Report Interpretation

Overview

The report examines whether Yonyou can convert its system-of-record position and growing AI product activity into stronger financial returns. J.P. Morgan recognizes a larger enterprise-AI opportunity and meaningful strategic scarcity, but assumes coverage at Underweight because incremental AI monetization, consolidated revenue acceleration, profitability, and cash conversion remain insufficiently demonstrated.

Core views

J.P. Morgan argues that enterprise AI can expand the software value pool because agents may perform work that was previously supported by human labor. Traditional enterprise applications are monetized through licenses, subscriptions, modules, and employee seats, whereas agent-based products may also charge for usage, completed workflows, or transactions. In China, where enterprise software has historically captured a relatively limited share of corporate spending, this could allow vendors to address budgets linked to labor and services as well as conventional IT spending. The opportunity is not automatically incremental, however: agents may reduce required seat counts, general-purpose models may absorb reporting or configuration functions, and customers may redirect existing module or implementation spending toward AI. The report therefore focuses on the net change in total customer spending after cannibalization rather than standalone AI revenue. Yonyou has structural advantages because its BIP and ERP systems hold current enterprise records, permissions, business rules, integrations, and transaction history across finance, HR, procurement, supply chain, manufacturing, marketing, R&D, projects, assets, and collaboration. Agents still require this context and need systems capable of executing payments, procurement actions, and accounting entries. Customer-specific workflows and integrations with banks, tax systems, suppliers, logistics providers, and other software also create switching costs. J.P. Morgan consequently regards Yonyou's system-of-record role as defensible even if agent interfaces become more important than traditional software interfaces. The central uncertainty is who captures the incremental AI wallet. Existing ERP vendors such as Kingdee and global providers are adding their own agents; AI-native vendors can redesign selected workflows without maintaining legacy architectures; model and cloud providers are moving downstream; and large enterprises can build internal applications as AI lowers development costs. Yonyou appears better protected where work requires broad enterprise state, complex permissions, and transaction execution, but that does not ensure it owns the agent workflow or receives the associated spending. J.P. Morgan believes Yonyou is more likely to retain the underlying ERP relationship than to capture the full incremental AI value pool. This AI transition is beginning before the earlier cloud transition has fully matured economically. By 2025, cloud service revenue had reached approximately Rmb7.1bn, about 77% of group revenue, and cloud ARR was Rmb2.9bn. Cloud ARR rose 18.1% and BIP revenue increased 15.1% that year, yet group revenue grew only 0.3%. The contrast illustrates the report's concern that progress in strategic products has not consistently flowed through to consolidated growth. Yonyou now offers BIP 6 and agent-oriented products supported by 65 vertical agents, 623 Skills, and more than 26,000 application and service APIs, but disclosure remains limited on AI ARR, usage, attach rates, incremental customer ACV, AI gross margin, component reuse, and deployment efficiency. Early adoption figures are meaningful but do not yet prove wallet expansion. Yonyou reported Rmb464mn of AI-related revenue and Rmb906mn of AI-related contracts in 1H26, with AI revenue representing approximately 12% of reported revenue. Nevertheless, group revenue increased only 4.1% YoY to Rmb3.7bn, BIP revenue rose 15.0%, cloud service revenue grew 4.6%, and cloud ARR increased 8.2%. J.P. Morgan therefore places greater weight on consolidated revenue, total ACV per customer, existing-customer ACV or NRR, and incremental gross profit than on reported AI activity alone. Its 2026-28 revenue forecasts are 3%, 5%, and 10% below consensus because it assumes AI revenue translates more gradually into net additional customer spending. Financial conversion is the main constraint on the rating and valuation. Gross margin improved 2.6ppt YoY to 50.9% in 1H26, but attributable net loss remained Rmb928mn and operating cash outflow reached Rmb1.0bn. Roughly Rmb503mn of development spending was capitalized in the half, compared with about Rmb535mn recognized as current-period R&D expense, making cash flow an important complement to reported earnings. The report looks for revenue reacceleration to produce higher gross profit and operating leverage, followed by durable improvement in cash flow from operations and free cash flow per share. Revenue per employee, implementation duration, product reuse, gross profit per customer, and capitalized development costs are highlighted as indicators of whether BIP standardization and AI-assisted delivery are improving productivity. J.P. Morgan forecasts revenue of Rmb9.568bn, Rmb10.142bn, and Rmb10.750bn for 2026-28, representing growth of 4.2%, 6.0%, and 6.0%. Adjusted net income is forecast at a loss of Rmb636mn in 2026, followed by profits of Rmb79mn in 2027 and Rmb470mn in 2028; adjusted EPS is forecast at negative Rmb0.19, Rmb0.02, and Rmb0.14. Adjusted free cash flow to the firm is projected at negative Rmb220mn in 2026, positive Rmb635mn in 2027, and positive Rmb974mn in 2028. The report also raises its 2026 adjusted EPS estimate from negative Rmb0.21 to negative Rmb0.19 and its 2027 estimate from negative Rmb0.06 to positive Rmb0.02, while retaining a cautious view of the pace of consolidated recovery. The Dec-27 price target of Rmb6.00, reduced from Rmb10.00, is based on 45x 2028E P/E. J.P. Morgan retains a premium multiple consistent with the historical framework for leading A-share enterprise-software companies, recognizing Yonyou's strategic position and scarcity. Its weaker growth and financial conversion relative to higher-quality software peers are reflected mainly through a lower earnings base rather than a discounted multiple. The current price cited in the report is Rmb9.50 on 23 September 2026. Historical performance reinforces the burden of proof. Valuation compressed in 2022 as global growth multiples fell and domestic enterprise-IT demand weakened; during 2023-24, organizational changes, sales and delivery adjustments, weak revenue, and widening losses shifted attention toward execution and financial returns. BIP progress, enterprise AI, cost restructuring, and improving gross margin restored strategic optionality in 2025-26, but the report argues that a durable rerating requires simultaneous gains in customer spending, consolidated revenue, operating leverage, and free cash flow. Yonyou paid no cash dividend for 2025 following the loss, and shareholder distributions are expected to remain secondary while profitability and cash generation recover. The report would become more constructive if group revenue and cloud ARR reaccelerate while deployment intensity declines and free cash flow improves. Higher customer ACV or NRR, monetized machine usage, shorter implementation cycles, improved revenue per employee, and stronger retention would indicate that AI is expanding the customer wallet and changing the economics of the business rather than merely shifting the revenue mix. Conversely, weak enterprise IT budgets, cannibalization, competitive loss of the AI interface, or limited delivery-efficiency gains would reinforce the Underweight thesis.

