Kingsoft Office In-Depth Research

Longitudinal Company Development and Financial Review

Longitudinal Company History

Kingsoft Office's origins go much further back than the listed entity. Development of WPS 1.0 began in 1988, when Qiu Bojun wrote 122,000 lines of code on a 386 computer, and it was officially released in 1989. At that time, the core problems for Chinese-language office software were straightforward: Chinese typesetting was hard to do well on PCs, and the local software ecosystem was just starting, while imported office suites were unfriendly in both price and compatibility. WPS appeared because early PC adoption in China met the demand for Chinese word processing. It grew out of the foundational need for Chinese office work first, rather than following a foreign product to make a "local version." The company website puts WPS's starting point in 1988, and Kingsoft Software also positions itself as a veteran software company founded in 1988.

Early WPS was once almost synonymous with Chinese-language office work itself. Media retrospectives have noted that between 1988 and 1996, WPS once held more than 90% of China's office software market. This history matters not because it was "glorious," but because it established the deepest layer of Kingsoft Office's product DNA today: document compatibility, Chinese typesetting, and usage habits among government and institutional users. These things were later repackaged in the mobile and cloud eras, but the roots were never cut.

The real turning point came after Microsoft Office fully entered the Chinese market. WPS did not extend its absolute DOS-era advantage into the Windows era. It had to face a rival with stronger global standard formats, operating-system bundling, and enterprise IT ecosystems. Precisely because Microsoft squeezed it, Kingsoft's later strategic path became very clear: first, emphasize compatibility; second, build a low-priced or even free entry point; third, make multi-platform capabilities lighter and broader than the competitor's. This strategy was hard to use for a complete comeback in the PC era, but it reopened the field in the mobile internet era.

The listed entity itself was established on December 20, 2011, and completed its overall conversion into a joint-stock company on September 27, 2016. This milestone means WPS was gradually supported from a major business line of Kingsoft Software into a business entity capable of independent financing, governance, and listing. It is an old product with a new capital structure, not a new company that appeared out of nowhere. Later, it also became one of the first important software leaders on the STAR Market, and a classic case of a Hong Kong-listed Kingsoft Software spin-off to the STAR Market.

If Kingsoft Office's development is divided into stages, I would split it into four.

The first stage was the "Chinese-language office tool era." Growth in this stage was driven mainly by the scarcity of Chinese typesetting. Management's most important decision was to keep the product alive first, rather than to build a platform. The constraints were that China's PC software industry was not mature, standards were unsettled, and operating systems were changing quickly. The long-term impact of this stage was that WPS became deeply linked with Chinese-language office work and formed the user memory of a generation.

The second stage was "mobile internet reconstruction of the entry point." WPS relied on mobile, multi-platform, and free strategies to expand its installed base again. Looking back today, this was almost the only viable path at the time. Microsoft was a heavy suite; WPS had to be lighter, cheaper, and more suitable for phones and cross-device distribution. The most important thing in this stage was the rebuilding of active device count and user mindshare, not revenue. The long-term impact was that the company could later build personal subscriptions because it had made the entry point large enough first, rather than relying on a single hit product. By the end of 2025, WPS global monthly active devices had reached 678 million, including 329 million on PC and 349 million on mobile. That is the long-term return from that strategy.

The third stage was "subscriptionization and institutionalization after the STAR Market listing." In 2019, the company listed on the STAR Market at an IPO price of 45.86 yuan, raising 4.632 billion yuan, closing at 126.36 yuan on the first day, and reaching a market cap of 58.2 billion yuan. The story told to the capital market at listing was "domestic office software leader," but what the capital market repeatedly tested after listing was whether it could convert a huge active-user base into recurring payments and upgrade traditional licensing revenue into SaaS revenue. From 2021 to 2025, revenue rose from 3.280 billion yuan to 5.929 billion yuan, driven jointly by personal subscriptions and institutional business. The long-term effect was that the capital market gradually began to view it less as an "information-technology-application-innovation concept stock" and more as a software subscription company with a user base.

The fourth stage is "AI and organization-level platformization." This is the current stage. The 2025 annual report is dense with information: WPS AI 4.0, Office Agent, WPS 365's KAG framework, digital employees such as "Little K," "Big Cousin," and "departed-employee digital avatars," plus formal integration with OpenClaw. It shows the company is trying to connect AI to documents, meetings, email, organizational knowledge, and automated execution, rather than treating AI as an add-on that writes copy or makes PPT slides. The significance is that if WPS stays only at personal document generation, it will struggle to escape its valuation ceiling between Microsoft and free collaboration products. Only when it truly enters enterprise-side knowledge management and organizational processes can the valuation logic move from "tool" toward "platform."

Several key milestones deserve to be singled out. WPS 1.0 in 1988-1989 determined why the company exists; the establishment of the entity in 2011 and the joint-stock conversion in 2016 determined why it could move independently toward the capital market; the STAR Market listing in 2019 determined how it turned from a business line into an asset; and the AI product line and WPS 365 upgrades from 2023 to 2025 are now determining whether it becomes a "high-gross-margin office subscription vendor" or a true enterprise work platform. In hindsight, the first two milestones changed its fate, while the latter two changed its price.

Governance is also clear at present. The Q1 2026 report shows Kingsoft WPS Corporation Limited holds 238.39 million shares, or 51.45%, giving it a very stable controlling-shareholder position; the company has no differentiated voting-right arrangement. At the board level, Qiu Bojun remains on the board, and the Kingsoft Software system background remains deep. For ordinary shareholders, this structure has the advantage of stable control, while the downside is that strategic direction remains strongly embedded in the Kingsoft system. So far, this control has not turned into an obvious governance discount, but it means Kingsoft Office is not a fully "de-parented" independent software company.

Longitudinal Financial Review

The revenue curve is attractive, and the sources of growth have differed over the years. From 2021 to 2025, operating revenue grew from 3.280 billion yuan to 5.929 billion yuan, a four-year CAGR of about 16%. Initially, growth came more from personal-subscription penetration and the cloud migration of older products. In the past two years, WPS 365 and AI became new growth, while software licensing benefited from government and enterprise procurement and information-technology-application innovation, maintaining recovery growth. In 2025, personal business revenue was 3.626 billion yuan, up 10.42% year on year; WPS 365 revenue was 720 million yuan, up 64.93% year on year; and software licensing revenue was 1.461 billion yuan, up 15.24% year on year. By structure, personal business remains the largest revenue source. By direction, WPS 365 is the segment lifting the valuation logic.

