Media friends, investors, good morning. Thank you very much for joining AsiaInfo Technologies Limited's 2024 annual results announcement. Apart from this on-site meeting, we are also presenting results via conference call and online webcast. You can watch this webcast on the phone conference as well as the online platform. Apart from myself, we are actually using Mandarin. There is also English simultaneous interpretation. Before we begin, please switch your mobile phone to silent mode or switch it off. Let me introduce to you our management in attendance. Mr. Gao Nianshu, Executive Director and CEO. Ms. Huang Ying, Senior Vice President and CFO. Besides, this year is a very special year of the company. It is a year of transformation. This time we have made some special arrangements. The vice presidents of the three major business segments are invited to join.
They are Ms. Fu Wei, Vice President and General Manager of Cloud Business. Ms. Chen Haixia, Vice President and General Manager of Energy Business. Mr. Li Yunchuan, Vice President and CEO of iDigital. Today's presentation will first begin by Mr. Gao presenting the results and operations. Ms. Huang will go through financial performance. In the Q&A session, you can ask Mr. Gao or Ms. Huang questions. At the same time, you can also ask our three vice presidents questions. Now let me defer to Mr. Gao.
Yes. Good morning. I am going to present our 2024 operation results first. There are some impact from a number of factors. Our revenue declined. There is some slight increase in our profit. Revenue, CNY 6.646 billion, down 15.8%. For new business, it is down 10%. [Gross profit] margin [37.4%]. There is a slight decline. Net profit, CNY 516 million, up 0.7%.
Net profit margin 7.8%, up 1.3 percentage points. Earnings per share, CNY 0.60. Dividend per share, HKD 0.412, including HKD 0.252 normal dividend and a special dividend of HKD 0.16. Apart from the macro external environment, the operators in 2024 suppressed costs significantly. Some operators compressed their operating cost by a larger margin than in the past. For vertical industries in the past, we enjoyed rapid growth. In 2024, we monitored a lot of risk, so we took the initiative to slow down the pace of vertical industries development. This is mainly because of worry for some cash recovery or bad debt risk. We have sacrificed some market. As a result, for vertical industry segment, the growth is not as in the past. Revenue declined. The main reason is because of traditional business. BSS business declined quite a lot by 18.9%.
OSS business, basically it is flat, down 1.8% year-on-year. For digital operation business, there is 0.4% year-on-year at CNY 1.106 billion. Vertical industries, as such as now, down 30%. This is quite a big decline. As such as now, for ourselves, we took the initiative in order to avoid risk to sacrifice some vertical industries-related orders for traditional business down 18.9%. This is because operators, to a large extent, suppressed costs, and they wanted to maintain their profit. Their growth was also weak, so they put into place some measures. Some operators adopted measures to lower their cost. We also see that in order to offset the decline in traditional business, we did some moves. For example, AI is being included into the overall layout. In this way, we are able to lower the impact from traditional business operators.
For overseas operators, there are opportunities, and we actively expanded. For example, in Hong Kong, HKT, and there are operators in overseas. We increased cooperation with them. In China, traditional business was also affected, so we wanted to develop overseas business. For OSS business, there is some room in the market. There is weak decline by 1.8%. Basically, it is quite stable. We can see that our revenue declined mainly because of the traditional BSS business. When it comes to digital operation, it is rebounding. In the past three half-year periods, it has been rising. We have already overcome AsiaInfo and also AISWare pain. Business has gone back to the normal track. For non-operator area, our growth is fast. For example, in automobile business, we are up 35%, consumer up 90%, finance increased by 300%.
We have already got away from the situation of relying on single type of carriers or operators. In this way, our digital intelligence business has risen fast and there is quite a good growth. This fast growth can offset the lack of growth in telecommunication business. We are happy to see that with more data source companies like Volcano Engine, Alibaba, Tencent , [WeCom], we have deep cooperation. Together with these strategic cooperators, we serve the automobile consumer finance industries by offering data plus AI service. In the future, digital intelligence business will be our main driving force in growth. The fee-paying customers have accounted for more than 25%. 25.4% of customers are result-based. This is very good. Next, we will continue to focus on automobile consumer finance industries.
At the same time, we strive to be leading in result-based charging model. With this new model, we hope to be a leading top company. This is pay for performance. Here we hope to be a top business or company because of our digital capability, where we have Volcano Engine's data, Tencent, WeCom data, we are working together with our application scenario capabilities and AI modeling capabilities. We serve different industries. We have set out many benchmarks, and these are in automobile, consumer, e-commerce and also finance. They are top companies. Because of the customers' confidentiality, we did not disclose their names, but they have entered into good working relationship with us. In the future, digital intelligence operation will be our fast-growing direction. For vertical industries in 2024, it seems that there is quite big decline. CNY 676 million, down 30%. There are two reasons here.
