After SaaS Profitability, Kingdee Bets on AI to Recreate Itself

Ask AI · How does Kingdee Cloud’s transformation lay a solid foundation for AI strategy?

Kingdee International (00268.HK) has delivered its most important results after cloud transformation and embracing AI.

On the evening of March 17, Kingdee announced its full-year 2025 performance, with total revenue of 7.006 billion yuan, up 12% year-over-year. Net profit attributable to shareholders was 93 million yuan, and adjusted net profit was 232 million yuan, both turning from loss to profit. Net cash from operating activities reached 1.097 billion yuan, a 17.6% increase, hitting a record high.

NVIDIA CEO Jensen Huang once mentioned in a presentation that future enterprise management systems will increasingly be called upon by Agents, reducing human learning costs, which will lead to explosive growth in internal software tool usage.

Kingdee is now reflecting Huang’s judgment in its reports.

The first profit turning point in China’s SaaS industry

This turning point is highly significant.

For Kingdee, it marks the completion of the most critical loop in its 13-year cloud subscription transformation that started in 2012: from strategic long-term cultivation to a new stage of comprehensive value realization.

For the entire Chinese enterprise management software industry, this is a benchmark moment.

Over the past decade, domestic vendors shifting from traditional licenses to cloud subscriptions have faced ongoing tension between profitability and growth quality. The market repeatedly asked: Is this path feasible? Kingdee is the first to give a definitive answer through a full annual report.

Breaking down the data, the core indicators point to the formation of a positive subscription cycle.

Subscription revenue reached 3.556 billion yuan, up 20.9%, surpassing 50% of total revenue for the first time. This milestone, where over half of revenue comes from subscriptions, is crucial. It indicates that Kingdee has substantively completed the revenue structure shift, truly becoming a company primarily driven by recurring income.

Gross margin on subscription software is 96.5%, boosting overall gross margin to 67.1%. This figure has room to grow; as the proportion of subscriptions continues to increase, margin improvement is assured.

ARR (Annual Recurring Revenue) is 4.09 billion yuan, and contract liabilities are 3.773 billion yuan, up 19.2% and 20.7%, respectively. These metrics lock in future revenue: even without new customers, renewal fees alone can sustain a substantial revenue base. Operating cash flow is 1.097 billion yuan, with sales expense ratio and R&D expense ratio decreasing by 1.1 and 2.9 percentage points year-over-year, and per capita revenue reaching 600,000 yuan.

The profit improvement stems from structural efficiency gains, ensuring sustainability.

Looking at business lines, large enterprise products Cangqiong and Xinghan generated 1.94 billion yuan, up 28%, with a net renewal rate of 110%. Mid-sized enterprise Xingkong subscription revenue was 1.52 billion yuan, with a renewal rate of 97%. Small and micro products saw a 22.7% growth in subscription revenue. All three lines are growing healthily, indicating customer stickiness is strengthening.

New clients include AI newcomers MINIMAX and Zhipu, as well as China Aerospace and Alibaba PingTouGe.

These industry figures go beyond Kingdee itself.

They provide the first complete example of China’s enterprise management software transitioning from cloud investment to profitability, and inject confidence into the industry’s future development.

Embracing AI comprehensively is Kingdee’s next goal

But Kingdee’s ambition goes beyond being a profitable SaaS company; it is redefining itself as an enterprise management AI company.

The success of cloud transformation has laid the foundation, but AI is the direction Kingdee is actively advancing.

Kingdee is reconstructing all product lines, organizational structures, and operational logic with AI technology, aiming to become an AI-native company.

In 2023, it launched Cangqiong GPT and a large financial management model; in 2024, it released an AI management assistant, providing personalized intelligent solutions for finance, HR, procurement, and development scenarios. By 2025, it aims to become a “corporate management AI company” and has upgraded its brand from “Kingdee Cloud” to “Kingdee AI,” launching China’s first enterprise-level AI-native super portal “XiaoK” and nearly 20 independent intelligent agents.

In March 2026, it will release an AI super suite with 26 intelligent agents covering finance, HR, supply chain, and manufacturing scenarios.

Internally, 41% of new code is generated by AI, some new products require 100% AI-assisted development, reducing R&D delivery cycles by 21%, with performance evaluations increasingly AI-oriented.

This path aligns closely with global enterprise software industry trends.

A report from HSBC at the end of February clearly states: Enterprise systems need to be repeatable, auditable, and error-free. Large foundational models inherently lack these attributes. Therefore, the best way to implement AI in enterprise scenarios is through embedded intelligent agents integrated into mature software platforms.

The companies best positioned to do this are those deeply engaged in enterprise management scenarios. Kingdee has solid foundations for enterprise management AI:

• Over 30 years serving more than 7.4 million enterprises.
• Deep understanding of complex scenarios like finance, supply chain, manufacturing, R&D.
• Relies on real-world scenario insights for B2B AI; its intelligent agents are closer to customer needs than general AI companies.
• Extensive digitalization has accumulated structured, governed enterprise data—crucial for AI products, as cleaner data enhances AI effectiveness.

Kingdee China Vice President and R&D Center General Manager Liu Zhongwen clarified at the earnings conference: Large model technology is now open-source. The next competition will focus on who has the highest data quality. Business processes and rules in ERP, finance, CRM, and supply chain systems are the foundation. Kingdee encapsulates these into AI-native systems on top of these platforms, enabling instant API calls by intelligent agents, creating natural entry barriers. Kingdee leads in Windows ERP and cloud-native transformation, and now pioneers AI integration across all product lines, proposing the Live AI route to help customers smoothly upgrade from digital systems to AI-native systems.

AI deployment data is already emerging.

In 2025, AI contract signing amounted to 356 million yuan. Shenzhen Energy processes 140,000 documents annually, with AI-assisted high-risk business interception rate increasing by 40%. Tongwei’s 50,000 employees fully use AI applications, with “three expenses” intelligent review accuracy reaching 97%.

AI is also opening new markets for enterprise management software.

Traditional ERP capabilities were limited to standardized functions; many long-tail needs were too costly to cover. AI significantly lowers the threshold for meeting these needs.

IDC data shows China’s AI market will reach about $63.1 billion in 2025, growing to over $200 billion by 2029. HSBC’s assessment is straightforward: The software industry is on the eve of a large-scale TAM (Total Addressable Market) expansion, which will last 5 to 10 years.

Kingdee is precisely at this intersection.

Cloud transformation provides stable cash flow and customer stickiness, while AI unlocks incremental markets. Its 2026 goal is double-digit group growth, with AI suite revenue exceeding 1 billion yuan, and adjusted operating profit margin rising from 3.3% to around 7%.

The 2030 vision is for AI + SaaS revenue and AI-native revenue to each account for 50%. Meanwhile, Kingdee is advancing globalization, establishing local service networks in Southeast Asia and the Middle East, with 463 new overseas clients signed in 2025.

This annual report’s significance extends beyond Kingdee. It confirms two things: Chinese SaaS software can be profitable, and enterprise management AI companies focusing on industry scenarios can become the core carriers for AI implementation.

The 2026 goal of 1 billion yuan in AI suite revenue will be the next key validation point.

_The above content does not constitute investment advice, nor does it represent the views of the publishing platform. Markets carry risks; please exercise caution, make independent judgments, and decisions.