The table below summarizes the company's main financial data from 2021 to 2025, sourced from the 2023, 2024, and 2025 annual reports.

This set of numbers has three features. First, cash-flow quality is very good. Operating cash flow exceeded net profit in each of the past five years, averaging about 1.5x net profit, which in a subscription software business usually implies decent prepayments, renewals, and collection discipline. Second, R&D investment has not been sacrificed for margins. In 2025, R&D investment as a share of revenue rose to 35.34%, still a very high level. Third, the company does not rely on capital expenditure to stack up growth. The software business itself is asset-light, R&D is largely expensed, and R&D capitalization in the annual report is zero. Therefore, the gap between accounting profit and "owner earnings" mainly comes not from capitalized R&D, but from revenue recognition and the treatment of investment income.

Profit quality does not look bad on the surface, but Q1 2026 reminds us that this company requires a clear separation between "making money" and "core business making money." The Q1 report disclosed operating revenue of about 1.613 billion yuan in the first quarter of 2026, up 24.63% year on year; personal business revenue of about 975 million yuan, up 13.80%; WPS 365 revenue of about 244 million yuan, up 60.79%; and software licensing revenue of about 355 million yuan, up 20.99%. The operating line is healthy. But the same Q1 report also shows net profit attributable to the parent of about 2.202 billion yuan, far above quarterly revenue. The explanation of profit changes is also direct: the sharp increases in total profit, net profit attributable to the parent, and adjusted net profit mainly came from large investment income generated by some external investment fund projects. At the same time, the non-recurring gains and losses table totals only about 24 million yuan. This mismatch means that the statutory "adjusted" figure did not perform the intuitive investor function of stripping out everything outside the core business in this quarter. Q1 profit metrics are clearly distorted. Research on this company should return to revenue, paying users, ARPU, renewals, and cash flow, rather than focusing on one-quarter net profit.

User data supports this judgment. By the end of Q1 2026, the company had 48.995 million cumulative annual paying personal users, including 46.15 million domestic users, up 11% year on year, and 2.85 million overseas users, up 64%. At the end of 2025, WPS AI domestic MAU was 80.13 million, up 307% from 19.68 million at the end of 2024; the paid conversion rate of free users who enabled AI rose by about 2.3x. Looking at these numbers together with personal business revenue, the key operating variable is now whether AI can lift a large low-ARPU base, rather than whether there are users. Based on 2025 personal business revenue and paying users, overall annual ARPPU is only about 74 yuan, with domestic at about 72 yuan and overseas at about 99 yuan. This shows the company's largest room lies in deepening value per user, not further enlarging the user pool.

The balance sheet is very stable. At the end of 2025, total assets were 18.156 billion yuan and net assets attributable to the parent were 12.853 billion yuan; by the end of Q1 2026, total assets had further risen to 20.217 billion yuan and net assets attributable to the parent to 15.130 billion yuan. The company has no heavy debt expansion and no heavy-asset capacity burden. What truly requires continuous investment is R&D, people, and channels, not factories and equipment. For this kind of company, the biggest financial risk is that profit metrics are jointly amplified by one-off investment income, accounting classifications, and market sentiment, causing investors to misjudge true earning power. It is not leverage.

Share Price and Valuation History

Kingsoft Office's capital-market story is almost a miniature of valuation sentiment toward Chinese software stocks. When it listed in 2019, the IPO price was 45.86 yuan, corresponding to a post-issuance PE of 78.37x; the first-day closing price was 126.36 yuan, up 175.51% from the issue price, with a closing market cap of 58.2 billion yuan. What the market paid for at the time was a head start on domestic software scarcity and long-cycle substitution logic, rather than the pricing of a mature profit stock.

The drivers of the following years' gains changed several times. Early on, it was information-technology-application innovation and domestic substitution; in the middle stage, cloud collaboration and subscriptionization; later, AI. The share price reached a historical high of 530.50 yuan on June 20, 2023, a period when the market priced "domestic Office leader," "core information-technology-application-innovation asset," and "AI office entry point" together. After the high, the share price declined for a long period, closing at 215 yuan on June 12, 2026, with a market cap of 99.625 billion yuan. The price has fallen by more than half, but it should not be simply interpreted as fundamentals being cut in half. A more accurate statement is that the valuation center has moved from thematic frenzy back under the discipline of operating delivery.

Using P/S makes this compression most intuitive. Based on the current market cap and 2025 revenue, the company's TTM P/S is around 16x; at the 2023 peak market-cap range, P/S was once pushed above 50x. In other words, the market today does not deny that it is a quality company, but it no longer accepts the logic that a story alone can sustain extreme multiples indefinitely. This change is crucial for valuation analysis because it tells us that over the next year, Kingsoft Office is more likely to rise on revenue and ARPU delivery than on concept re-expansion.

Another point is that reported PE has lost readability. Reuters' "PE excluding special items" looks only 27.5x, but this is highly likely distorted by Q1 2026 fund investment income. If we more conservatively use 2025 adjusted net profit of 1.803 billion yuan, current operating-basis PE is still around 55x. For an A-share software leader, this is not absurd, but it is certainly not cheap. Its current valuation is neither a bubble peak nor deep undervaluation. It is closer to a normally expensive range for a high-quality asset.

Business Model, Moat, and Industry Position

Business Model and Revenue Structure

Kingsoft Office now has three main revenue pillars, but they correspond to three very different businesses. The personal business is a typical consumer subscription business, converting a huge active-user pool into continuous payments. WPS 365 is enterprise SaaS, centered on organizational penetration, collaboration depth, and renewal capability rather than installation. Software licensing is closer to traditional government and enterprise software procurement: stable, predictable in cash flow, but less elastic than subscriptions. In 2025, these three segments generated 3.626 billion yuan, 720 million yuan, and 1.461 billion yuan in revenue respectively, accounting for about 61.2%, 12.1%, and 24.6% of total revenue. Structurally, this is still a personal-end-led company. Directionally, it is trying hard to become a company with a higher enterprise-side weight.