First, by 2023, there were some big projects coming in, but in 2024 there was no such big project. There is a transport hub building. In 2023, large order was signed in the amount of CNY 600 million-CNY 700 million. There was big growth in 2023. There was no such big order in 2024. This is cyclical in nature. As a result, there was this impact. More importantly, our company has a mature risk control management system. We see that some projects have higher risk. We took the initiative to sacrifice those projects. The orders are in the range of CNY 1 billion. When we identify risk, we would rather sacrifice our growth speed, hoping that we can maintain a safe, high-quality development. That is why we have sacrificed some business. For vertical industries, we are focused in two big areas. 5G private network.
We strive to become a top company in the area. Second, large model, application and delivery. Basically, the cooperation with Baidu, Alibaba Cloud and so on, hoping to become a top company in these two areas. For 5G private network, energy is a leader for us. We hope that in 5G private network we can be a top company hoping to expand the strengths of energy business. For example, nuclear power, we have already built our unique advantage. 102 sets of engines. There are 25 of them with 5G coverage. For wind power, for large wind powers, there are more than 5,000. We are covering 200 and above. For PV, 7,500 large-scale PV facilities. We are covering those in six provinces with 15 [total] provinces PV coverage. The distribution is spread. There are PVs in various countries.
For mines, there are 1,000 intelligent mines of scale in China, especially the coal mines. We have benchmark companies with five cases related to coal mines. In terms of energy, coal mines, new energy, nuclear power, our 5G private network has become top companies. We hope that in the other areas we can have bigger breakthrough. For example, in power grid, petrochemical, airports, ports, we hope to be leading company in 5G private network. There are many benchmarks. There are traditional nuclear energy, new energy, as well as mines. These benchmark companies are set out. This is how we are going to expand our foundation in 5G private network. For large model, we hope that we can be a top company in large model application and delivery. In the past, we had long-term strategic cooperation with Alibaba. Apart from operators, we have developed some large-scale battlefields.
With cooperation with Alibaba, there are 30+ large model delivery benchmark companies. They are large SOE central enterprises. We have already developed 30+ benchmarks with Alibaba Cloud. We are the most important strategic partner of each other. We are the first company having this kind of joint innovation center with Alibaba Cloud. Large model will be used in various top SOEs and central enterprises. We have many benchmarks, power, energy, government. The government companies, airlines, petrochemical, oil, these are benchmark industries. In the future, apart from large- scale central enterprises, we hope to work for listed companies and private companies, especially those in the [mid-market] and also SMEs. We hope to build cooperation relationship for large models. All along, we have accumulated some very good large model delivery tools. We have tools to lower our delivery costs. We have long-term working relationship with Alibaba.
We have strong delivery team, which has long-term accumulation. These are our natural strengths and advantages. In the future, they will become our major growth engines in revenue. If you have questions related to large model delivery, our colleague is here to answer your question. In terms of R&D, we continue to invest into cloud, network and IT. CNY 906 million, 13.6% of revenue. We are a leader in telecommunications and AI. This is Gartner's analysis. We are on the top right, we are a leader. At the same time, for 37 times, we are included by Gartner in their assessment. Gartner, 37 times, IDC, 24 times, Forrester, 19 times, and so on. Examples like 3GPP, ITU, ETSI, IEEE. Altogether, there are 330 international domestic standards, including 60 new ones. We are internationally leading. Recently, we joined AI-RAN Alliance.
You know that this organization is initiated by NVIDIA. AI being added to network, forming a new organization. We are an early company in China to join AI-RAN Alliance. We are facing next- generation telecommunication network plus AI. Overseas telecommunication operators and chip companies pay a lot of attention to this area. In the future, we will continue to introduce AI- native network system, 5G private network, AI- native core network, and wireless equipment network. These are important R&D areas or directions. We will be internationally leading in these areas. For digital intelligence, we continue to be leading internationally, especially in data plus AI products. In IT, we maintain tier one position in the domestic market. IP patents continue to increase. Looking into 2025, we can say that in traditional business areas for BSS, in the long term, it is going to be stable with weak decline.
For traditional BSS, that is the case. Growth will be very difficult. However, we will continue to increase our effort of cost control to maintain our profit margin. When traditional business growth is weak, we will lower our delivery cost to maintain our fundamentals of profitability. Then we have enough strength to develop our three new business. Large model application and delivery. This is one main growth driver. 5G private network in energy, mines, oil and petrochemical, aviation and ports. Third, digital intelligence-driven operation. It has started to rebound in 2024. In 2025, we will increase growth in this segment. These three new business will bring bigger growth to us. We hope to be a top company in large model application and delivery, 5G private network, and result-based digital intelligence-driven operation business. These will be our main driving forces.
