Yonyou's $900M Loss vs Kingdee's $60M Profit: Decoding China's ERP Giants' Divergence

Capitalization Myth: Today's Losses Are Yesterday's Bills

The market focuses on Yonyou's 650 million RMB R&D amortization, often labeling it aggressive accounting. The article argues this reflects large-customer project-based R&D: customized solutions for state-owned enterprises, Xinchuang adaptation, and complex governance scenarios have 5-8 year reuse cycles. Capitalizing costs across multiple revenue periods aligns with accrual accounting. The 650 million amortization today stems from 2020-2023 cloud transformation peak investments, now coinciding with slowing revenue growth — a cycle mismatch, not operational collapse.

Conversely, Kingdee's 25-35% capitalization rate isn't more "conservative"; its mid-market focus yields standardized products where R&D is naturally expensed. Kingdee front-loaded costs in earlier years, so current subscription scale yields profit. The profit gap is a timing difference: one deferred costs to now (repayment phase), the other expensed early (harvest phase).

Subscription Penetration: Speed Difference Rooted in Customer Physiology

Revenue scale is similar but growth differs fourfold. The core driver is market transformation difficulty: mid-sized firms treat cloud ERP as a product decision (owner approves, IT executes, 3-month rollout), while large enterprises face institutional barriers — bidding, annual budget cycles, multi-subsidiary architecture adaptation. Yonyou's slower subscription conversion reflects customer decision-chain length, not execution weakness. Kingdee's speed benefits from mid-market's natural fit for standardized subscriptions.

However, slow doesn't mean safe. Subscription's anti-cyclical value matters in downturns: one-time license and implementation fees vanish, while recurring revenue provides a floor. Yonyou's pain this cycle stems from high traditional revenue share. Yet large customers' switching costs are 10x+ mid-market; once deeply embedded, they form an ironclad base. The long-cycle domestic substitution tailwind still favors Yonyou's hand.

Valuation Reset: New Script, Industry Still En Route

As of July, both stocks have fallen over 80% from peaks. This reflects a third paradigm shift in enterprise software valuation: from traditional PE steady-state, to SaaS-era PS+ARR growth multiples, to today's hybrid "profit verification + AI commercialization expectations." Each shift brings systemic valuation compression. The 2020 rally priced a "China Salesforce" fantasy — ARR growth justified any loss, PS reached 10x+. Now markets demand subscription quality, profit delivery, and tangible AI progress. The 80% drawdown isn't operational failure; it's the evaporation of unrealistic fantasies, returning to industrial pricing. Future gains require real profits and visible AI monetization, not just "cloud transformation" rhetoric.

AI: Cure or Next Money Pit?

Both bet on AI with notable contract signings: Yonyou 910 million RMB AI-related contracts in H1, Kingdee 230 million RMB AI-native product contracts targeting 1 billion RMB full-year. Yet Goldman Sachs maintains Sell on Yonyou, Neutral on Kingdee, while cutting both's multi-year profit forecasts. The consistent logic: AI monetization remains uncertain and requires heavy continued investment.