The table below lists only core operating numbers to clarify the pace of the company's shift from "personal subscription" toward "personal plus enterprise."

What matters most in this table is which segment is changing the revenue structure, not which segment is largest. WPS 365's revenue share was only about 8.5% in 2024 and had risen to 12.1% by 2025, an increase of about 3.6 percentage points in one year. It is far from becoming the company's first curve, but it is already large enough to affect valuation. Conversely, if WPS 365 growth falls below 30% over the next two years and revenue share stalls around 15%, the market will reclassify the company as a model of "high-quality personal subscription plus stable licensing," and the valuation ceiling will fall noticeably.

Advertising and other revenue are no longer core today. They still exist, but the share is small. Their importance lies more in proving that the WPS entry pool can support diversified monetization than in determining company value. What Kingsoft Office truly needs to prove has always been whether AI can keep expanding subscription and collaboration businesses, not whether it can earn a bit more advertising money.

Cost Structure and Operating Leverage

This is a typical software company, not a typical internet advertising company. Its main costs are R&D, sales, and continuous product refinement. The R&D expense ratio was 35.34% in 2025, 33.11% in 2024, and 32.32% in 2023, showing that the company's strategy over the past two years has been to continue investing in AI and enterprise capabilities rather than squeezing profit. R&D capitalization is zero, which also means management has not "beautified" the income statement by capitalizing R&D.

Operating leverage exists, but it is not linear. Once conversion rises in the personal subscription business, marginal cost is low and profit improvement can be relatively clear. WPS 365 has similar characteristics, but it requires more sales, delivery, and ecosystem-building investment upfront, so profit release often lags revenue release. Operating leverage in licensing depends more on government and enterprise procurement rhythms, and is less smooth than subscription. For Kingsoft Office, the ideal state is steady ARPU uplift in personal business, a rising share from WPS 365, and licensing holding the base, rather than explosive growth in only one segment. If these three segments move together, margins will naturally rise. If only licensing supports the business, profit may not be bad, but valuation will come down. If there is only an AI concept without enterprise payment delivery, revenue cannot support high multiples.

The hardest cost to cut when revenue declines is clearly R&D. The office software track is not a platform business of "occupy the spot first, optimize slowly later." It requires continuous work on format compatibility, multi-device experience, model integration, knowledge search, permission control, and data security. If investment stops, the product will not collapse immediately, but competitiveness will clearly decline two or three years later. In other words, Kingsoft Office does not have the large capex pressure of traditional manufacturing, but it has a softer yet equally rigid cost: continuous R&D.

Moat and Its Boundaries

Kingsoft Office's most real moat is first the installed base and usage habits. At the end of 2025, the company's major products had 678 million global monthly active devices, including 329 million on PC and 349 million on mobile. Installed base does not equal pricing power, but it at least means WPS has become one of the default office tools and is not relying on paid traffic to maintain its presence. Especially in Chinese-language documents, mobile office work, government affairs, and education, user acceptance of WPS is far higher than that of many overseas peers. This moat is valid on the personal side.

The second moat is localized compatibility and information-technology-application-innovation adaptation. The biggest pain point in office software is whether formats are stable and whether the software can mesh seamlessly with existing workflows, not whether it can "write text and make spreadsheets." What WPS has accumulated in China over time includes not only Chinese interaction and template ecosystems, but also adaptation to domestic software and hardware systems, government and enterprise processes, and document standards. The annual report defines the basic office software industry as a technology-intensive field and emphasizes capabilities in intelligent document processing, collaborative office platform architecture, data security, and privacy protection. This is less propaganda than an explanation of why it can still hold the leading position in China's office software market.

The third moat is price and distribution efficiency. Compared with Microsoft, WPS has long used lower prices, a more flexible free entry point, and stronger mobile coverage to exchange for installation in China. This remains effective today. The official enterprise price for Microsoft 365 Copilot is USD 30 per user per month. Based on the USD/CNY central parity rate of 6.8109 published by China Foreign Exchange Trade System on 2026-06-14, that is about 204 yuan per user per month. This means that as long as Kingsoft Office can offer a "good enough and localized" experience in AI office products, it will always have a chance to take part of the market with a lower total cost of ownership.

But all three moats have boundaries. Installed base is not a network effect; having many users does not mean enterprises will not migrate away. Localized compatibility is also hard to use against organizational collaboration platforms that reconstruct workflows. Price advantage is most vulnerable when the industry enters "free AI-fication." If Feishu, Tencent Docs, and DingTalk embed AI capabilities deeply into collaboration, knowledge bases, and approval flows, then subsidize prices through platform ecosystems, WPS will increasingly struggle if it relies only on low prices. What it most needs to strengthen is organizational capability, not document capability.

Management and Governance

At the management level, today's Kingsoft Office still clearly carries the mark of the Kingsoft system. The 2025 annual report shows Zou Tao is the company's legal representative; Qiu Bojun still serves as a director; and the governance team completed a transition during the reporting period, emphasizing a "smooth transition and orderly handover." This kind of wording sounds formal, but it matters for Kingsoft Office because it means the company did not suffer a management vacuum during AI transformation and organizational adjustment.

The equity structure is relatively stable. The Q1 2026 report shows controlling shareholder Kingsoft WPS Corporation Limited holds 51.45%, and together with several employee shareholding platforms, control is firm. The company has no dual-class shares or differentiated voting-right arrangement, a positive for A-share investors. On the other hand, this structure also means the will of management and the parent group is highly aligned, and capital allocation will emphasize long-term product strength more than short-term financial optimization.

Among historical blemishes, the most worth recording is that during the 2019 IPO feedback stage, the Kingsoft PowerWord App received rectification notices over privacy policy, permission explanations, and account cancellation issues. The company subsequently completed rectification and submitted reports. This did not develop into a major penalty or ongoing litigation, but it reminds investors that once office software moves toward cloud collaboration and AI, data compliance and permission management become part of the main product, not peripheral issues. Today, Kingsoft Office connects AI to knowledge bases, meetings, email, and digital employees, and the governance requirements will only be higher than they were then.