In traditional business, we will lower our delivery cost so that our organizational structure, in the past, there was high input. It will become a pyramid structure. It will change from an olive structure to pyramid structure. Some high- standard people will be the leading personnel, and there are lower- tier technicians. As a result, our cost can come down. We will bring in AI tools. With large use of AI tools, we can lower delivery costs so that we can maintain fundamentals of traditional business. There won't be growth, but we hope that profit margin can improve by lowering delivering cost. Within three new business growth, we can have better growth in revenue. In 2025, we hope to be improving in our business growth in 2025. That is my report in relation to operations.
After Ms. Huang reports on financial performance for large model application delivery, 5G private network, digital intelligence, if you have questions, we can explore in greater depth. Thank you.
Thank you, Mr. Gao. Now we will talk about financials. I will talk about cost, cash, and internal control. Here you can see some numbers which have been presented by Mr. Gao already. Revenue, gross profit, net profit. You can see that this year, operating cash flow for the whole year is net outflow, more than CNY 100 million. This is the first time since our listing. But I would like to say that behind this, it is a short-term policy related fluctuation. Our main customers are being assessed by SASAC. In 2024, there was this performance appraisal indicator related to cash. It is not that they have no money, but with this appraisal indicator, there was this impact.
In the past, receivable turnover is 20 days or 60 days. All of a sudden in 2024, it increased to 100 days and above. With such change in appraisal, we know that in 2025, the authority has optimized such appraisal. Looking at January to February cash collection, we can see this trend. Cash collection was up 30% year-on-year. This is just a short-term factor. Our [cash] exceeded CNY 2 billion. We still have a lot of credit facility. Our fundamentals are very good still. Our main customers are very high-quality customers. We think that this is just a short-term policy-related impact. You don't need to worry too much. Next, our revenue. I won't say too much because Mr. Gao already presented a lot. Let me talk about cost.
When revenue came down and the business is at a downturn, well, this is a test of a company's internal control. When business develops, can we do internal risk control and cost control? Can we make timely adjustment? If you look at our various costs and expenses, there is double-digit decline. Selling and marketing expenses down 15.7%. Later on, I will also elaborate on what we have done and what systems we have put in place. R&D expenses, we attach a lot of importance to R&D. It's down 17.3%, but investment is still more than CNY 900 million. As percentage of revenue, 13.6%. Internally, we have strategic guidance. That is, R&D should account for 13.14% of revenue following scale changes. But we have to maintain R&D investment to support future business development. Selling expenses, this is structural adjustment.
Traditional business has smaller scale, and marketing efficiency has to go up, and limited costs will have to be invested into the three new businesses. Market development. Overall speaking, selling expenses down more than 15%, 15.7%. As percentage of revenue, 7.9%. Administrative expenses, SG&A, was low. It's always below 5%. In 2024, it's up 2.3%, but it is only 5% of revenue. In 2024, we did a lot of manpower structure adjustment. We optimized our manpower establishment. As a result, administrative expenses was down more than 5%. We are very confident, even though this factor accounts for 4.1% of revenue. For this slide, this is the first time that we are introducing our internal control. As Mr. Gao said, when business declines, it is test of the management's refined management capability. If you look at business results, we have sacrificed some orders.
It is not just lip service. Internally, we do not have much arguments. When the economy is declining, do we want speed of development or quality? Do we want the present or future? We want high-quality development. There is a system mechanism to ensure the order quality. In 2024, we gave up more than CNY 1 billion. In 2023, we started the work, CNY 1 billion or above of business opportunities were given up. This is about our internal control and management. On the left, you can see that we have a set of procedures. Order quality assessment procedures. We have already got 20 big categories with 70 points and risk thresholds. According to market changes, we can tighten or loosen them. It's like a hose. When the economy is not good, we will tighten them.
With this mechanism, for each order, it will be identified online and precisely assessed. That's how we gave up CNY 1 billion. This is a way we ensure quality in terms of our business results. Secondly, how do we control cost? We are an asset-light company. Our staff cost is the biggest asset and expense. 70% of our cost is staff cost. There is a flexible and effective system. It is a rolling control mechanism. We adopted color card rules. The annual budget is divided into different quarters, different months. One month, two months, three months. If it cannot be satisfied, then it is on rolling basis. The budget will be released in this way. For example, there should not be increase or there should not be replacement. There should be optimization of existing manpower system.
In our daily production, this is already incorporated. In 2024, staff cost was down 15%, in order to maintain stable gross profit. Third part. Expenses. The way we manage expenditure has been built in the past three years. It is a one-stop official consumption platform. It includes senior executives like Mr. Gao and myself. The rules are migrated into this platform. Supplier, third-party platforms will also be connected. On this platform, usage rate is more than 95%. Official expenses decreased by CNY 80 million or more than 15% year-on-year. It is gradually online in March, April, May. It was not in one go in January. There is no cash or no bill in terms of reimbursement, so efficiency is much higher. Official expenditures can be seen on this platform, and we can eradicate some adverse condition. That's how we do cost control.