Analysis framework

J.P. Morgan first assesses how AI changes the addressable enterprise-software spending pool and possible pricing units. It then evaluates Yonyou's system-of-record assets, competitive defensibility, and ability to retain incremental AI spending. The report tests product progress against consolidated revenue, customer economics, productivity, profitability, and cash flow; compares its forecasts with consensus; and applies a forward P/E framework to derive the target price.

Methodology notes

• Industry AnalysisUpstream-Midstream-Downstream Transmission

  AI value-pool and value-capture analysis

  The report separates growth in the overall enterprise-AI opportunity from the portion captured by ERP vendors, AI-native applications, model providers, cloud platforms, and customers' internal development teams.

• Competition & strategyEconomic Moat and Competitive Advantage

  System-of-record scarcity and defensibility

  Yonyou's enterprise data, permissions, workflow complexity, integrations, and transaction-execution capabilities are treated as scarce assets that create switching costs, while the report separately tests whether those advantages produce incremental revenue.

• Corporate Fundamentals and FinanceFree cash flow analysis

  Cash conversion after development investment

  Because a substantial portion of development spending is capitalized, the report evaluates operating cash flow and free cash flow alongside reported earnings to judge whether product growth creates sustainable economic returns.

• Valuation methodsP/E and PEG Valuation

  Forward earnings valuation

  The Rmb6.00 Dec-27 target applies 45x to 2028 estimated earnings, with Yonyou's strategic scarcity reflected in the multiple and weaker execution reflected primarily in the earnings base.

• Other

  Net customer wallet expansion after cannibalization

  The report evaluates AI monetization through the net change in customer ACV after allowing for possible reductions in seats, modules, services, and other legacy spending.

Asset mapping & comparison

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

• Yonyou Network Technology Company Limited (600588.SS, 600588 CH)

  Primary covered company and Underweight-rated China enterprise-software provider exposed to the transition toward agent-based enterprise workflows.