Horizontal Peers and Current Fundamentals

Horizontal Peer Analysis

If one insists on finding the "most similar" listed comparable for Kingsoft Office, it is actually difficult. A more accurate statement is that it sits at the intersection of office suites, collaboration platforms, and document AI companies. It differs from a global platform productivity giant like Microsoft, from an independent PDF tool vendor like Foxit, and from comprehensive platforms like Tencent and Alibaba that use super-traffic entry points to move backward into collaboration. For that reason, horizontal analysis cannot be limited to parameter comparison. It must look at what each company truly lives on.

Microsoft has become the global office system itself. Its advantage is that these products have become part of enterprise work standards, rather than Word, Excel, or PowerPoint being individually strong. In the AI era, Microsoft has put Copilot on top of the entire Microsoft 365 suite, with an enterprise price of USD 30 per user per month. What it sells is essentially a productivity layer that can directly connect to Teams, Outlook, SharePoint, and enterprise permission systems, not a "smarter document tool." For Kingsoft Office, Microsoft defines the ceiling; it is not merely a local competitor. As long as large Chinese enterprises are willing to keep paying higher prices for global standards and mature IT systems, Microsoft will remain the hardest rival to shake in the high-end enterprise market.

Tencent's route is different. Tencent Docs' official description is very direct: it supports multi-person online editing of Word, Excel, and PPT documents, real-time cloud saving, and permission settings based on QQ and WeChat contacts. Its biggest weapon is lightweight collaboration and instant distribution, not complex functionality. One step further, Tencent's AI workbench ima has already turned "knowledge base plus AI search-and-answer plus document understanding" into a new entry point. After Tencent's full-year 2025 results disclosure, related reports said ima had more than 13 million MAU and more than 420 million knowledge-base files. Tencent is more like a company starting from social relationships and instant collaboration and cutting backward into office work, rather than starting from an Office suite like WPS. For small and medium teams, project-based collaboration, and light-document scenarios, this route is threatening.

Feishu represents another path, one that moves more "deeply into the organization." Its official definition of itself is already "ByteDance's AI work platform," not traditional collaboration software. It integrates AI into knowledge, meetings, spreadsheets, workflows, and enterprise assets; it has also launched an official OpenClaw enterprise deployment solution and a low-code platform deeply integrated with Feishu. Feishu's real strength is turning documents into a node in organizational operation, rather than making documents. Its pressure on WPS is in enterprise scenarios, not personal users: if management, projects, approvals, knowledge bases, and Agents all run inside Feishu, the document tool may deteriorate into an attachment editor. Kingsoft Office has clearly recognized this, which is why WPS 365 moved obviously in 2025 toward digital employees, KAG, and OpenClaw compatibility.

Foxit Software is more like a mirror. It also began as a domestic office-document software company, also does subscriptions, and also does AI, but its commercial focus is more on PDF editing, electronic signatures, and document processing. Its existence helps investors judge one question: if office software only does "document tools" well and cannot become an organizational platform, what valuation will the capital market give it? By mid-June 2026, Foxit's P/S was about 5.48x and its market cap about 6.0 billion yuan, far below Kingsoft Office. The gap lies in who is closer to the "default work entry point," not who has more feature buttons.

The table below includes only numbers, not judgments. Cross-market market caps are converted using China Foreign Exchange Trade System central parity rates as of 2026-06-14: USD/CNY 6.8109 and HKD/CNY 0.86926.

The business meaning behind this table is clear. Kingsoft Office is far more expensive than Foxit because the market believes it is more than a PDF or document tool. But it is also more dependent on a single track than Microsoft because it lacks outer moats such as operating systems, email, CRM, and cloud infrastructure. Compared with Tencent, Kingsoft Office is much more expensive because Tencent's office collaboration capability is only one component within a large platform, and valuation will not independently price document collaboration. Therefore, Kingsoft Office has only one premise for maintaining a high valuation: it must prove that in China's office scenarios, it can both defend the suite entry point and gradually capture organization-level collaboration value. As soon as this stalls, valuation will quickly fall back toward the tool-software range.

From an ecological position, Kingsoft Office is the leader in China's general office software field, but in the AI collaboration era, it is still a leader transitioning toward platform status. The gap it fills is "an office suite in China's localized environment that is compatible enough, light enough, and more price-friendly." The profit pool it most directly attacks is Microsoft's in Chinese-language office work and SMEs. The profit pool most likely to be taken from it comes from platforms such as Feishu and Tencent Docs that bundle collaboration, knowledge bases, and Agents together, rather than from single document software. If technology substitution occurs, Kingsoft Office's position will not automatically strengthen. Its position becomes steadier only when WPS 365 truly embeds itself in organizational processes.

Current Fundamentals

Looking only at business lines, Kingsoft Office's current state can be summarized in one sentence: revenue is accelerating, the structure is changing, and the income statement is distorted. In 2025, quarterly revenue was 1.225 billion yuan, 1.188 billion yuan, 1.214 billion yuan, and 1.494 billion yuan; in Q1 2026, it further reached 1.613 billion yuan, up 24.63% year on year. This shows the company began to accelerate again in the second half of 2025 and did not stumble in Q1 2026. The strongest driver is WPS 365, followed by the recovery in software licensing, while the personal business shows resilience rather than explosion.

Putting the 2025 annual report and Q1 2026 report together, two different curves appear. One is the operating curve: paying personal users continue to grow, overseas paying users grow rapidly, WPS AI MAU expands, WPS 365 maintains growth around 60%, and software licensing returns to double-digit growth. This curve is good. The other is the profit curve: Q1 2026 fund investment income caused profit metrics to diverge from operating reality. If investors mix these two curves together, they will reach the wrong conclusion that the company has suddenly entered a "big profit-release year." In reality, it has entered a stage of "operating acceleration, but profit needs to be redefined."

There are three positive operating signals worth the most attention. First, WPS AI is no longer just showing off technology. The company explicitly disclosed that the payment rate of free users who enabled AI rose by about 2.3x, which means AI has entered the commercialization chain. Second, overseas users are no longer just a traffic story. Overseas annual paying personal users grew 64.43% year on year, consistent with the company's international version upgrade and deployment nodes in Southeast Asia, Europe, and the United States. Third, WPS 365 is moving beyond the positioning of a simple "enterprise Office" toward knowledge enhancement, digital employees, and automated execution, which is crucial for enterprise stickiness.