We have this mechanism to face up to management resilience, given economic downturn. Mr. Gao already talked about our gross profit, net profit, and earnings per share. They are basically stable. We can lay a solid foundation for future developments. You are most welcome to ask questions. That's all in my presentation. Thank you.
Thank you very much. Let's see whether there are questions on-site here. This lady, please. Can you state your name and your organization?
Good morning, Mr. Gao, Ms. Huang. I'm from Citi, [inaudible]. I have two questions. First, in relation to the three growth drivers, as mentioned just now, your core business will be stable, but you will be focusing on profits. For the three new businesses, they will gradually drive growth. For these three parts, how much are their share? What do you think will be their growth? That's my first question.
Second question, from this year onwards, if you talk about cash collection, receivables and inventory are quite high. How do you see will be this year's recovery or normalization? In the past two months, cash collection increased 30%. Is there some risk that will affect business results in 2025? Will there be impairment risk? Thank you.
What's the second question?
Cash collection.
Right. Thank you for your questions. For the three businesses, looking at the present, digital intelligence, the business volume is big, CNY 1.106 billion, 12% of our revenue. In the past, we had new business share. It already exceeded 30%. It's almost 40% now. For the three business, it is relative to BSS. There is also OSS. We do not talk much about OSS. OSS is part of operator business. In the future, there will still be growth in OSS.
It is different from BSS. For BSS, it is showing a decline. For OSS, there is a small part of growth because OSS is very fragmented. There are many small manufacturers, which we have integration opportunity. For new business, I think the main growth will come from non-operator business. I think it will grow faster. We talk about data operation business. Now 12% of total is share. Large model last year in 2024. I do not include operators. They have a certain scale, not reaching CNY 100 million, CNY 50- odd million. For non-operator this year, there will be a few times of growth. There will be rapid growth. I believe that there will be growth of a few times. For 5G private network, it is the first time that we disclosed the data. For core network, there are 50 sets, 12,000 base stations of various types up to now.
This is the first time we disclosed this figure. In 2025, this will continue rapid growth. There will be even more core network and wireless base stations. We believe that wireless base stations will exceed 7,000. There will be like 7,500 or 10,000 base stations. For core network, 20- 30 core networks are also in high- growth stage. These three businesses, digital intelligence businesses, will see faster growth than in 2024. There will be better profit margin. For large model, there will be a few times growth. For core network and base stations, rapid growth. These three businesses are like that. For new business, accounted for 40% of revenue. That is for sure. That is new business according to our past way of calculation.
Okay, let me talk about cash collection and cash. First of all, just now based on our analysis and prediction, 2024 showed short-term impact related to policy. From January to February, looking at our full- year estimate, it will go back to the normal level. Unless if there are other unforeseen factors. Our customers are good customers. They are trustworthy. On this point, I have confidence. In 2025, operating cash flow will go back to normal level. It won't continue to be negative. Besides, on the financial statements, you can see all these real numbers. Accounting standards are very prudent and stringent. When our customers are so good, but in 2024, the credit period was longer, so we have provided enough provision for bad debts. If you look at our credit period, within one year, customers' credit period, receivables within one year, 95% or above.
60 % of receivables are within 30 days. Within one year, 95%. Overall speaking, we are healthy. Our cash flow and also we have bank loans right now. We have very adequate credit facilities for our future business growth and resilience. I'm not too worried.
Thank you.
Thank you. Next question. Now we will take question from the phone or online. Investor online or on the phone. Operator, please give instructions.
Thank you. If you have question, please press star one on the keypad. This question is from CICC. [inaudible], please go ahead.
Congratulations on the excellent results. The results are really good. My question is about your business, your financial cost. For large model business, you are working with Alibaba, Tencent, and Baidu. You have good working relationship. For this business, what is the outlook?
Also, can you give a detailed explanation on the breakdown? Besides, concerning financial cost control, there was pressure from traditional business, and at the same time, we can see that you have exercised strong financial cost management so that your business is stable. In 2025, you are going to develop more your new business. If you do a comparison with your revenue, how are you going to control your financial cost further? How are you going to achieve your financial cost target? Thank you.
Thank you. Mr. Gao?
First question. I won't say too much about professional areas. We have large model experts here. For large model, for the three new business, they are related to the first one. I can explain more clearly about the data, and let me also supplement to the question raised by the Citi representative.
For digital intelligence business, we divide the business into operator and non-operator. For non-operator, high double-digit growth. For operators, it is low single-digit growth. For non-operator, high double-digit growth. For a large model, again, there are operators and non-operators. I won't mention operator business. For non-operators in 2024, as such as now, there's CNY 50 million. I won't include operator. For non-operator, CNY 50 million. 2025, large model, CNY 200 million- CNY 300 million orders at least. So there will be growth by a few times. So this is high- speed growth. We hope that we will exceed CNY 200 million in scale. Basically, this is our cooperation with Alibaba and Baidu. So we face large central enterprises and top companies, and in the future, we will be facing SMEs as well. We have set the benchmarks. There are many benchmark cases. For non-operators, more than 30 benchmark cases.
