Comparison: sell-side consensus 2026/2027 revenue 70.46/82.51 billion (ours slightly higher, given H1 already +24.7%), apparent attributable net profit 36.29/26.33 billion (includes fund gains/losses; not used as an anchor).
Conclusion: Valuation is fair; score quality and price separately. Quality: gross margin 85%, subscription 73%, net cash RMB 15.68 billion, ROIC approximately 15%, full cash conversion — excellent. Price: 64.5x TTM on the adjusted basis, 39.5x EV/2027E, at the upper end of the base range, margin of safety -9.6% — not cheap. Target price RMB 225–252 (base scenario); the current price has fully bought the base case, with upside hinging on AI delivery (bull +30.5%) and downside on definitional cleanup and AI falsification (bear -38.7%). If the 10-28 Q3 report shows adjusted growth ≥25% with a smooth basis transition, or the share price falls below RMB 225, the odds turn favorable again.
1. Industry Size: The company operates in a triple-stacked market of "office suite + organization-level collaboration SaaS + AI office value-added services." Figures must be presented separately: the narrow-caliber China collaborative office platform market was approximately RMB 10.2 billion in 2023 (+28.1%, iResearch, first-tier institution source); the broad caliber (platform + adjacent services) was approximately RMB 33.01 billion in 2023 and an estimated RMB 41.48 billion in 2025 (iiMedia Research, implied CAGR ~12%); the global office software market was approximately RMB 457.78 billion in 2025 (QYResearch as cited by brokers—third-hand source, to be verified, for cross-reference only); "AI + office" was approximately RMB 30.86 billion in 2024, up 135.6% YoY (Industry Information Network aggregated caliber, low reliability, to be verified). Structurally, IDC projects China collaborative office SaaS share rising from 65.4% in 2024 to 77.6% in 2028—cloud-based subscription is irreversible. Traditional suites grow at single digits globally (Microsoft M365 Commercial cloud FY2025 +15%), with incremental growth almost entirely driven by two structural variables: "AI office" and "Xinchuang (IT application innovation) substitution." Quantified demand inflection chain (AI): The global anchor is Microsoft's M365 commercial paid seats exceeding 450 million, with Copilot paid seats rising from 15 million (Jan 2026) to 20 million+ (Apr 2026, +5 million in a single quarter), penetration of roughly 4.4% (conservative caliber; partner caliber ~5%, management's "qualified seats" caliber 35%—wide divergence). Copilot pricing at USD 30/month equals 2.4x the base seat price—AI adopters see ARPU uplift of 83–140%. On the company side: 48.25 million paying users × 10–20% AI membership penetration × incremental ARPU of RMB 50–100 = incremental annual revenue of RMB 0.24–0.97 billion (equivalent to 7–27% of FY2025 personal business); realization below the midpoint is part of the base-case scenario. Xinchuang chain: Document No. 79 requires 100% replacement of eight categories of systems at central SOEs by end-2027, with office suites among the earliest categories. Orders recovered from 2025Q3 (confirmed by the company's WPS software single-quarter growth of +50.52%), with 2026Q4–2027 as the concentrated procurement window.
2. Industry Chain and Value Distribution: Upstream comprises basic software/hardware (Kylin/UOS OS, Kunpeng/Phytium and other chip adaptations), cloud infrastructure (Alibaba/Tencent/Huawei Cloud + self-built), and the AI layer (in-house development + purchased domestic LLM APIs, with Xiaomi MiMo already integrated)—upstream concentration is high but office suites can switch among multiple suppliers, giving medium bargaining power; AI inference compute is a new cost item. The midstream is Kingsoft's segment (WPS Office / WPS 365 / WPS AI) and retains the highest gross margin across the entire chain (blended gross margin 85.95%, of which licensing 95.02%, 365 91.49%). Downstream comprises party and government bodies (Xinchuang centralized procurement), central/local SOEs (sector Xinchuang), private enterprises (365 subscription), and individuals (free + membership). Downstream bargaining power is differentiated: strong on C-end (free-tier barrier + paying users +15.44%), medium-strong with party/government and SOEs (very few catalog players), medium-weak with private enterprises (suppressed by DingTalk/Feishu free strategies).
3. Supply-Demand and Competitive Structure: Global suites are a duopoly (Microsoft + Google, CR2 > 90%). In China, WPS dominates suites alongside the Xinchuang long tail; third-party aggregated data put WPS licensed-market share at approximately 32–41%, first place, with Microsoft second at ~28.5% (confusing caliber, low reliability, to be verified), while the company's annual report claims "continued Xinchuang share leadership" (first-hand but self-reported)—by revenue, Microsoft remains first in China's high-end commercial market; both calibers are presented side by side. In collaborative platforms, a three-horse pattern exists—DingTalk 32.7% / WeCom 23.4% / Feishu 18.9% (per 36Kr, to be verified)—and none are profitable (DingTalk subscription ~RMB 3 billion+, Feishu ARR RMB 2.1 billion+). Entry barriers: 30 years of format-compatibility accumulation, Xinchuang catalog and security certifications, government/enterprise channel switching costs, network effects from 676 million MAU devices, and an annual R&D threshold of RMB 2.095 billion. Substitution threats: LLM-native conversational office (Doubao/ChatGPT directly generating documents) is the largest long-term variable, though AI office is currently still delivered via suites; Tencent Docs/Shimo siphon off light users.