Market Narrative and Bull-Bear Divide

The current market mainly trades three things: AI monetization, recovery in government and enterprise information-technology-application innovation, and overseas incremental growth. AI is the part with the greatest imagination because it affects both personal paid conversion and enterprise collaboration value; information-technology-application innovation is the valuation floor because software licensing and domestic substitution give the company a base that is not easy to completely destroy; overseas is the new elasticity because it gives the market the imagination that "WPS is not only eating the Chinese market."

The bull evidence is clear. First, the funnel from users to payment is improving. WPS AI domestic MAU more than quadrupled in one year, and the payment rate of AI-enabled free users rose 2.3x. Second, the enterprise-side second curve is becoming visible. WPS 365 revenue grew 64.93% in 2025 and another 60.79% in 2026Q1, while functionally moving clearly toward knowledge and Agents. Third, cash-flow quality remains stable. Operating cash flow exceeded net profit in each of the past five years, showing this is not a software company telling a story through accounting tricks.

The bear evidence is equally solid. First, enterprise business is still too small. However fast WPS 365 grows, it accounted for only 12.1% of 2025 revenue, still far from rewriting the company's valuation structure. Second, AI office is a crowded track. Microsoft sells standards, Feishu sells organization, and Tencent sells collaboration plus light knowledge bases. WPS does have a chance, but it faces strong opponents across three different dimensions. Third, current valuation does not provide enough margin of safety. On 2025 operating-basis profit, the company's PE remains around 55x; on rough owner earnings, free-cash-flow yield is only in the low single digits. Fourth, the distortion in the Q1 income statement may lead some market participants to overestimate core earning power, which is both a source of mispricing and a risk of subsequent financial-report repricing.

In my view, the market is most likely to misjudge where AI will land the value, not whether the company has AI. If AI mainly raises personal membership conversion and pricing, Kingsoft Office will become a better subscription company. If AI ultimately shifts the center of competition to organizational processes, knowledge bases, and Agent orchestration, while WPS 365 penetration fails to keep pace, the market will find WPS's moat narrower than it imagined. The true dividing line is enterprise-side retention and ARPU, not model parameters.

Valuation, Risks, and Tracking

Valuation Analysis

To value Kingsoft Office, the first step is to rule out methods that should not be used. First, reported TTM PE cannot be used. Q1 2026 investment fund income made the net-profit metric almost incomparable. Second, statutory adjusted net profit cannot simply replace core-business profit, because Q1 "adjusted" profit did not fully strip out all fund income that affects operating judgment. Third, SaaS thinking is reasonable, but it cannot be mechanically copied from U.S. stocks because China's office software pricing environment, government and enterprise procurement cycles, and competitive landscape are different.

Start with cash-flow-through analysis. From 2021 to 2025, operating cash flow / net profit attributable to parent was about 1.79x, 1.43x, 1.56x, 1.33x, and 1.36x, respectively, with a five-year average of about 1.50x. The annual report also shows R&D capitalization is zero, so the company does not have the problem of "putting large amounts of R&D onto the balance sheet to make the income statement look better." Because the software business is asset-light and maintenance capex is very low, I approximate 2025 operating cash flow at 95% as owner earnings in the valuation, arriving at about 2.376 billion yuan. That corresponds to a current owner-earnings multiple of about 42x and a free-cash-flow yield of about 2.4%. This yield is higher than the most expensive layer among consumer blue chips, but for a growth software stock, it does not represent a meaningful margin of safety.

The historical valuation position also supports this. At the company's 2023 peak, implied P/S exceeded 50x; today it has come back to around 16x, which is certainly much lower, but still meaningfully above platform giants such as Tencent and also above tool-type document software companies such as Foxit. In other words, the market no longer treats it as a pure concept stock, but still does not treat it as an ordinary software stock. Its valuation center has fallen from "thematic frenzy" to "high-quality growth with demanding delivery requirements," but it has not fallen to "clearly undervalued."

The table below gives my three valuation scenarios. This is "scenario analysis under a research framework," not investment advice. The price range corresponds to the reasonable range over the next 12 months, based mainly on combined assumptions for 2027 revenue and owner earnings, cross-checked with target P/S and owner-earnings multiples.

In this framework, the current price of 215 yuan is closer to the lower end of the base scenario than to the "safety cushion" of the bear scenario. This shows two facts. First, the market has already paid for "the company is good." Second, the market has not fully paid for "the company will definitely turn AI and WPS 365 into a platform." Therefore, it is not as overheated as it was in 2023, but it also does not give investors the thick protection of a typical undervalued stock.

The expectation gap is likely to appear in four indicators: WPS 365 renewal and new-signing rhythm, ARPU of personal AI members, net additions of overseas paying users, and whether the company continues to show large investment income that interferes with the income statement. If in the next financial report revenue continues to grow quickly while profit metrics return to normal, the market will find it easier to accept higher-quality pricing. If revenue remains stable but enterprise-side growth slows and profit is again distorted, valuation is more likely to be pressured.

The conclusion of the margin-of-safety review is cautious. Based on the bear range above, the current price is at a premium to conservative value, and the margin of safety is close to zero. The most fragile assumption is "AI can continue to raise ARPU, and the uplift will not be offset by free competition," not revenue growth itself. If that assumption is cut by 30%, my base-value estimate would be revised down from 210-245 yuan to about 185-210 yuan. A more direct test: if core earnings show zero growth over the next three years and valuation does not expand, investors will probably receive only returns close to the dividend yield plus limited buybacks, with annualized returns meaningfully below the roughly 1.74% level of China's 10-year government bond yield on 2026-06-12. My conclusion is: this is a good company, but at the current price, the margin of safety is not obvious.