In 2025, there is fast replication. So this brought very fast revenue growth for the three new business, base stations, wireless base stations, more than 7,500, core networks, more than 30 stations. These are increments on top of our original foundation. So there is vast room for growth. So these are some data I'd like to share with you for cost. Let me supplement. In the future, I'll ask Ms. Huang to supplement. For delivery, it is dynamic control. Over the years, for cost and revenue, why is it that during difficult times, we still achieved good profitability? Because our business and our financials are highly integrated. There are very few companies that can bundle up the cost as well as each contract financials. So by doing this, we can see dynamic rolling of our costs.
When I see that the orders and business are not good, then I will control my costs. These two are closely related, and they are real-time. The integration is real-time. In 2025, our delivery organizational structure is adjusted as well. This is important. We'll bring in more AI tools and those people who know AI. So we will replace some personnel. The original people, if they don't understand the new delivery model, the large model delivery model, they have to understand more about data governance, data clearance, model fine-tuning, adjustment, and so on. If you don't know them, you can't do large model delivery. They are different from traditional delivery, so you have to train more technical people. Traditional delivery people, some of them can migrate, some will find it difficult. So we'll bring in some new personnel, so there will be more organizational structure change.
There are some who did original operators delivery. They will still do the same thing now, but there will be new technical people who will do new type of delivery. In 2024, we started the work. We did overall training and also AI certification before taking on work. Only with that, then technical people can continue to do product development delivery. On this, there will be increase in some costs, but in the future, our company will be better. Training, learning, replacement of old people. When we bring in new people, there will be increase in costs. Definitely, staff costs will change. But all such costs will be closely related with our projects, and there will be good integration. Ms. Huang can supplement to the second question.
Actually, what Mr. Gao said is more or less it. For our current business development, we definitely have to strictly control our costs. Staff cost is the biggest asset and cost. This is regular control and management on that basis. This year, in 2024, there is a decrease by 700-odd employees. But this is not for the sake of reduction. There is adjustment to our manpower establishment structure as well. For new businesses, concerning input of manpower, there is adjustment as well. For old business, it has come down. Staff costs or the staff structure in the past was olive shape, now pyramid shape.
With AI tools, there are the conditions to change the olive-shaped structure. We use AI to replace some people. The overall structure is pyramid structure. Number of employees does not come down a lot, but in terms of our cost and gross profit increase, there is good impact. We control in a dynamic way, and there is gross profit requirement for each business line, and we dynamically adjust the staff structure and staff cost structure and technology structure.
Thank you. Next question, please.
Thank you. Second question is from CITIC. Please go ahead.
Hello. Leaders, good morning. I have two questions. Just now, Mr. Gao also mentioned about the 5G private network. Those will be your main areas. My first question is about large model operation and delivery. In the future, your business model, the large model business model will be important. Now on the B end, in terms of telecommunications, energy and transport, what is your service experience? Now you are doing more in relation to large model. You work with Alibaba and Volcano Engine. You have deep cooperation. You have this B-end customers service capability. On the right-hand side, you have the model capability, right?
My second question is: how do you integrate your two strengths? In the future, how do you look at your large model B end delivery? Second question, 5G private network. For 5G private network in the market, while I know that you have been paying attention to it for a number of years, many companies are as well, and there are larger- scale operators. When you compare yourself with other companies in the development of 5G private network, what are your strengths? How are you going to solve customer demand pain points in 5G private network area? Thank you.
Our two VPs, they are experts in the two areas, will answer the two questions. The first question will be taken by Mr. Fu, the second one by Ms. Chen.
For large model delivery, we started quite early. Early the year before last year, we deployed the large model strategy. In April, we signed strategic cooperation agreement with Alibaba. In June, we started large model delivery. When large model develops, the whole thing evolved in terms of chips and also tool sets. There is constant advancement. It is in dynamic change process. In the process, there are some core key points. First, your large model capability understanding. In application, what are the expected outcomes? Can you catch up with the pace of development and iteration? Besides, for customers' complicated scenarios, do you have the capability to solve the problem? Large model under new quality productive forces will be very valuable. How can you seize the customer's value? This is one key point in relation to large model delivery.
Customer value will be very influential to all the complicated scenarios, improvement of efficiency, control of cost, and so on. There will be also governance of the knowledge base. All these are some applications. With advancement in large model capabilities, how much value can be brought to customers? This is our core capability, and there are a few related points. First, in the past years, for large customers' complicated scenarios, we have good understanding. Besides, we are earlier than others by more than one year. We started in this segment. We have first-mover advantage. In terms of computer and base station vendors, in large models, we have in-depth cooperation. We will pay more attention to various customers' ever-changing situation in terms of their quality and their base.