4. Cyclicality and Regulation: Three layers of cycles overlap—mature traditional suites, mid-stage growth in subscription/cloud collaboration (company subscription mix at 73%), and early-stage transition from introduction to growth in AI office (single-digit penetration). Regulation is broadly favorable: Document No. 79 creates certain demand, data compliance raises barriers for overseas products, and the Generative AI Measures plus content-labeling rules (effective Sep 2025) raise compliance costs but benefit leading vendors that have completed filing; the main disturbance is fiscal budgeting rhythm causing volatility in party/government procurement (declined in 2023–2024, recovering from 2025). Leading indicators: central/local SOE Xinchuang centralized procurement tender volume, the company's WPS software/365 single-quarter growth rates, individual paying-user growth, and Microsoft Copilot quarterly seat additions (a global synchronous demand indicator).
5. Peer Benchmarking:
| Company | Revenue Scale | Growth | Gross Margin | ROE | Key Differences vs Kingsoft Office |
|---|---|---|---|---|---|
| Kingsoft Office | FY2025 RMB 5.929 billion | +15.78% (2026H1 +24.69%) | 85.95% | 15.2% (2025) | Full stack of subscription + licensing + AI; Xinchuang government/enterprise franchise |
| Microsoft (Productivity segment) | FY2025 ~USD 121.0 billion | +13% (M365 Commercial cloud +15%) | Not disclosed | 34.0% | Global No. 1; Copilot seat-based monetization at scale; Kingsoft is roughly 1/30 its size |
| Foxit Software | FY2025 RMB 1.067 billion | +49.99% (consolidation-driven) | 91.58% | ~0.5% | Pure PDF single product, overseas-focused; no suite or Xinchuang franchise |
| Weaver Network | FY2025 RMB 2.287 billion | -3.17% | Not obtained | 12.6% | OA workflow software, growth stalled—underscoring the scarcity of Kingsoft's subscriptions |
| Cailun (Racent) | FY2025 RMB 1.830 billion | +10.76% (2026H1 -24.33%) | ~35% | 10.1% | Carrier mailbox project-based customization; non-standard delivery with low margins |
6. Company Industry Positioning: The absolute domestic leader in office software and a second-tier global challenger. User base: 676 million global MAU (domestic ~430 million + overseas 244 million), No. 1 domestically; absolute dominance in institutional Xinchuang licensing (2026H1 added benchmark clients including CNNC and PetroChina); 48.25 million individual paying users with a ~7% payment rate and ARPU of ~RMB 76/year vs. Microsoft M365 seat prices of USD 12.5–36/month—a tenfold gap in customer spend that represents both a gap and AI-era pricing headroom. Share trends: rising on the institutional side (Xinchuang + 365 above 60% for six consecutive quarters), stable-to-rising on the personal side, slowly climbing globally (overseas personal +43.79%). Four layers of moat: 30 years of format compatibility, Xinchuang access qualifications, device network effects, and R&D scale. Weaknesses: organization-level collaboration ecosystem lags DingTalk/Feishu (WPS 365 revenue RMB 720 million vs. DingTalk subscription RMB 3 billion+ / Feishu ARR RMB 2.1 billion+), and no separate disclosure of AI monetization.
The decisive battleground for subscription software lies not in the income statement but in the "users → paying → ARPU → retention" funnel. Kingsoft Office's latest readings:
Rating: Neutral; confidence 0.58; time horizon 12 months. Kingsoft Office is a scarce A-share software asset with "true subscription + true profitability + net cash," and its core business quality—28% adjusted growth—is solid. However, the current price of RMB 248.96 (2026-10-23 close) has fully priced the base case (fair value RMB 225–252) and prepaid about half of the AI upside, leaving a margin of safety of -9.6% and odds slightly tilted to the downside. Strategically, neither chase nor short: the Q3 report on 10-28 is the nearest inflection point—if adjusted growth ≥25% and the reported loss shock is absorbed by the market as a "caliber switch" (rather than a double-kill), and the share price falls back below RMB 225, the odds turn favorable and the stance turns positive; if the Q3 report confirms continued suspension of AI data disclosure and adjusted growth falls below 20%, the probability of moving toward the bear-case path (RMB 145–160) rises.
Key monitoring items (next 6–12 months): ① Q3 report on 2026-10-28: adjusted growth, investment income line item, WPS 365/institutional licensing single-quarter growth, whether AI data disclosure resumes; ② 2026 annual report (~Mar 2027): whether AI revenue/AI membership numbers are separately disclosed for the first time, and whether "subscription + AI usage" billing lands; ③ Lingxi's promised "order-of-magnitude growth" in active users (verification Dec 2026 to 2027Q1); ④ Xinchuang orders: whether WPS software single-quarter growth returns to 40%+ (from 2026Q4) and the frequency of government/enterprise contract-win announcements; ⑤ overseas Agent product launch (management committed for H2 2026); ⑥ whether sell-side consensus switches to the adjusted caliber and cuts target prices after the Q3 report (current median RMB 338.8); ⑦ governance signals: buyback execution progress (plan of RMB 250–500 million), changes in directors'/supervisors'/executives' holding platform interests, and terms of the new 2027 incentive plan. Risk warnings: Q3 reported loss exceeding expectations, pressure from AI-native competitors (Workbuddy et al. with media-reported users at the 20 million level), AI inference costs eroding gross margins, two-way quarterly volatility in fund gains/losses, and continued mobile MAU decline.
All prices and valuations in this report are based on the 2026-10-23 closing price of RMB 248.96 (frozen anchor); financial data are as of the 2026 interim report (balance sheet stock data as of 2026-06-30); the 2026 Q3 report is expected to be disclosed on 2026-10-28, at which point key calibers will be updated.