Distinguished investors, analysts and our friends from the media, good morning and I'd like to welcome you to join us for Kingdee [audio distortion] 2025 Interim Result Announcement. First of all, please allow me to introduce the management team with us here today. Mr. Xu Shao Chun, Chairman of the Board and CEO of the company, Zhang Yong , the President of Kingdee, [Represent and] CFO of the company, Mr. Lin Bo, Vice President and GM of [R&D], Mr. Lio Zhong Wen, and the General Manager of IR Department, Wu Chao. Ladies and gentlemen, coming next, we're going to welcome Mr. Lin Bo to work us through the financial performance and the future outlook, p lease.
Distinguished investors, analysts and friends from the media, good morning . I welcome all of you to join us for 2025 interim result announcement. By the end of 2024 or at the beginning of 2025, when we were formulating the annual plan of 2025, management team has already anticipated persistent microeconomic challenges while at the same time we also see the opportunity from artificial intelligence. That's why we adjusted our strategy. On one side we're seeking more opportunities, while on the other side we also well control our internal wastes. Here is the financial result I'd like to show to all of you. In H1 of this year, revenue increased by 11.2% on a worldwide basis, reaching RMB 3.192 billion. We on one side introduced more AI talents, while at the same time we continue to leverage technological innovation to continue to further downsize the team.
Our team size being decreased by 9% compared with the beginning of this year, close to 1,000 headcounts. Capital income generated per capita grew by 90.4%. As many of you may know, income expenses are a key component of our overall components expenses. Team structured development can actually help to further grow our business, where the cloud subscription revenue grew by 22.1% on a worldwide basis. For product implementation and other income, there is positive growth. We actually changed our way of information disclosure because we really want our investors and analysts to understand our business more clearly rather than mixing the business together. Cloud business is a cornerstone business we have. It's also a high margin business and you can see that the margin is more than 95%.
We do hope this win business and the subscription business connect to the high revenue contribution of the company as a whole. For product implementation, we really want to outsource it to the third party and make sure it's [keeping] a reasonable presence within our overall presence of the income and revenue. On the right side, a s you can see, because of the business changes, our gross profit margin grew by 2.4%, reaching 65.6%, still having ample room for further growth. That's the reason we are very confident for our future business growth. At the same time, the expenses and the promotion fees ratio has been further optimized.
The proportion of the subscription business company increased, the renewal rate also improved, the same as the sales and the promotional expenses are all being decreased. As you can see, we made heavy investment for R&D. On the right side, m ajority of the products are indeed on Cosmic. No matter for large, small, or medium-sized enterprises, w e are all using Cosmic platform to continue to improve the R&D efficiency. Even if you can see expenses down by 3.8%, we continue for investment and we also introduce many of the AI patterns to continue to lead to the business expenses when hot.
Let's also take a look at the SMG. It was growing up by 2.1%, which is very much different from what I see internally. The key reason is because we further optimize our team. Here is a one-time composition for the employees and it is close to RMB 80 million. This is an interpretation we'd like to share with you on that one-time expenses, where last year the last of the same period was RMB 480 million, where for this year it's actually been reduced to RMB 98 million significantly. At the same time, our operating cash flow was very decreased significantly.
I can say for software people, actually the human cost used to be very high and the bonus need to be paid in H1 of this year. That's the reason our operating cash flow is being stressful and in H1 of this year, actually revenue generally accounted for 50% - 55%, where 40% - 45% in H1 of this year, but it's a reversed story of expenses. We should believe for this year we're going to have a good profitability. Operating cash flow should be more than RMB 1 billion. Just now I walk you through the operationals in H1 of this year. I also would like to specifically talk about the long-term guidance. I have to say the external environment is volatile and for our forward-looking statement, it's also subject to our feelings and our external environment. Please just take it as a reference rather than coin it 100%.
I have to see the guidance we provided in data and a certain degree of the subjectivity where now we are a cloud service company based upon the subscription model as we have already mentioned, this is what we have in H1 of this year. Actually, cloud subscription revenue grew by 22.1%, the ERR grew by 80.5%, and the contract availability of the cloud subscription grew by 24.7%. At the same time, you can also see the GP margin was 96.2%, a 0.9% growth compared with last. The reason the margin was growing is because we are using third party AIs, but as we continue to develop the high-tech technology with ever improving computing power and our understanding of the market competition, even if we believe GP margin of a subscription is already 96.2%, it still has every opportunity to be further improved.
Let's also take a look at the cloud subscription. In the past, the majority of the large worried about whether the cloud subscription of the Constellation is looking right or not. If this line is not huge, the growth rate is looking very well. A ctually for Kingdee Cloud Cosmic and Constellation, and the subscription revenue grew by 41.1% and its return rate was 180%. At the same time, the net order retention rate for the corresponding Constellation subscription was 108%, while at the same time for the startup, its MDR was 23.8% and 93%, still with a good performance. That is also the reason why after the business transformation we'll still be able to maintain a very steady growth with our large customer base, same as what we used to plan. We have already changed the traditional one-time license transaction business model.
Cloud business or cloud subscription has become a key stream of our total revenue. Cloud subscription contributes more than 50% of our revenue. We are joining hands with our clients as partners for co-developments. All those indicators could actually be fairly same from the indicator as long as we maintain a good customer relationship, as long as we actually maintain a good MDR and retention, that would be the key priority of the company. In the past, we only worked on customer development, but now we continue to improve the service and mobilizing our clients to get more of our product, prolong the life cycle of the customer with ever increasing customer returns. In H1 of Asia, as you can see with artificial intelligence, with advanced management tools provided by Kingdee, and for Constellations, we constructed more than 500 corporate clients in [H1 of this year] .