We will offer very stable application value, so that for customers' own development, we will be helping it. Then customers will set aside some costs to this industry. Then as a result, there will be a virtuous cycle. For end-to-end closed loop, this is a very good outcome by us.
Second question, Ms. Chen.
Thank you. Let me stand up to answer your question. First of all, to answer your question on 5G private network, unique strengths.
Yes, please continue.
When I comment on our strengths, I think I need to explain first. Our team is not a team working on traditional telecommunications. We used to work on power. For power and cost control, we have much understanding. Given this backdrop, our team has a unique characteristic. We know electricity, so for energy plants, we understand what they are like in operation.
When they are migrated into the network, what should be done in terms of operations? We understand these customers. Secondly, we need to understand telecommunications. All the 3GPP standards and rules. Based on existing rules, what kind of change and innovations can we generate? When retention is coming down, how can we be more flexible? It is totally different from the traditional way. Our team is an integrated team. We know energy, electric plants, and we also know telecommunications. When these two are integrated, then we have completed our innovation and transformation. This is an important background. Then the second logic. With the two capabilities, we are not only an equipment manufacturer for 5G private network in the industry, it is a complete one-stop service or integrated service.
If you only give customers a core network with base stations and software, that is not meaningful. That cannot solve any problem. For those customers, it is most important thing about two dimensions. One, horizontal. We offer one-stop service, comprehensive service. We do business advisory service. We have to understand the problems that customers have to solve. Different industries, different businesses have different demand on the network. Besides, we have to do network design and planning. We have to understand the client's complexity. It is totally different from the C network. After that, we need to offer professionalized products and on-site delivery and operation and maintenance. That is one whole set of service. It is not only software, hardware, equipment manufacturer capable of doing.
From end to end, we have to ensure that there will be sensors, intelligent terminal and newly added robots, drones, all these are smart terminals. All these have to be connected. On the bottom layer, risk control, all the way to 5G operation and maintenance, optic fiber, backup, replacement, all the way to big data production and operation. We need to offer complete end-to-end service. Then we can use that network to solve users' problems. Then they can reduce cost, improve efficiency, and they can do better management. For horizontal and vertical, we have to consider many other dimensions. So full lifecycle operation and contingency. So full coverage. Many scenarios are very special. They are not as what you have imagined. We have to do full coverage. Besides, we have to ensure that in the process, there are different degrees of flexibility involved.
These are in-depth understanding. Apart from horizontal, vertical, full lifecycle and so on, we have to solve customers' wireless and wired scenarios. We cannot only solve wireless issues. Can the optic fibers be improved? Can the connection be better? There may be a few hundred plants or sites for power companies. Can there be better connection? We offer a super compatible connection and data extraction from wired to wireless. For the energy plant sites, there are two layers of logic. One from wired to wireless, all the data will be gathered and connected. Another logic is by AI and large model, we can form our management model. The nodes in between will be less. In the future, our users will become indispensable parts in the overall intelligent network. Back to your question.
For our whole team, because we have the two DNAs being integrated, then we can achieve transformation. We are different from traditional hardware, software vendors because we offer complete horizontal, vertical services and products. Apart from the detailed products, we have also done a lot of customization and improvement. There are hardware, software that are totally different from the C network or core network equipment. We offer end-to-end service. Secondly, it is absolutely safe. It is not only telephone call, but equipment system, and there is very good value for money. It is inexpensive and 5G and integration with overall control, customization, industrial customization, made in China and so on. These are our differences from traditional vendors.
Thank you. Let me supplement to these questions because this involves our new business development. I would like to add some points. Many people may be interested, why is it that we are moving into large model delivery? This does not happen overnight. When it comes to large model delivery, first of all, we have some general advantages, and then we have professional advantages. Besides, with Alibaba and Baidu, we have long-term cooperation. General advantages.
For large model delivery, do you have advisory planning capability? After acquiring AISWare, we have the capability. Then for data clearance, data labeling, we need that capability as well. In the past 30 years, we did BSS. As a result, data was generated. CRM system, customer relationship system. There is huge amount of data generated. So we have trained our capability. How do we clear them, govern them, and how do we label? Then for large model, there is fine-tuning and refined adjustment. That is our capability. Safety of models.
After AsiaInfo became our majority shareholder, we have to make sure that things are safe. Then operation and maintenance. These are our natural strength. Every year, we have more than 1,000 projects. As a result, we had huge amounts of data, and so we have trained our large delivery capability. I think these are general capabilities. Based on these, we have accumulated some delivery tools. So you want to be fast in delivery, cost has to be low and so on. Then you have professional advantages. We work with Alibaba Cloud and Baidu AI Cloud. They are experts in the profession. So with airlines, we also cooperate. We cooperate with petrochemical and oil companies that are not related with telecommunications. So the team personnel are experts in different fields. They are not telecommunication experts. You need in-depth understanding of the industry so that you can integrate with large models.