  Strengths

  Scaled A-share enterprise-software franchise with an installed base, system-of-record ownership, enterprise data, permissions, complex workflows, integrations, transaction execution, and BIP-based AI capabilities.

  Weaknesses

  Strategic-product growth has not consistently translated into strong consolidated revenue, profitability, operating leverage, or cash generation; disclosure on incremental AI economics remains limited.

  Comparison

  The report says Yonyou has weaker growth and financial conversion than higher-quality software peers, while existing ERP vendors, AI-native applications, model providers, cloud platforms, and customer-built agents compete for incremental AI spending.

  Risks

  Enterprise IT weakness, AI cannibalization, competitive loss of workflow ownership, slow productivity improvement, and weak cash conversion could prevent value capture.

• Kingdee

  Named direct competitor embedding AI and agents into its installed enterprise-software workflows.

  Strengths

  Existing enterprise-software position provides access to installed workflows.

  Comparison

  Cited as an incumbent ERP competitor participating in the same AI transition.

Key data

• Current ratingUnderweightPrevious rating was Neutral.
• Price targetRmb6.00Dec-27 target, reduced from Rmb10.00 and based on 45x 2028E P/E.
• Current priceRmb9.50Price as of 23 September 2026.
• 1H26 AI-related revenueRmb464mnApproximately 12% of reported revenue.
• 1H26 AI-related contractsRmb906mnEvidence of adoption, but not yet proof of net wallet expansion.
• 1H26 group revenueRmb3.7bnUp 4.1% YoY.
• 1H26 cloud ARR growth8.2%Compared with 15.0% BIP revenue growth and 4.6% cloud-service revenue growth.
• 1H26 gross margin50.9%Improved 2.6ppt YoY.
• 1H26 attributable net lossRmb928mnOperating cash outflow was Rmb1.0bn.
• 1H26 capitalized development spendingRmb503mnCompared with around Rmb535mn of current-period R&D expense.
• 2026-28 revenue forecast discount to consensus3% / 5% / 10%Reflects a more gradual assumption for AI-driven wallet expansion.
• 2026-28 revenue forecastsRmb9.568bn / Rmb10.142bn / Rmb10.750bnCorresponding growth forecasts are 4.2% / 6.0% / 6.0%.
• 2026-28 adjusted EPSRmb-0.19 / Rmb0.02 / Rmb0.142026E was revised from Rmb-0.21 and 2027E from Rmb-0.06.
• 2026-28 adjusted FCFFRmb-220mn / Rmb635mn / Rmb974mnThe model anticipates positive free cash flow from 2027.
• AI product building blocks65 vertical agents, 623 Skills, and more than 26,000 APIsThe report says usage and customer economics matter more than product counts alone.

Impact & implications

The report concludes that Yonyou's enterprise data and execution position should preserve its relevance as AI moves deeper into business workflows, but strategic relevance alone does not establish equity value creation. A durable rerating depends on AI expanding total customer spending while reducing implementation intensity and producing stronger consolidated revenue, margins, operating cash flow, and free cash flow.

Risks

• Weaker enterprise IT spending could delay contracts, reduce customer expansion, and slow cloud ARR growth.
• BIP or AI growth may replace legacy seats, modules, or services rather than expand total customer spending.
• AI-native applications, ERP rivals, model providers, cloud platforms, or customer-built agents could capture the incremental AI workflow and spending.
• Agent adoption could compress traditional seat demand before usage- or workflow-based pricing offsets the decline.
• Slow improvement in implementation efficiency could leave strategic-product growth dependent on heavy headcount and delivery spending.
• Reported earnings improvement without stronger operating cash flow and free cash flow would provide weak evidence of structural recovery.
• Upside risks to the Underweight view include stronger enterprise IT spending, faster BIP and cloud ARR growth, meaningful AI-driven ACV or usage revenue, and greater operating leverage than forecast.

What to watch

• Track consolidated revenue, contract liabilities, cloud ARR, total customer ACV, and existing-customer NRR for evidence that AI expands rather than reallocates spending.
• Monitor active agent usage, workflows executed, cross-module penetration, third-party integrations, and usage- or workflow-based revenue.
• Watch gross margin, revenue per employee, implementation duration, product reuse, gross profit per customer, and capitalized development costs for productivity gains.
• Assess whether operating cash flow and free cash flow improve alongside revenue and margins, with the report becoming more constructive if growth accelerates as deployment intensity falls.