We actually upload our slides to the website, but as you are here in the RFI channel, let me just share with you what did we do for our corporate clients. For Kingdee Cloud Constellation, w hat did we do and what we offered to each company? Many people know we do provide the business management tools, but what are those value? Let me just demonstrate to you. For example, at [China Merchants Group] we can help to further reduce the expenses for their accounting people by using artificial intelligence. At Hisense, the end-to- end process efficiency improved by 41%. We are at the same time at the national fleet and we also help them to bring great cost saving where for CGM and we also help them to further improve their savings by fighting the corresponding taxation policies.
You can also see even for Weichai we actually help them to further empower their domestic alternative upgrading. We hope all those customers have said improving their operational efficiency. While for Tongwei they actually have the double layer ERP, they need ERP for group and ERP for subsidiary. Both layer of the ERP is being well connected, making sure the entire group and umbrella is just using one system. This is also a co-innovation can be made with a customer where they also supported some of the specialized and high growth enterprises by bringing them great efficiency gains. For example, you see [Duo Mei] Holdings and [Fei Zhong]. Those are all efficiency gains we provided to each company including the turnover efficiencies settlement and the sales complete data delivery. We do have the one order one profit beside consolidation the cost for each of the company.
Coming next, please allow me to walk you through what we presented to you aside from our 2023 interim results. In the past, we believe the licensed ERP model going to be shifted to cloud management. Those are all the slides showcasing what has been done by Robert. We are actually shifting from B2B and B2C and also continue from the licensing to the subscription model. You can see what has been achieved by us. In 2015, revenue was RMB 1.5 billion, where by end of 2024 it was already RMB 6.2 billion. Where for today, I surely believe the interim result revenue going to be even higher for cloud subscription revenue. For the cloud service revenue, cloud service revenue has already accounted for 82.6% of our total revenue of the group.
If you take a look at the track record from the group or from the total revenue perspective, the CAGR used to be 50% where for the cloud service the CAGR was around 47.3%. There might be some low baselines, but we also have to further prove the success of the cloud transformation. In next stage, we're going to continue to transform for artificial intelligence. This is a new slide. We start from SaaS, continue to building our EBC capacity and now we emphasize on enterprises AI management. Our revenue tell us our SaaS business going to be continue to be strengthened. When we transform from SaaS to cloud and AI is a new opportunity for SaaS. For SaaS revenue is going to have a very healthy growth for the next five to 10 years. There's also a new opportunity from AI agent and also from the data service.
Generally speaking I should say for AI Cosmic platform in the past we do have the EPC enterprises capacity which indeed complementary to our existing AI agent that can help to leverage and mobilize resources from EPC for business management. What are those scenarios we are targeting now? Let's take a look at our product portal. Let me just make it simple. The more the enterprises use SaaS application, the more data it's going to generate and the more likely that AI going to be further applied. The more AI you use, the strength of the SaaS could also be further improved. This is leveraging SaaS to generate more data. Data support enterprises in a complementary role and AI will to some extent further strengthen the performance of SaaS. We are actually grabs this great opportunity. We have already provided data to profit cloud business from many years ago.
What are those typical AI scenarios we can provide? Let me just give you an example from the manufacturing industry. AI is every barrel for us. Majority of the time we were talking about financials but talking about manufacturing industry. Let me just show you two very typical applications in the future. No matter for product R &D, for manufacturing, for quality management, for supply chain management, for sales, to human resources management. In the entire flow of the company procedure you will be able to hide more applications of having AI. What is our key property? When we design the procedure to the companies, we will be able to have an interaction with PM, and PM thus has the collaborations with our design software for the smart schedulings, for quality management, for supply chain management, and human resources and financials. All those applications need artificial intelligence.
Artificial intelligence to some extent has been seamlessly embedded into every part of the process. When we talk about AI, it is not only a concept or technology. We hope that AI can truly unleash its power and value to indeed helping everyone to further improve the productivity. Let me just show you a customer case. It's a drone company, and this is actually the requirement statement from the client. After receiving this requirement statement, artificial intelligence can help to identify the highlight of the requirement and also have internal sort for our database and also identify the very similar model machine to become the prototype and then to actually identify the performance gap between the requirement and the model we have and be able to identify the indicators including the flight speed and cost requirement.
It will be able to identify the existing database with the samples and then be able to make adjustment in the software design. Indeed, AI is empower every industry, but someone, especially the media, was questioned whether AI is taking revolutionary changes to software. Let me just tell you it's not because AI majority of the time it actually use the data debris from the margin where for raw enterprises you need a complete set of the data. Especially the completeness of the ERP data means a large if you ask AI different questions you may get very different answer. Any company you need to have a very strong underlying data which can make sure you have a good data consistency. AI can also help us get access to the backstage data, and it can actually help to translate this four wing robot into a six wing robot.
This is how AI empowered the design. Let's also take a look at how AI supports sales process. It is an agent help the sales improve the response when there's any inquiries being made. We need lighting properties projects. Lastly, for John Xiaomi received inquiry rather than having a lighting fixture, she doesn't want to give a raise order to someone else. No worries, now we have the Quotation Agent as well as your inputs in barbie information into the PDAI quotations timing vacation is filled in. Product access will help the salespeople to understand what the client needs potentially and allocate the right solutions. By understanding the program, the AI will help to refer to the lighting design principles and product parameters, identify ties to you and then provide the product configurations and price, the whole process being done autonomously by AI.