Third, with Alibaba Cloud, we have eight years of cooperation. We have gone through many big projects. We have built very in-depth trust. In other words, we have done a lot of hard work. We have overcome a lot of pain together. We have trust. We know each other's strengths and weaknesses. How can we fill the gaps and so on? So we have both general and professional advantages and deep relationship with working partners. Alibaba has been our strategic partner, and we have set up joint laboratory. That is our cooperation experience built in the past eight years. As large model delivery company, this is not done overnight. So we have much cooperation in cloud. We have trained ourselves in different industries. We have formed our own tools and teams.
For 5G private network or private network, how come we can become a top 5G private network company? First, we have low cost. Our cost must be lower than others because many other enterprises have worked in this industry for many years, but their cost has not come down. When we entered this industry, we insisted on being low cost. Second, you have to get into the enterprises. You work on energy, you have to understand the pain points of the energy. We can't be generalized. For energy, I have go deep into the energy business. This is different from other companies. Many other companies are generalized, but I can do a better job in energy. In the area of energy, I put down my own position. For a long period of time, I work with two big customers.
We have strong awareness of the businesses. We get into their business system. The third point is we are more open. I started late. I do not want to do closed loop or a black box- like operation. Our core network can do cloud deployment. We joined AI-RAN Alliance, we joined the [O-RAN ALLIANCE], so we are more open. We are integrated into international organization standards. Number four, we have no burden. We have things to defend. In this way, I can get into different industries and companies more easily. I can satisfy their needs. These are our strengths and advantages. Externally, as I just now, on public cloud, we can deploy our core network. Our wireless equipment can be integrated into the [O-RAN] organizations. In AI-RAN and wireless network, we are the advocates and leader. I do not have legacy burden.
I do not need to think of the old products. I create new generation equipment. These are my strengths. From 2019 till now, we are in rapid growth stage in 5G private network. Thank you.
Because of time, we will now pass the floor to the online question. Please read out the question.
There are still questions on the phone.
From online investors?
From the telephone side.
Okay, operator, please.
Yes. Thank you. Charlie from HSBC, please.
Hi, Mr. Gao, Ms. Huang. Congratulations on the stable results. I am Charlie from HSBC. I have one question. Recently, you are expanding or developing international business. You have achieved breakthrough with HKT. Can you tell us the progress of your international business?
Thank you, Charlie. It is regrettable that you are not here. It is true that we are doing more with the international business development. There are two directions. First, HKT. This is in traditional BSS and OSS area. This is development of traditional business overseas. We hope that the traditional OSS, BSS system experience from China can be brought to overseas market, especially in Hong Kong and Southeast Asian market.
So that is our traditional business. For BSS, we are the number one leader in China. Internationally, we are only second to Amdocs. We are number two. We have rich BSS experience. This time we joined MWC, we realized that in BSS area we are leading already. We have a lot of opportunity to replace the original, smaller manufacturers. We can replace them with newer system, lower cost. When there is this opportunity, we will seize them, like HKT right now and other operators in the Philippines. We do have traditional BSS opportunities.
Besides for standard products being sold to non-operators overseas, for example, our edge intelligent equipment, our 5G private network equipment, these can be sold to Middle East and Southeast Asia. This is one direction that we are working hard. We adopt this two-pronged approach, hoping to develop overseas market. But our strategy is sound and stable. We did not invest a lot this year for traditional business. There may be bigger orders in BSS in the range of a few CNY 10 million. Standard product sales, the ASP may be smaller, so in the future, number of customers may be more. But ASP for selling one standard or edge equipment, the cost may only be a few $10,000 . But I hope that number of customers will be bigger buying our standardized products.
We rely on our working partners and channels to sell our edge equipment, 5G private network equipment in Southeast Asia and Middle East and in other markets, our BSS business. We have both. That is also an important development opportunity for us.
Thank you, Charlie. Thank you, Mr. Gao. Please take the last question on the phone. Next question. Next question, please.
Leaders, good morning. I'm [inaudible] [Yang]. I have two questions. One is about large model. This year, we can see DeepSeek developments and for delivery, I think that the popularity is rising. Based on your observation, no matter whether we talk about progress of operators, what is your observation? In 2025 and within the coming three years, how big will be the business volume? Can you give some views?
We have a vice president specializing in cloud and large model. Can you take the question, please?
Good morning. In June, the year before last year, we communicated with many customers on large model delivery. In the process, a lot of things happened. Technology iteration, model change, computing power change, clearer customers' demands. The whole market is in the period of cultivation and explosion. You can see that in 2024 for large models, application and delivery orders, on average, [CNY 1 million] orders. The number is limited. In the whole market, a few billion scale. In 2023, 2024, people focused on computing power and basic model. From 2025, large model application is an important large model monetization direction. On the B end, there will be rapid growth. From our point of view, we can see that growth as well. First, in the whole industry, there is full scenario demand.
