Report Interpretation

Overview

JPMorgan assumes coverage of Kingsoft Office at Overweight, viewing AI as an extension of WPS’s productivity-software growth runway rather than merely a feature upgrade. The thesis rests on stronger personal monetization, WPS365’s enterprise growth, workflow-based competitive advantages and earnings-led upside.

Core views

JPMorgan argues that advances in general-purpose AI enlarge, rather than simply threaten, the addressable market for productivity software. Traditional office applications monetize document creation and editing, while AI expands the work that can be performed in WPS into research, analysis, retrieval, writing, presentation creation, collaboration and spreadsheet processing. The report expects monetization to move beyond seat-based subscriptions toward “subscription + AI usage” for individuals and “seat + AI usage” for enterprises. Lower model costs could improve the economics if WPS retains the customer workflow and customer spending rises faster than inference costs. The personal business already shows evidence of deeper monetization of an established user base. Personal revenue rose 15.2% YoY to Rmb2.0bn in 1H26, accelerating from 8.4% in 1H25 and 10.4% in 2025. Paid individual users increased 15.4%, compared with only 3.9% growth in global MAU, which JPMorgan interprets as evidence that conversion and spending per user are becoming more important as the installed base matures. WPS AI MAU exceeded 80mn at end-2025 and daily token consumption exceeded 200bn, up more than 12x YoY. The report expects paid penetration, membership mix and AI usage to support mid-teens personal-revenue growth over its forecast period. Enterprise SaaS is the second structural growth engine. WPS365 revenue increased 60.8% YoY to Rmb497mn in 1H26 and has delivered more than 60% growth for six consecutive quarters; it rose 64.9% in 2025. WPS365 combines documents, collaboration, enterprise knowledge, permissions and security, while Comate connects models to enterprise data and business systems. JPMorgan expects growth to moderate as the base rises but remain materially above group growth through 2028E. It also highlights WPS365’s Rmb599/user/year flagship package, versus Rmb199-399 lower tiers, as a route for AI-driven package upgrades before usage-based charging becomes material. The report sees durable differentiation shifting from standalone AI features toward distribution, persistent document state, native execution and enterprise context. WPS has 676mn global MAU, and its ability to turn model output into native editable files keeps it embedded in users’ workflows. WPS365 and Comate add customer-specific knowledge, permissions and system connectivity that may deepen switching costs. However, the key competitive question is whether WPS retains the primary interface to knowledge work as general-purpose agents improve. If agents become the interface, WPS could remain the execution layer but capture less of the customer relationship and incremental spending. Microsoft is cited as the long-term benchmark: Copilot is priced at about Rmb232/user/month in China-facing commercial channels, or around Rmb2,784 annually before the Microsoft 365 licence, indicating a potentially larger AI software budget but also a competitive benchmark. JPMorgan believes Kingsoft Office currently benefits from local deployment, domestic infrastructure, data-control requirements and Xinchuang compatibility. Financially, Kingsoft Office reported 1H26 revenue growth of 24.7% to Rmb3.3bn and adjusted attributable net-profit growth of 28.3% to Rmb1.1bn, which the report views as early operating leverage while the company continues AI investment. Operating cash flow was Rmb643mn, down 13.0% YoY, mainly due to employee spending, tax payments and prepaid income tax related to unrealized investment gains; the report therefore emphasizes normalized CFO and FCF relative to adjusted earnings over multiple periods. Reported attributable net profit of Rmb2.5bn in 1H26 included about Rmb1.9bn of investment gains. The 2025 cash dividend rose to about Rmb580mn, or Rmb1.2 per share, from about Rmb416mn in 2024, while approximately Rmb100mn of shares were repurchased in 2024 primarily for employee incentives. JPMorgan’s Overweight call is principally an earnings call rather than a rerating call. It forecasts 2026-28E revenue 5%/11%/8% above consensus and core profit 10%/2%/2% above consensus. The Rmb325 Dec-2027 target is based on 45x 2028E P/E. Although the shares trade at around 40x 2027E P/E and at a premium to China software and many global SaaS peers, the report notes historical forward P/E of approximately 94x, 77x, 76x, 71x and 53x in 2021-25. It considers 45x a meaningful discount to the company’s historical 60-80x one-year-forward range, while recognizing the larger earnings base, evolving AI economics and competitive uncertainty. Sustained 20%+ revenue and earnings growth with improving AI monetization could support further multiple expansion, but the base thesis assumes broadly stable relative valuation and earnings-driven upside.

Analysis framework

JPMorgan starts with AI’s effect on the productivity-software opportunity, then assesses Kingsoft Office’s personal and enterprise monetization engines, competitive assets and workflow ownership. It tests the thesis against operating metrics, AI usage and pricing evidence, compares WPS365 with Microsoft’s enterprise AI proposition, evaluates cash conversion and applies a historical growth-to-valuation comparison to set a 45x 2028E P/E target multiple.