We maximize and empower our salespeople, reducing the lending cost, continue to improve accuracy and success of product recommendation. If client is not happy with the result, Kingdee AI quotation system can also make adjustments. Time and energy being spent in this process being greatly reduced compared with manual paper. Besides the professional product recommendation in the quotation system, it also has the fresh texture and flash surface. It actually has the multimodal OCR effects and image. It will be able to help them identify any pictures that have been taken by the client over the handbrake statues or lesion regional purchase form right on the page. It can help respond to different customer inquiries, reducing the menu error and identify the right product portfolio and helping to further reduce the time for quotation.
AI implementation provided by can actually provide the pricing and according to the pricing system, once the client clicks on the position it will actually generate presentation for confirmation. It can also be shared to the WeChat of the clients. Since you conclude quotation journey, a journey that used to engage multi department collaboration has already been signified as AI plus social salesperson for highly incentive collaboration. We are transforming the business and service grid. Thank you very much and hope that for both cases you will have an intuitive feeling on how we use AI to further improve the production design, reducing the sales cost and also further improve the internal collaborations between different departments. For sure, as I have already mentioned, there will be more AI and AI agents to support enterprises for their business.
Well in 2025, b esides cloud subscriptions, AI, and globalization, we surely made some good progress in globalization. On the slide, we have already set up our subsidiaries in every country listed here, including we support the China Merchants Group and Luoyang as well as Hisense for their international expansion. Yesterday, we newly signed a group customer in Malaysia, and actually today we have already received the order, which could be converted into millions of RMB per year. So our business model in Malaysia is truly recognized by the local customer. We believe our international expansion is going to be of great success in the near future. I think many of you may very much like to know how can we leverage cloud subscription to rebuild Kingdee, and for the target of this year, as we promised to the investors, we're going to make a profit for this year.
Operating cash flow should at least be more than RMB 1 billion, where we also hope of discussing with the board of directors after the calculation. Even if the guidance to the market may have some uncertainties or subjectivities, we have every confidence in 2030 the AI revenue accounted for 30% of our total revenue. As I'm the one in charge of the finance, in 2030 by then our operating cash flow, the forecast here is quite conservative. We believe we'll be able to make it more than RMB 3 billion. This is actually a conservative assumption we have based upon the business growth rate we have. AI revenue will also account for 30% of our total revenue in 2030. We really would like to become an enterprise management AI company, the world leading one.
Coming next, we will get into the Q&A session, but right before the Q&A session, let me just welcome you to test our AI-empowered product, that is the Golden Key Financial Report application. It can actually help you to do the financial report analytics of these companies, including those ones on Asia market and the U.S. I will personally experience that, which is quite innovative and very powerful. For sure, I know the app still needs to be further improved. It can actually be complementary to our existing internal financial analytics. I would like to welcome you to join us to have the hands-on experience of the AI Gold Key Financial analytics hub. Thank you. Thanks for all the investors. Now let's welcome the management team to be rotated on the stage. We will start the Q&A now.
Okay, ladies and gentlemen, we're going to proceed to the Q&A session. First question. Let's welcome [Dai Li] sitting in the first front row, please.
Thank you very much. Thanks for the management team. My name is [Dai Lili] . I'm an analyst from Bank of America. Congratulations to the company on your wonderful performance in H1 of this year. Your AI business and subscription revenue was growing very good, the same as the profitability. Two questions from me. The first question, you talked about large, medium-sized, and small-sized enterprises. What do the trends of the demand look like? Because I think for this year your Constellation business was growing very fast, especially for subscription, the same as the mid and small-sized subscription business. What is the subscription trend do we see, especially for large enterprises?
Many of the companies are taking profitability as a key, so whether the competition landscape be improved. My second question is regarding AI, and in H1 of this year your AI backlog is actually RMB 150 million. In May, after releasing the AI product, including the Recruitment AI assistant, AI agent, as well as the Golden Key Financial Report analytics app, may I just kindly ask about the pricing model you have or even the feedback from the customers. What are the businesses that may have some good revenue or order contribution in H2 of this year? Thank you.
Thank you very much and thanks for your question. Let me respond to your first question. In Asia, for mid, small, and large-size enterprises, there are some demand trend changes.
For large enterprises, we do have SOEs plus POEs, where for SOEs or central SOEs, I think you probably know that for indigenous self-innovation, we do have the 2027 timeline where SE SEC and also asked for the penetrating management. The central SOE and SOC are actually having a penetrated management of all the companies within just one umbrella. While for large companies, at least from their demand, I see many of them are integrating their different business lines, making the supply chain management more transparent, and supporting their business worldwide. Those are indeed the highlight of the demand we see. For middle-sized enterprises, they still need faster deployment, lower cost, standardized product best practice, along with the industrial-specific attributes being built into the system. Those are the changes we see from the large and medium-sized companies. Not too much changes for the SMEs. Your second question is regarding AI.
On 4th of May of this year we launched some AI-powered product. We received a very good response from the market. For example, for Recruitment AI, for our ChatBI product, we do have some large company who come here to ask for a test pilot of the product. Where for microsize and a small company, what they need is a full system, but for large companies, they ready us to break down the tasks of the AI agent and they're going to have the AI agent for different scenarios with different applications. As we discuss with the client and we identify some new demands, we're just helping the clients to fine-tune their existing solution to take care of the customer needs.