Last year, when customers did budget, all of their budget is related to AI, we can feel that. If you look at top companies in a waist or the lower companies, they are moving towards the edge. Now there is the budget. Why is it that market growth is not that fast? Customers' knowledge in this area is not even. They need a lot of resources. Resource needs are complicated. The first- mover customers sometimes find that the outcome is not as good as expected. A lot of things are stuck in the delivery process. When it comes to large model, outstanding new quality productive force, how can it be changed into customer value? This is complicated. In the past one year or so accumulation, we think that first of all, delivery chain is long, from computing power to chips.
Different chips have different support effect to different scenarios, and the choice of model is very important. Is it a large language model or multimodal model? They perform in different ways in different scenarios. There are very few people with such experience when you go up the tool chain. Hallucination of large model or instability problem of large model. Every time the thinking chain is different, how can we control these issues so that we can offer to customers stable and highly effective delivery results? From customer scenario structure, if we look at the [B2C] scenario, they are simpler. For [B2B], we have, for example, Q&A scenario and other scenario, and the final delivery accuracy is very difficult. Let's say for Q&A scenario, we can reach accuracy of 85%. There is a lot of optimization and industrialization.
For large model, no matter how you adjust, it only gives 65% accuracy. For chat BI scenario, for many questions from customers, we will interpret them. Behind the customer's questions, we will do a search, and we will do different ways of verification to enhance accuracy. Behind customers' use experience, we have done a lot. For complex scenarios, like in power stations deployment, the development and operation and maintenance, planning scenarios. Also, earth, physical and seismic scenarios, model building, geological structure. These are very complicated scenarios. In these, there will be multiple intelligent chains in order to come up with a solution. When it comes to large model delivery, the capabilities, experience required will be huge as a result. Many customers in the first round of scenario building encounter many problems.
Based on our past experience, we have touched on these problems, and we have some solutions. As a result, in large model delivery, we have our strengths and advantages up to this stage. We work with our working partners to develop rapidly in order to reinforce and form our leading advantage. Thank you.
Thank you, Ms. Fu. Let's see whether there is any final question on site.
Well, I am Matt from [inaudible]. I have two questions. First is related to AI. Just now, we talked a lot about AI and new business opportunities from AI. I would like to ask, for your internal use of AI, what is its effect on cost control and efficiency improvement? How many calls can be completed by AI autonomously?
The second question is related. Just now, we said that in 2025, for large model delivery team, there will be more people in it. In 2024, how big was that team? In 2025, what is your plan? How many more people will be added? How will you do your manpower control and management? Last question is related to gross margin. Your business composition has changed. You have added the AI delivery new business. Overall speaking, when it comes to gross profit margin in 2025, what do you think? Thank you.
For AI tools, well, basically, there are two major dimensions. On the first point, product research and development and delivery, we use automatic programming tool. We have a [inaudible] automatic coding program. Based on different people, they use different programs. They can also get into DeepSeek. They can use Lingma, developed by Alibaba. They can also be ChatGPT based.
The whole thing forms an automatic coding tool, which is used by us. Different teams use them with different efficiency. On average, 10%-15% efficiency gain. When with the use of codes, efficiency can improve by 10%-15%, it is not really amazing. This depends on the experience and people in different team. But internally, we already widely use it for front office, delivery, product development, testing, and programming. In the back office, we started to bring that in. For example, staff recruitment, finance, contract management, taxation. They have started to bring in AI tools, especially with the emergence of DeepSeek for back office management and back office people. They will learn, train, and certify the use of such tools. Actually, we started last year in programming with DeepSeek.
Last year in May, we started the use of DeepSeek as a fundamental tool for coding. Many are incorporated into our systems. For back office, we started later. Manpower recruitment, contract management, legal, taxation, tax return, such tool is being used. We are in the process. Back office is later. Front, mid, and back office all use AI delivery model of tools. This is the big trend. In the future, gross profit margin will be between 36%-38%. It will be stable. That is the gross profit margin. At the beginning of new business development, gross profit margin may not be high, but when they stabilize, it will go up. For digital intelligence business, GP margin is rising year after year. 5G private network, GP margin is higher because we are talking about self-owned products.
For large model delivery, we had to do a lot of deployment at the beginning, like development of tools and also experiment points. We have to pay some costs and price. At the beginning, gross profit will be lower. When we go on large model business, gross profit margin will rise. Traditional business, gross profit is stable. There are a lot of changes, but for our manpower, dynamic cost of staff costs, dynamic control of projects, so gross profit is stable. In the long run, it will be within the range of 36%-38% in terms of GP margin. Any supplement from Ms. Huang?
Okay, thank you. Because of time, today we have to conclude our results announcement here. Thank you for your participation. I am [inaudible] Zhang. I am the person in charge of IR. You are most welcome to contact our team. See you next time. Thank you.