Risk Analysis

The first risk that could truly cause permanent capital loss is AI office becoming free too quickly. I assign a "medium to high" probability and a "high" impact. The reason is simple: Microsoft, Feishu, Tencent Docs, and DingTalk will all view AI as a weapon to improve user stickiness and platform activity, and will not necessarily treat it as a profit center that must be charged separately. Once the industry begins to view "AI generation, AI summarization, AI search-and-answer" as standard office platform features rather than add-on packages, Kingsoft Office's most imaginative ARPU uplift logic will be compressed. Observable indicators are personal AI paid penetration, pricing of AI-related membership packages, and whether competitors announce that core AI capabilities are being merged into base versions. The transmission path would be direct: slower ARPU uplift in the personal business, lower commercialization elasticity for WPS AI, reduced market tolerance for high multiples, and a stock price that first de-rates before growth is reassessed.

The second risk is that WPS 365 enterprise penetration is not as fast as the market expects. I assign a "medium" probability and a "high" impact. Whether Kingsoft Office's valuation can hold over the next few years depends largely on whether the company can move from a personal subscription vendor to a true "organization-level SaaS platform." But as of 2025, WPS 365 still accounted for only 12.1% of revenue. If renewal rates, new customer signings, digital employees, and other capabilities fall short over the next few quarters, and WPS 365 growth drops from the 60% range to 20%-30%, the market will reclassify it as a "high-quality office software stock with a limited growth ceiling." What needs to be tracked is WPS 365 revenue growth, management's disclosure density around large-customer cases, and whether the product keeps advancing toward KAG, OpenClaw, and digital employees rather than staying at document SaaS.

The third risk is continued profit-metric distortion causing capital-market misjudgment. I assign a "high" probability and a "medium to high" impact. It already happened once in Q1 2026: operations were good, but the income statement was distorted by large investment income, and even the "adjusted" figure did not fully solve the problem. This type of risk may not hurt company operations, but it hurts the valuation anchor. If similar situations repeatedly occur in future quarters, the market will become less willing to trust PE and instead look only at P/S and cash flow. Once a growth stock loses its profit anchor, volatility will expand noticeably. Observable indicators are the share of investment income and fair-value change gains in the income statement, and whether management proactively restates operating-basis metrics in earnings communications.

The fourth risk is disruption from policy and government/enterprise IT budget cycles. I assign a "medium" probability and a "medium" impact. The market often treats information-technology-application innovation as a long-term positive that can only go up, but for a single company, budget rhythm, tender delays, fiscal constraints, and changes in procurement standards can all cause periodic volatility in licensing business and some enterprise projects. Kingsoft Office can certainly benefit from the domestic software environment, but it is not the policymaker and cannot decide procurement realization speed. Observable indicators include software licensing revenue growth, government and enterprise project delivery rhythm, and changes in government procurement rules and data-security requirements. In transmission, licensing revenue and enterprise signing pace are affected first, followed by profit and valuation.

The fifth risk is data security and product compliance. I assign a "low to medium" probability and a "medium" impact. Once office software connects to knowledge bases, email, meeting minutes, and digital employees, permissions, audit, and privacy management become core capabilities, not peripheral functions. The company experienced App privacy rectification during the IPO stage, showing such issues are not remote. If enterprise customers worry about permission boundaries when AI calls enterprise knowledge, or if regulators impose more detailed requirements on enterprise data and AI-generated content, Kingsoft Office will need to keep strengthening product trustworthiness. This risk will not show up in revenue as immediately as a price war, but if an incident occurs, it usually hits enterprise customer trust first and valuation second.

Catalysts and Tracking Indicators

Among positive catalysts, I care most about three. First, WPS 365 maintains growth above 40% for several consecutive quarters and management discloses more real deployment cases for digital employees and knowledge-enhanced generation. This would push market expectations from "concept validation" to "platformization starting to take shape." Second, the personal business shows clearer ARPU uplift, not only paying-user growth, meaning AI is truly raising value per user. Third, the company proactively separates core-business profit from investment income in financial reports and communications, helping the market rebuild an earnings anchor.

Negative catalysts are equally clear. If WPS 365 growth quickly falls below 25%, or overseas paying-user growth slows sharply, the second curve is decelerating. If Microsoft, Feishu, and Tencent further merge core AI office functions into base versions, the market will lower expectations for WPS AI monetization. If the company again reports "stable revenue and surging profit, but mainly due to investment income," valuation may come under pressure before fundamentals do.

The table below is the dashboard I think should be tracked over the long term. The table includes only numerical thresholds, with explanations after it.

Why these indicators matter can be said in one sentence. Personal business revenue growth and paying-user growth determine whether the base is loosening; WPS 365 growth determines whether the second curve is still being delivered; overseas paying users and AI domestic MAU determine whether new growth is only a story; operating cash flow / revenue determines whether profit is cash-backed; and P/S is the most direct thermometer of market sentiment. For a company like Kingsoft Office, valuation often heats up first and then forces fundamentals to chase, rather than fundamentals turning bad first.

Key Data Table

The table below puts together the most important capital-market numbers for quick review.

Together, these numbers show the stock's core capital-market character: it has always been a growth stock with high expectations, strong narratives, and large valuation elasticity, rather than a cheap stock sold on static profit. It was true at listing, true when the AI theme lifted it, and remains true today. The difference is that the market's reason for paying a high multiple has shifted from "domestic substitution" to "AI and enterprise platformization."

Research Uncertainty

First, the company does not disclose ARR, net revenue retention, enterprise customer count, and seat count like a typical U.S.-listed SaaS company, so judging WPS 365 quality must rely on revenue growth and product progress, lacking more granular retention evidence.

Second, there is a clear mismatch between the accounting classification of Q1 2026 investment income and the "non-recurring gains and losses" table, so reconstructing core-business profit metrics involves a degree of researcher judgment. I have pointed this out clearly in the report, but it remains the largest metric uncertainty in the current analysis.

Third, non-independently listed businesses such as Tencent Docs and Feishu do not continuously disclose full operating data like listed companies, so horizontal competitive analysis must rely more on official product information, parent-company disclosures, and industry reports. It is hard to do precise financial disaggregation as with Microsoft or Foxit.

Fourth, overseas business growth is eye-catching, but disclosure remains coarse, especially with limited breakdowns by region and product. Public information is insufficient to determine whether rapid overseas paying-user growth comes from price increases in mature markets, penetration in Southeast Asia, or channel changes.