Okay, let me just make one more comment. As Mr. Zhang has already mentioned, in H1, we actually launched many AI-powered products. We received many feedback from the customer. For example, like our Golden Key Financial Report analytics app, we shared this product with many senior executives of the company. We received very positive feedback from them, especially they believe our product is quite intelligent with very accurate data. As you can see, that's artificial intelligence. People are still quite concerned of using it in financials because of the accuracy reason, but our product still receive very positive response from the market. My second point, as Mr. Zhang has already mentioned, we launched many AI-powered products in May of this year. They are quite intelligent. For example, like the Recruitment Agent as well as the travel agent, it's actually a fully intelligent one could make support.
for Recruitment Agent, it can help to generate job descriptions by just having one sentence and recommending the right candidate for you, but it is more fit for the mid-sized and small company. For large enterprises, need to have a very refined management internally. Each of our products are having multiple agents inside. For example, like Recruitment Agent, we quit one agent. We actually have multiple agents in the underlying structures to support the agent as a whole. For large enterprises, maybe the Recruitment Agent will be broken down into small agents to support the large enterprises for their refined operations and management b y providing a good alternative. For Golden Key Financial Report analytics, we embedded 70 agents inside. Our ultimate result is to deliver one financial report analytics of one company as a whole. We still provide a complete result rather than building it into different agents.
We are talking about the future planning. On one side, we're starting to work for the customer, take care of the customer needs to roll out more AI agent experience. For example, just as I mentioned, in May we do have one platform with five major agents being launched, and Mr. Lin also mentioned about product design, R&D agents, and also sales and quotation agents . Those are all new products launched in H1 of this year. Along with artificial intelligence, the performance and efficiency has been further improved. We will continue to launch more agents in H2 of this year. For sure, more artificial intelligence would be allowed into our key SaaS system. Thank you.
Thanks for the response.
Let's welcome Liu Yang from Morgan Stanley, please.
Thank you. Thanks for giving me the chance to raise a question. I have two questions here. My first question, Kingdee adjusted the product price in H1 of this year. Some used to be free of charge s ervice started to charge in H2 of this year. It's been implemented for one month. How's the feedback from the customer of the product being charged? Any price adjustment plan? My second question, Mr. Lin, you were talking about team optimization and what about the people who do project implementation? Whether the implementation, business science, and growth would be impacted because business implementation has everything to do with customer product launch. Thank you.
Thank you very much. Two points you mentioned. Let me just respond to them one by one.
In H1 of this year, I do see some of the media and investors are making such discussion regarding the price, and we are just separating a standard advanced service from a standard service, making sure we provide file access to all the customers. Some people said we're just increasing the price. It may be true, but we don't recognize that because it's advanced service. Many customers still take it. First of all, by so doing, it can actually make our subscription offerings be more standardized, and we will also be able to understand the cost and service we provide for some advanced or higher demand. I think right before this meeting we actually provide you the certain data. I do think the customer really like the separate charge of the advanced features. We have 800 enterprises who signed for this advanced service with a very nice growth.
The second question you were asking about team optimization I think I have already mentioned in my presentation. You see, we do notice the high growth of the subscription business. Team optimization is not only for the implementation team. The whole team is going to be more refined and follow the lean management. We're not going to sacrifice the customer experience. What we're trying to do is to have the tailor-made service and handle or outsource those implementation service to our certified third party partners to further improve our product delivery quality. Without customer satisfaction, other things are all going to be empty. I have a few things to share with you. If we talk about the service changes, I think for any of the company, the change of the service has everything to do with the value of the service. Recently we just refined our service value.
We'd like to provide high quality standardized service to our customer. In July we made a switch, and later we actually are going to let the customer feel the changes from our service by providing value to the customer. You also asked a question regarding team optimization. On one side, we improve the personnel efficiency along with organizational change. For this year, regarding delivery for large enterprises, we'd like to have more centralized management and delivery. Some of the consultants, they're going to be built into one team, further improving the work efficiency. We're also developing our eco partner team. We actually strengthened our certification and assessment system. We have our partners in different verticals and industries making sure they will help us to deliver the product to the customer. We are also working with some of the leading companies for project implementation.
For example, we are working with Xiamen Digital Technology, which is indeed a company who used to do the SAP delivery in the energy sector. Starting from 2024, we and this company actually worked with Yunnan Energy Investment team. We produce the team, they help us with implementation. They have already become a key partner we have in the energy sector where in Guangdong Energy Group. We are working with KPMG for the product deliveries and we are responsible for the product design. We are supporting some high end professional partners to help to take care of the large enterprises' need as one team. More importantly for Kingdee, as Robert mentioned, we have four highlights of our business including systematic, standardized, procedure driven, and data focused. These are all things that PND continue to do.
Where you know that for me, in software industry in China, we attach grading footnotes to our system building. By having all those measures in one phase, you will see a consolidated management of Kingdee with the efficiencies being further improved. Thank you.
Thanks for the response from the management team.
Next question. Let's welcome Guangfa Security representatives.
Thank you. My name is [Liu Xiefeng] from GF Securities. I have two questions. At the very beginning, you used to be the pioneer for the cloud business transformation. You show good progress. How likely the sustainability might be in the near future? Where for artificial intelligence, what would be your 2030 development goal? My second question, yesterday in your financial report, your large enterprises business grew by 30%, total revenue close to RMB 800 million. We all notice for the last few years for SAP, their growth in China was quite flat, always very small growth. When your business reaches a certain level and your market share continues to go up, what would be the final outer limit for the large enterprises business growth? Thank you.