Methodology notes

• Industry AnalysisSupply-demand framework

  AI expands the addressable knowledge-work workload while lowering intelligence costs.

  The report evaluates how broader AI-enabled tasks increase demand for productivity software and how declining model costs affect the economics of serving that demand.

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  Workflow ownership determines value capture between AI models, applications and enterprise customers.

  JPMorgan examines whether WPS retains the user interface, document execution and enterprise context as model capabilities become more interchangeable.

• Valuation methodsP/E and PEG Valuation

  Target price based on 45x 2028E P/E and comparison with historical forward P/E.

  The target multiple reflects the report’s comparison of expected medium-term growth with historical growth and valuation, adjusted for a larger earnings base and uncertain AI monetization.

• Corporate Fundamentals and FinanceFree cash flow analysis

  Normalized CFO and FCF conversion relative to adjusted earnings.

  The report treats cash conversion and incremental gross profit as confirmation that AI revenue is creating durable earnings rather than only higher usage.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Beijing Kingsoft Office Software, Inc. - A (688111.SS, 688111 CH)

  Primary covered company; expected to benefit from AI-enabled productivity workloads, personal monetization and WPS365 enterprise SaaS growth.

  Strengths

  Large-scale WPS distribution, persistent document context, native execution, WPS365 enterprise workflow integration, local deployment and Xinchuang compatibility.

  Weaknesses

  AI economics and the scale of willingness to pay remain unproven; AI revenue is not separately disclosed.

  Comparison

  Trades at a premium to China software and many global SaaS peers but at a discount to its own historical forward P/E range; Microsoft is the key long-term enterprise AI benchmark.

  Risks

  Earlier-than-expected personal monetization maturity, faster WPS365 normalization, interface displacement by agents, and AI inference costs outpacing pricing.

• Microsoft

  Competitive benchmark for enterprise productivity AI pricing and workflow integration.

  Strengths

  Copilot combines AI with Word, Excel, PowerPoint, Outlook, Teams, identity and enterprise data.

  Weaknesses

  China local-service feature differences may limit current competitive parity in some domestic deployment environments.

  Comparison

  Microsoft 365 Copilot is priced at around Rmb2,784 per user annually before the underlying licence, versus WPS365’s Rmb599/user/year flagship package.

  Risks

  Improvement in Microsoft’s China offering could narrow Kingsoft Office’s local-deployment and feature advantage.

Key data

• Personal revenue, 1H26Rmb2.0bn; +15.2% YoYAccelerated from 8.4% YoY in 1H25.
• Paid individual users, 1H26+15.4% YoYCompared with 3.9% growth in global MAU.
• WPS365 revenue, 1H26Rmb497mn; +60.8% YoYMore than 60% growth for six consecutive quarters.
• Global WPS MAU676mnDistribution base cited as a competitive advantage.
• 1H26 revenue and adjusted attributable net profitRmb3.3bn; Rmb1.1bnUp 24.7% YoY and 28.3% YoY, respectively.
• 2026-28E revenue versus consensus+5% / +11% / +8%JPMorgan forecast premium to consensus.
• 2026-28E core profit versus consensus+10% / +2% / +2%JPMorgan forecast premium to consensus.
• Target valuation45x 2028E P/ESupports the Rmb325 Dec-2027 price target.

Impact & implications

The report believes AI can lengthen Kingsoft Office’s growth runway if WPS converts its installed base and enterprise workflow position into higher ARPPU, paid conversion, AI usage revenue and incremental gross profit. It expects earnings growth, rather than a large rerating, to be the main support for the investment case.

Risks

• Personal monetization could mature faster if paid penetration reaches a ceiling, users will not pay separately for AI, or AI features become bundled into standard subscriptions.
• WPS365 growth could normalize faster as its base grows, especially if NRR, package upgrades, Comate adoption or conversion from pilots to recurring production use are weaker than expected.
• General-purpose agents or a stronger Microsoft China offering could shift the interface away from WPS, reducing Kingsoft Office’s share of customer relationships and AI spending.
• AI usage could pressure margins and cash flow if monetization does not keep pace with inference costs, usage charging develops slowly or workloads require more compute than expected.
• Slower-than-expected software localization progress and slower growth in paying subscribers are key downside risks to the rating and target price.

What to watch

• Paid conversion, ARPPU and AI usage revenue in the personal business.
• WPS365 growth, net revenue retention, AI attach rates, higher-tier migration and Comate production adoption.
• Whether AI usage produces incremental gross profit, alongside gross-margin trends and normalized CFO/FCF conversion.
• Evidence that paid-user growth remains above MAU growth and that personal revenue sustains mid-teens growth.
• Further consensus earnings upgrades and disclosure that clarifies AI usage monetization.