Cross-Sectional and Longitudinal Synthesis

Longitudinally, what Kingsoft Office has truly proven is the ability to survive three generational transitions, not merely the ability to "make Office." It first survived the Chinese-language office era, then was not fully squeezed out in the Microsoft-dominated PC office era, then rebuilt its entry point in the mobile internet era, and today it is entering the new migration toward AI office work. Many Chinese software companies can do the first step; far fewer can do the second; those that can continuously do the third and fourth are rare. The hardest point in Kingsoft Office's journey is that it has never drifted away from the core problem of "office work," rather than any single quarter's growth rate. Many companies chase new terms and find their main business drifting away after a few years; however much Kingsoft Office changes, its core remains documents, collaboration, knowledge, and workflows. That continuity is valuable.

Its past success had both era dividends and management and product-team judgment. Era dividends certainly existed: China's PC adoption, mobile internet diffusion, domestic substitution, and the AI boom each gave it opportunities. But opportunity is not outcome. What converted opportunities into results were two capabilities. The first was turning a "low-priced compatible tool" steadily into a mass entry point. The second was willingness to keep investing in product iteration and platformization before the money was fully visible. The R&D expense ratio staying around one third for the past five years is the best proof. One can say the market once valued it too highly, but it is hard to say the company was lazy on product.

Are these success factors still present today? Most are still present, but their form has changed. The personal-end entry point remains, compatibility and localization advantages remain, and the government/enterprise and information-technology-application-innovation base remains. What has truly changed is that the competitive center of office software has moved from "whose document tool is easier to use" to "whose organizational efficiency system is more complete." This is exactly the hardest and most critical question now facing Kingsoft Office. It has proven itself at the personal and licensing ends, but the enterprise platform end is still in the validation period. WPS 365 has grown quickly over the past two years, and the AI route is clear, but revenue share is still not large enough and remains some distance from becoming the company's valuation center. For that reason, looking at Kingsoft Office today requires attention both to what it has already proven and to what it has not fully proven.

Horizontally, Kingsoft Office's real advantages against peers remain "China-localized office entry point plus price efficiency plus multi-device compatibility plus government and enterprise adaptation." These advantages are effective on the personal end and in part of the government and enterprise end. But its weaknesses are equally clear: in high-end enterprise IT systems, Microsoft's standard status is still hard to shake; in organizational flows, approval flows, knowledge bases, and Agent scenarios, Feishu and Tencent attack more like platforms; and in PDF and document component scenarios, Foxit-like vendors can provide more focused niche solutions. In other words, Kingsoft Office has a moat, but that moat mainly surrounds the "office entry point," while the market's next-stage concern is the "organizational platform." This is not a problem that can be completely solved in the short term by adding one or two features.

The current valuation rewards what it did right in the past more than fully discounting the future. But this "reward" is still not cheap. On 2025 operating-basis profit, PE around 55x still requires sustained growth to digest; on P/S around 16x, the market is effectively prepaying part of the delivery from WPS 365 and AI. If enterprise-side growth genuinely accelerates over the next year, the current price may not be absurdly expensive. If AI monetization stays more on the personal side and the enterprise side remains slow to break through, this price will look quite full. My judgment is that the market is most likely mistaking "growth exists" for "growth is enough to support the current multiple," rather than misjudging whether the company has growth. These are two different things.

The key variables for the next 1 year, 3 years, and 5 years are layered but not the same. Over the next 1 year, the question is whether WPS 365 revenue growth can hold around the 40% center and whether AI continues to lift personal ARPU. Over the next 3 years, the question is whether the enterprise end can form a more stable renewal logic around collaboration and knowledge management. Over the next 5 years, the question is whether Kingsoft Office can become one of the "system-layer platforms" in China's office scenarios, rather than only a very strong tool-layer brand. If all three questions are answered correctly, Kingsoft Office will be a compounder that can be held for the long term. If the second question starts to get stuck, it will look more like a high-quality software stock with limited valuation elasticity.

For investors, under what conditions would it become a better target? The answer is that price and operations must both become more certain, not simply that "the share price falls a bit." The ideal state is: the share price falls back into my buy range, WPS 365 maintains high growth, AI paid conversion continues improving, and the income statement is no longer disturbed by large investment income. At that time, it would satisfy both "the company is good" and "the price is right." Conversely, if WPS 365 growth falls over the next two quarters, or the market begins to see signs that free AI competition is compressing ARPU, then the research conclusion should be revisited even if the share price is not high. The most dangerous mistake in studying a company is treating the old moat as eternal when the business logic starts changing, not buying at too high a price.

Bull and Bear Cases

Bull case:

• The personal subscription base is very solid. By 2026Q1, cumulative annual paying personal users had reached 48.995 million, with overseas users up 64% year on year. The company does not lack an entry point; the only question is value uplift.

• WPS 365 has upgraded from an enterprise document tool into a knowledge-enhanced, digital-employee, and Agent platform. Revenue grew 64.93% year on year in 2025, and the second curve is beginning to form.

• AI has begun to enter the monetization funnel rather than staying at the demo layer: WPS AI domestic MAU grew 307% in one year, and the payment rate of AI-enabled free users rose by about 2.3x.

• Cash-flow quality has consistently exceeded the income statement. Operating cash flow exceeded net profit in each of the past five years, showing strong cash collection capability.

• Control is stable and there is no differentiated voting-right arrangement, making the governance structure relatively clear among A-share software companies.

Bear case:

• Enterprise business is still too small. WPS 365 accounted for only 12.1% of 2025 revenue, not enough by itself to support a platform-type valuation.

• Q1 2026 profit metrics were heavily affected by investment fund income, and even "adjusted" profit does not fully represent core-business performance, making the earnings anchor unstable.

• Microsoft, Feishu, and Tencent are pressuring office AI simultaneously from standards, organization platforms, and lightweight collaboration. Competition is a three-dimensional encirclement, not isolated point pressure.

• Current valuation remains expensive. On 2025 operating-basis profit, PE is still about 55x, and the margin of safety is not thick.

• If AI is rapidly made free across the industry, Kingsoft Office's most market-anticipated ARPU uplift logic will be hurt first.