Artificial intelligence business growth, starting from the rollout of DeepSeek, people started to be more alerted on artificial intelligence. Still the market may have some variables, but was still very confident. Let me just mention one point, i f you take a look at our AI revenue. We're going to consolidate into the financial report. When we report the data to our investors, we take a very cautious and rigorous way of counting on the entire categories of our data. In the long run, a s I mentioned, in 2030 we hope that AI revenue accounted for 30% of the total group revenue and then by then how the company size might be. Roughly speaking, we hope that in the near future we'll reach a double-digit growth every year, probably 15%. 15%, a number to calculate what the figure might be for the past few years, for the future few years. Okay, Jason, would you mind responding to the second question?
Thank you. You asked me about the upper limit of the large enterprises business. In the short run, I think we'll still be able to grow our business sustainably. For subscription business, our market share was only 50%. Subscription business only accounted for 50% of our total. We hope we can make it to 80% - 90%. For large enterprises, the opportunities are there. In China, f or large enterprises, it used to be dominated by the international brands. In China, the market is indeed worse. Chinese companies are now having a fundamental understanding and needs for management. Once artificial intelligence comes, the work efficiency and AI adoptions will be further accelerated by large enterprises in China. There are more growth opportunities for us.
You know that this afternoon our go for travel and we have a major project in Malaysia and the value was more than RMB 10 million. Subscription is a few millions of the RMB. International market also presents us good opportunities. I think in the matter for China or international market, we face good opportunities. International, we are being challenged by large enterprises because you have many competitors. You have to take care of the local policy and the market, social, and compliance. For this year, we hope at least we will be able to continue to grow our business in China and also continue to develop our business internationally.
Thank you. Next question please. Ladies sitting in the back please.
[Shaodan] from CRCC . Ladies and gentlemen, thank you. Thanks for taking my question and congratulations on the company's AI priority, subscription-prioritized strategy with nice progress. My first question is regarding subscription. Mr. Lin has already shared with us in 2030 the AI is going to contribute to a certain ratio of your total revenue for the company. Would you mind to share with us what would be the gross profit margin of subscription revenue and how likely it's going to be contributed to your total revenue. Mr. Zhang, you also mentioned you have some international business in Southeast Asia and Middle East. What would be your plan for international business? Any of those strategies in different regions may differ from one to another? Thank you.
Thank you. Let me just respond to it first. I will have Jason to give more comments. Let me make it in this way. I hope that I will be able to combine the subscription revenue along with AI revenue for subscription data. We hope that we will be able to have a subscription from AI revenue together at least to be 80% - 90% of the total revenue of the company. The margin will surely provide us good returns. It really takes time.
It has everything to do with the global strategy where for the global strategy was told to follow Belt and Road business. We are going to follow Chinese company for international expansion on one side where at the same time you can see also in Singapore, in Malaysia, in Vietnam, and in Thailand in the Middle East. We also continue to deploy our business following the Belt and Road strategy.
Thank you. Let me just give some additional comment talking about international strategy for this year. Southeast Asia and Middle East would be our two key markets starting from 2024 to now for business opportunities and projects. They are looking very well. What we're doing to trying to do is to localize our product there.
We have already published the product pack in four countries and regions to fit into the local business application and be adapted by the local market. Some region they are famous for trade and manufacturing the supply chain. We just want to make sure our product truly reflects the local market needs. Each region and market has different requirements. For Middle East, for Southeast Asia, for Japan and Europe and North America for sure the market deepens a lot. As Mr. Lin mentioned, we are going to follow Northern Road initiative to continue our international expansion. In 2025, we want to consolidate our foothold in Southeast Asia and the Middle East. Starting from 2026, we will go for Africa to go for Mid Europe and Japan.
Coming next, let's welcome some media friends.
Hello, I come from the 21st century. I have a question regarding agent. People claiming this year would be the first year of agent. You also publish a corresponding product. Many people believe not many people would like to actually have a good review of the product for agent. How you're going to comment on that and DeepSeek actually reduce the AI development cost. How can you avoid the homogeneous competition? Thank you.
Thank you. Thanks for your question. Let me respond to your first question regarding the agent service renewal. First of all, for Kingdee, we are serving the customer no matter how the product we provide, no matter it's ERP software, SaaS, or the future AI-powered product. It only differs on the form of the products or technology. As long as we serve the customer, we still would like to join hands with the customer to continue to create value for customer.
This remains unchanged. For Kingdee as we are there for SaaS and for the subscription model, o ur renewal rate and our NDR is being quite high. For large enterprises, 108%, mid and small size company, 94% and 93% respectively. In the near future, when we provide AI agent products to customers, we still join hands with customer to create the value, be with the customer over time. No matter which technology or service we provide to customer, at least for Kingdee, we are there to provide the service to the customer. For the renewal rate or NDR, it will continue with a very high baseline. When the customer would like to renew the service or not, it truly depends on the product providing value to them or not. Still now, we are quite optimistic about it, about NDR. The second question is regarding DeepSeek is coming up.
How can we avoid the homogeneous competition? We actually internally made some reflections on that. When we choose a product, we always think about which products we're going to build in the market. There are many of the tech giants, they actually build AI agent company. We also have startup companies building AI agent company. Some of the tech subsidiaries of the large company, they also have a self R&D of the AI agent. What do we do? We still be quite cautious and prudent. Any product we launched at the very beginning phase, those products really close to customer needs and especially the high frequency demand of the customer. Why Kingdee can do so? Because when we provide AI agent or product, we consider multiple factors. First of all, you need to have the application or scenario. You can't blindly make a product with any future or potential application.
Tech giants company, they can release a product for faster iteration. B2B business not like so. If you don't have the customer with good application or scenario, you won't be able to iterate your product continuously. My second point is regarding data. We need to have the enterprises who are willing to join hands with you to hand over the data for you. That's the reason for Kingdee. We actually accumulated a very consolidated baseline by doing SaaS business. Customers are happy to share their data with us for business collaboration. At the same time, as I have already shared with you for the Golden Key Financial Report analytics app, it's actually our professional expertise for the past three decades. We can leverage DeepSeek.
We also recruited many of the algorithm and AI experts with industrial professional knowledge along with our financial expertise and the customer trust and the potential application that will be able to win. If a product can actually make a success, you have to consider multiple factors, weigh and pick the right applications we build with the right technology. Thank you. Just now we mentioned about the renewal rate, whether it is high or not. Because we just have the AI product this year, we still take time to see what the NDR might look like, the dollar retention rate. As long as the product provides good value to the customer, the dollar retention rate would be very high. Still a huge room for us to grow the business. One more point. I think if I talk about general large language model, how it's going to challenge our industry.
I think for Kingdee we do have our own moat because general large language model are for general ability where we do have in-depth industrial know-how because of our three decades accumulation. You need to know the taxations, financials, and business principles of each industry. That's why we will be able to differentiate from the general large language model. I also mentioned about the trusted custodies of the data. Enterprises, they won't be able to run their data on generalized large language model. They have to choose someone they trust and they can work with long term. They will always be able to find such a partner to work with by trusting their data to us. We can also use their data and their business to support guidance and support for the customer. Sadly, I think ecosystem and the data integration means a lot.
I think generalized large language model won't be able to generate such a good corporation opportunities in short run for Kingdee. We are a professional ERP solution provider. We do have a remote after many years development in the near future, I believe data is still the key. As in the slide we show you the data product. In the near future we may have the data product plus our AI platform. This is going to be the two major modes we have in the near future. You know that for me I was trying to experience many AI product because many of them being launched to the market. For AI product you have to consider the data source. Those companies who have the data source be more valuable in AI age.
As I demonstrated to you, we do have the end -to- end AI empowerment design to prototype production to sales. You need a complete set of the data. AI is more like a function, but you still need a complete set of the data to make the AI to be more supportive, the more versatile the generalized large language model would be it also going to support the business.
Let's welcome [Mao Xinjun] from Tianfeng Securities, p lease.
Management team, good morning. My name is [Mao Xinjun] from Tianfeng. Congratulations to the company with wonderful performance in H1 of this year. I clearly noticed you disclosed your subscription revenue for the first time and then we are very happy to see. A question regarding financials. A few years ago, Mr. Xu, as your proposed subscription prioritized strategy and for this year, your subscription business continued to be accelerated. Your GP margin, cash flow and net profit all show great momentum for growth. I have a question to you then. What would be the profitability the cash flow for this year and what would be the normalized or stable margin or the free cash flow rate in the next few years? I mean three to five years. Thank you. Any outlook you may have?
Thank you. Thanks for Tianfeng Securities of supporting the company all the time. As has been introduced just now, the company are improving the contribution from the high margin subscription business. Our financials would be improved continuously and we're also going to show very strong profit growth and cash flow growth in the next few years. I have already provided you conservative forecast. I won't be able to give you over optimistic growth.
Our operating cash flow surely going to be more than RMB 1 billion. To what extent is going to be? I don't disclose it here because we would like to play conservative. I was talking to my budget team and we believe by 2030 we have every confidence we can make sure operating cash flow is more than RMB 3 billion. Those are all conservative data. Because the growth rate was looking good for margin, we made some internal discussion. If margin is looking right, we will surely use a margin for further investment. As far as I believe, if we meet 8% of the margin, we believe we'll still be able to hit it, 8% margin, and we still need the money for more investment. 8% would be the sustainable margin. I still would like to focus on that.
According to my understanding, our operating profit margin, if we don't consider all the investment, it should be more than 15%- 20%. Going back to the operating profit, we hope 8% would be a conservative number. By referring to other large companies like Huawei on how to manage the financials.
Next question. Let's give the opportunity to friends from the media, please.
Hello management team. I come from the [Daily Economics] News. I have a question regarding the team o ptimization. It's a concern of the market. What would be your team optimization plan? Any plan you have? How many people need to be optimized? My second question regarding AI and the large language model, how the traditional ERP and the SaaS market competition landscape may look like. Is it going to see fundamental reshuffle? Thank you.
for team optimization. While not doing optimization for optimization reason, we also introduced many high quality titans. We actually proactively introduced high quality project managers and product managers, AI people to come to our company to wait. As far as I know, for team optimization, at least in H1 of ACN compared with H2 of 2024, our labor expenses would be further saved because we don't have the one-time compensation as what we see in H1 of this year. In the longer run, we'd like to stabilize our team or even with a slight team optimization. We'll still continue to optimize the team structure, putting more efforts on AI investment, c ontinue to optimize the team structure for our business growth.