Pre-mortem

The first loss scenario I worry about most occurs in the second half of 2026 to 2027. Microsoft keeps embedding Copilot deeply into enterprise email, meetings, knowledge, and permission systems, while Feishu and Tencent further turn Agents, knowledge Q&A, and light collaboration into platform standards. Enterprise customers begin to find that "good documents" are no longer enough to decide procurement. Kingsoft Office is forced to increase AI and sales investment, WPS 365 revenue growth falls from 60% to 25%-30%, and personal business ARPU does not rise meaningfully because of industry free competition. At that point, company revenue can still grow, but the market's multiple could compress from 16x P/S to 10-11x, implying a possible share-price range of 160-180 yuan, down about 20%-25% from current levels.

The second scenario is worse and occurs in 2027-2028. The company continues to have investment income or accounting-metric disturbances. The market initially overestimates core-business profitability, then over two to three quarters finds that enterprise-side delivery is below expectations and the high profit in the financial statements is not sustainable operating profit. In that situation, valuation would not merely drift down slowly; an "earnings-trust discount" would appear. If this is compounded by macro budget contraction and delayed information-technology-application-innovation orders, the share price may seek a new anchor at 140-160 yuan, a decline of more than 20%, equivalent to another 25%-35% drop from the current price. What would truly cause a 50% investment loss is the simultaneous occurrence of free competition, stalled enterprise penetration, and valuation compression, not any single one alone.

Final Research Conclusion

My final judgment is that Kingsoft Office remains a Chinese software company worth tracking over the long term and with high fundamental quality, but buying it today means buying a stage where "operations can still move forward, while valuation may not return to high heat." It is neither a bubble stock nor a cheap stock. It has proven that it can survive and grow in the brutal office software track, and it remains at the table in the AI cycle. But it has not yet proven WPS 365 into an enterprise platform large enough to rewrite the company's valuation center. In other words, company quality is clearer than stock attractiveness.

I do not doubt WPS's position in China's office market, nor do I doubt that it can capture some incremental growth from the AI office trend. My real concern is that the market has already seen that "AI can improve conversion," but has not fully priced the possibility that "AI may also push industry pricing down." Once office software enters the Agent era, the variables determining victory will spill beyond single-document capability into collaboration, workflows, permissions, and knowledge bases. Kingsoft Office is filling this gap, and not slowly. The problem is that the current market price does not leave much room for trial and error.

If I were to become more positive in the future, the conditions are clear: first, the share price returns to a range with a better margin of safety; second, WPS 365 maintains high growth for several consecutive quarters and discloses more enterprise-customer deployments; third, the income statement returns to a basis that can represent core business operations. If all three happen together, Kingsoft Office would move from "good company, average price" to "good company, clear entry point." Until then, it is more suitable for holders to track patiently than for chase-style additions.

【Company Profile Score】

• Fundamental quality: High

• Growth: Medium

• Moat: Medium

• Financial stability: Strong

• Management credibility: Medium

• Valuation attractiveness: Low

• Risk level: Medium

• Suitable investor type: Long-term growth

【Investment Rating】

• Rating: Hold

• One-line investment thesis: The core business is solid and AI is being delivered, but enterprise platformization has not yet reached the stage where it can support valuation re-expansion.

• 【Ideal/Fair Buy Price】170-190 CNY

• Basis: It should correspond to the bear scenario and leave enough margin of safety, at least covering the risks of free AI competition and weaker-than-expected enterprise penetration.

• Holdable price: 210-245 CNY

• Clearly overvalued price: 280-320 CNY

• Current price classification: Holdable

• Is it worth waiting for a better price: Yes; the more suitable trigger range is 170-190 yuan, while requiring WPS 365 growth to remain above 35% and AI paid conversion not to deteriorate. The opportunity cost of waiting is that if enterprise-side delivery exceeds expectations, the share price may rebound before reaching the buy point.

• Target holding period: 1-3 years; more suitable for medium-term holding by existing shareholders than for a pure 6-12 month trade.

• Expected annualized return: bear -21% to -12%; base -2% to +14%; bull +30% to +49%.

• Maximum loss risk: about 25% to 35%; triggers are slower enterprise penetration, intensified free AI competition, and valuation multiples compressing from mid-to-high levels to 10-11x P/S.

• Signals triggering reassessment: If WPS 365 revenue growth stays below 25% for two consecutive quarters

• If cumulative annual paying personal-user growth falls below 5%

• If ARPU or conversion rate for AI-related memberships stops improving

• If investment income again dominates the income statement, making core-business profit hard to identify

• If competitors merge core Agent/knowledge-base capabilities into base versions at scale and the company is forced to follow pricing

【Valuation Range】

• current: 215.00 (as of the 2026-06-12 close)

• bear (conservative, ideal buy zone): [170, 190]

• base (reasonable, acceptable holding zone): [210, 245]

• bull (optimistic, above the clearly overvalued line): [280, 320]

References

This report is mainly based on the following public materials for judgment and cross-checking: the company's 2023, 2024, and 2025 annual reports and Q1 2026 report; listing, governance, and announcement information disclosed by the Shanghai Stock Exchange and the company website; official product and financial report/IR pages of competitors including Microsoft, Tencent, Feishu, and Foxit; China Foreign Exchange Trade System central parity rates; and market data pages such as Reuters, Google Finance, HKEX, Yahoo Finance, and TradingView for the latest share price, market cap, and multiples.

Other Securities Mentioned in the Report

• MSFT.US — The benchmark for global office suites, Copilot pricing, and organization-level platform capabilities.

• 00700.HK — Tencent Docs and ima represent the "light collaboration plus knowledge base plus AI workbench" route.

• 09988.HK — The group behind DingTalk, an important alternative ecosystem reference for enterprise collaboration and organizational workbenches.

• 688095.SHG — Foxit Software is a comparable domestic document/PDF software company, reflecting the valuation boundary of "tool-type document software."

• 03888.HK — Kingsoft Software is the controlling shareholder and historical spin-off parent, affecting control and governance structure.

• ADBE.US — An international reference for functionality and pricing in PDF, document creation, and AI-driven subscription software.

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