Thank you.
Let's welcome the investor sitting in the middle, please.
Thank you. Thanks for the company. I have already noticed your international revenue a few years ago is still in the infancy stage. It seems that your international strategy has been showcasing very strong growth momentum. I noticed from your financial report, besides Singapore company, you also engaged more than 200 companies home and abroad to strategic cooperation. So Mr. Zhang and Mr. Lin, I'd like to ask you for your future global business. Besides Belt and Road initiative as your key strategy, is there any other plan you have for international market for the next one to two years? What would be the international business revenue growth? My second question is also have something to do with globalized business. When you go for going global, are you going to take on the capitalization strategy or financing strategies for your strategic business deployment worldwide?
Thank you. Let me respond to a question a few points has already mentioned. Let me see. Indeed for Kingdee, we actually start our international business in early start from Hong Kong. Beside that, there's no further movement. Starting from 2024, we accelerated our international business. For the past two years, we are following Belt and Road initiative to expand our international business. It's not easy for Chinese social company for international journey because you have to know the regulations, data, securities and compliance in every country and region. On one side, we made many efforts to be adaptive to the local laws and regulations. Otherwise you won't be able to build a business in that country or region by then. We hope in the next three to five years, international revenue could be 5% - 10% of the total revenue of the company. That was our assumption by then.
That is also indeed what our international team is working on to catch up. I hope we will be able to realize this goal as early as possible. That would be great. At the same time I also place new requirement over my product team. For the past few years when we were in international market, we spent many time to study the product features and the international needs. We now also have international business product unit to support our international business, continuing to help them to advance their business in regions and countries outside China. We now provide fully country region product package with 40 languages available. In the next three to five years, we hope we will be able to make sure international business accounted for 5% -1 0% of our total revenue.
Thank you. Your second question is regarding the capitalization movement. We still keep an open attitude to that. We're also talking to brokers and investment banks, but we're surely going to take a very cautious way of doing that. Thank you.
Thank you.
Next question please.
Management team. My name is Guo Ya Li, coming from Huatai Securities. I have a question. It's an industrial question. In 2025, people believe AI application in China is lagging behind those in the U.S. It seems that we don't have any further progress in China, but I know that each company does have very different cadence. I'd like to ask you, as you advance your AI agent product, what are those factors or constraints? Is it the capacity of the large language model, or do we lack the enterprises' data, or the customer perceptions, or the ROI reason? What are those constraints that may limit the AI applications in China? Thank you.
A few sentences from me, two points. Let's first talk about demand. In the short run, we all go for cost initiative, reducing cost, improving efficiency. Secondly, leveraging artificial intelligence to empower decision making of the operationals and predict the business for sure. Thirdly, we do need a business model that can overturn the traditional business model. That's for the demand. I think you were talking about the hurdles of AI application in China. There are a few. First of all, hurdle from the data. For Chinese enterprises, the data quality is not looking right. For the past few years, we don't have a very good data quality from Chinese enterprises, and many of the companies are indeed having the data silos. My second point. Many of the enterprises have their security concern over the data.
They don't want to mount their data to cloud and are worried about the leakage of the data. This is all going to constrain the AI application for Chinese companies. The second point is on technology and talents. We are indeed short of the top notch AI product. Algorithm and business are not being well connected. We don't have any hybrid talent who knows the business and AI. At the same time, we are short of such talents. When AI is being applied for business scenario, enterprises know nothing about it. No right candidate to help implement the project. They just want to make some simple attempts rather than just continue to have AI. Another way is the business transformation. Many of the companies are worried once AI comes, the existing organizations and human resources would be replaced. Organization itself is going to quite reject AI introduction.
For enterprises, they have to think about how to use AI to liberalize their people for more valuable and meaningful jobs. Finally, we're also facing ecosystem and standard constraints. In China, we don't have the standardized API interface, we don't have any good benchmark cases in one industry, and even the multi-agent schedulings or calls also are a big hurdle. Those are all the hurdles I believe may have constrained AI from having good implementation in China.
A few more comments from me. Mr. Zhang has already made a complete answer, but let me make it in this way. I'd like to emphasize on AI ecosystem. Indeed, enterprise service in overseas market has been quite complete. Many of the companies indeed have the API interface, which is a common practice. In China, many vendors, their product is not actually interoperable compared with overseas vendors.
In China, let's just make it simple. The so-called entry thinking is more important than interconnected thinking. Another reason is because the time frame. In international market, when ChatGPT 3.5 was available, that was in November of 2022. It's been around for a few months, and Chinese market also being constrained for ChatGPT 3.5 for a few months. In China, we waited till the beginning of this year when DeepSeek became available. Then artificial intelligence has seen explosive growth. We just have a half year after DeepSeek introduced. I believe in the near future there will be some nice ecosystem built as part of artificial intelligence in China.
Thank you.
Any further question from the onsite audience?
I have a question to you. What are your contacts on AI business?
For a company, we make some investment, but it's not going to pressure us because the company is not building the computing power. We just have the computing power on leasing model, where as you can see, we have already defined our margin. So AI investment not going to be a huge investment or not. That will be a key part of the complex.
Analyst sitting in the back, please.
Management team. My name is Luo Yin Yang from Dongxing Securities. Just one question from me regarding agent development. I have already noticed agent industry is still in the infancy stage. We still need to have the development partnering different applications. What about your R&D team composition? Are you going to change the algorithm ease to use agent? How long the development might be in the near future, w hether the agent development or product launch is going to be sustained at a few agents per year or are we going to have accelerated product release? I mean for agent. Thank you.
Thank you. For our company, our overall development is still the generalized or generical purpose agent. If we are embracing AI, we have to leverage a standardized agent particular to the customer needs. If you do agent development to every customer need, they're not going to be right. For tailor made agent, as Jason has already mentioned, while developing our ecosystem partner, they will be able to help to run the tailor made AI development. For the company, we hope what we provide is standardized products. I have one more comment. It surely depends on how you see the program.
Maybe the agent we developed, the AI products targeting on a few applications, if you ask me to count on the number of the agent, then you have to consider some of the applications with deep engagement of agent, maybe five, 10, a few dozens in the near future. As I have already shared with you, for each agent we have multiple small agents inside. For example, Recruitment Agent, we have the agent inside including job descriptions, candidate matchmaking, the candidate profiling, and AI interview agent. Some company just have one AI interview as a separate product. In our Recruitment Agents, we have small agents in the underlying layer. For example, for Golden Key Financial Report analytics, it has 70 agents in the underlying layer and we need the product team, engineering team, and algorithm team working together to have a professional task force to develop a specific agent product.
Where for Cosmic platform, we also have the agent development [SAP] platform. Our developers, our business people, they can access the Cosmic platform using the low code or zero code and be able to build many agents they need. On such agents, the number would be huge in the near future. This is how Kingdee and our ecosystem would be able to present. We will surely be able to have it into our SaaS key applications. Those are the two agent products we have. The first one is AI native agents and the second one is AI enhanced SaaS agen ts. Thank you.
Thank you.
Thanks for the management team. Due to time, ladies and gentlemen, now let's welcome Robert to give the final closing remarks, please.
Dear investors and our friends from the media, thank you very much. Thanks everyone for keeping an eye on Kingdee and your always support. Three sentences from me. My first sentence, at the very beginning of 2025 I have already mentioned the golden one decade of AI plus SaaS is already here. Till now, I still believe in this statement. AI plus SaaS's golden decade is not only here, it also will be accelerated for more iteration and evolution. You can see from the Kingdee H1 data, our subscription revenue still keeps a 22% growth. At the same time, the contracted value of our AI service is already RMB 150 million. I also noticed for any enterprises, if you don't embrace cloud, it would be very hard for one company to develop its AI service. No cloud, no AI. This is my statement.
I surely believe artificial intelligence will continue to advance the adoption of SaaS among enterprises. At the same time, because as we have data accumulation from SaaS, it will also enable the development of AI. This is my deep conviction. I believe the golden decade of SaaS and AI is already in progress. My second sentence to all of you. In the next one decade, artificial intelligence and plus management will surely become a key competitive strength of the company. AI plus management would become the key strengths of any enterprises. If we describe the core competence of any enterprises, we need to consider quality, cost, and speed. Artificial intelligence transformation for any company will surely help to improve the product quality. It will also help to further reduce the cost and also accelerate the speed of the product development.
AI plus management will surely become the key competitiveness of the company. W hereas we are in the financial adjustment stage, all companies are seeking for transformation including the operational transformations, business transformations, ecosystem transformations. Ultimately the most important one is still the management transformation. Management transformation means we have to embrace artificial intelligence. My third point, in artificial intelligence age, any advanced company, advanced enterprises need to count on management. Advanced management counts on AI. Advanced management need to count on AI. Using Kingdee would be your value enablers for this journey. We are leveraging advanced management. Using Kingdee is the great enabler. It is not only the commitment we made to the customer, but also the common pursuit we have. We can see that the transformation for the past one decade, cloud subscription has already accounted for 52.8% in April this year.
Our team size is already 11,000. In 2030, i f we continue to control our team size as 10,000, then if we have a double digit growth for the business, by 2030, our AI revenue of the company would range more than 30% as a whole. At the same time, we at least have an output of RMB 1 million per capital. More importantly, we're going to have more enterprises no matter from China or abroad to adapt Kingdee enterprise management AI. For Kingdee, we have a pursuit. We hope that we can make the Chinese management model rise. AI technology from the best enterprise management practice from China along with Chinese management philosophy will surely be able to make our product adopted worldwide. For Kingdee, we are indeed a company of great value. I was talking to my colleagues all the time.
If we take a look at the foreigners all over the world, if they all use the Chinese smartphones, driving a Chinese EV, that also showcases the great strength of China. Imagine if one day they don't use the Chinese management software, it cannot showcase the true strength of the Chinese model. I do believe the strength of China not only depends on our economic strength but also the soft power. Management software is the soft power. I always have a pursuit in my mind leveraging our solution to make Chinese management model wise. This is our belief. If we are a group of people with firm determination, with our great belief , with trust, with concrete actions, then you can imagine we use advanced management with Chinese management software for advanced management. This will one day be realized and ultimately Kingdee would become a company of great value.
I would like to thank all the investors and friends from the media for your long term support. We have experienced three transformations. Now we have undergone the AI transformation. In the next five years, we have more opportunities to go for. Thank you very much.
Thank you very much and thanks for the closing remarks from Robert. Here comes to the end of the interim announcement. We have arranged you the exhibition hall visit on the 30th floor of the Kingdee Cloud Building. The demo shown in our slides all have the on site demo with our staff on site for interpretation. For our investors and friends from the media, please go to the 30th floor to visit our exhibition hall. Thank you very much.