Ladies and gentlemen, good afternoon. Welcome to the VSTECS Holdings Limited 2026 interim result presentation. First, please allow me to introduce the members of the group's management team joining us today. They are Mr. Li Jialin, Chairman and Chief Executive Officer of VSTECS Limited. Mr. Ong Wei Hiam, Executive Director and Group Chief CFO. Mr. Gu Sanjun, Executive Director and Chief Executive Director of Greater China. Mr. Louis Lau, Group Chief Technical Officer. Mr. Li Yue, Executive Director. We would also like to thank all investors, analysts, and members of the media for joining us today. A warm welcome to you all. To begin, please watch a video highlight of the group's result.
AI's wave sweeps across the globe, while digital infrastructure development is rapidly advancing. The global IT distribution market boasts of vast capacities in the trillions with regional growth accelerating.
In Asia Pacific, a leading digital technology and ecosystem platforms in the region been listed in 2002. The group's average revenue growth CAGR has been 23%. Average net profit growth rate is 28%, both exceeding 10% and 15% for 14 years. We cover nine countries in Asia Pacific with more than 50,000 downstream channel partners, serving a population of 1.9 billion. We have deep cooperations with technology giants such as Huawei, Alibaba Cloud, Amazon Cloud, Microsoft, Apple, and HP, with anchorage corporations of more than 20 years with top 10 manufacturers. VSTECS is largest partner for Huawei in enterprise and overseas market, plus general distributor of Nvidia in Thailand and Singapore, and also largest general distributor for Alibaba Cloud. We also are the authorized training partners of Amazon. We have been developing and cultivating the Southeast Asia market for many years, achieving rapid growth and performance.
Our profit has increased fivefold in 10 years. Relying on a sound channel of networks and the profound group corporations capabilities, we continue to help Chinese company manufacturers to explore in Southeast Asia market. We work with partners to achieve rapid business growth. Our group manages the largest cloud in China, the Nationwide Cloud Environment, and we provide computing power operations and platform services to seven of them. By 2026, the self-developed computing power scheduling and management systems of our group has passed the standardized inspection requirement for token cloud service resources hub service platform of China Academy of Information and Communication Technology, and we've won advanced certifications. We adhere to refined operation and management, and we continue to promote self-developed management system, iterate, upgrade, and relied on digital means. We build compliance business online and effectively prevent all kinds of risks.
As of June 30, 2026, the revenue of our group was HKD 52.6 billion with a year-on-year increase of 15.6%. Net profit attributable to shareholders of the group is HKD 915 million, presenting a year-on-year growth of 50%. It is just the beginning. Our possibility is of immense future.
Thank you for your attentions. Up next, may I invite Mr. Gu Sanjun, Executive Director and CEO of Greater China, to present the group's 2026 interim result and development strategy. Mr. Gu, please.
Distinguished guest, good afternoon. I would like to talk about our 2026 interim result performance. In the first half, our revenue reached HKD 52.6 billion, presenting a 16% year-on-year growth. Profit attributable to shareholders recorded HKD 915 million, up 50%. Earnings per share, 65.95 cents. Return on equity reached 18.1%.
Let's look at our consistent steady growth in revenues. We have been through different economic cycle, while at the same time, we maintain stable growth of revenue. From the curve, you can see that we have some corrections back in 2022 and 2023 only, mainly due to COVID-19, where workers were required to work from home, resulting in significant increase in consumer electronics during the year, but also it overdraw the sales of next two years. After COVID-19, we returned to normal growth curve. Since listing, our CAGR has reached 43%. From the revenue of HKD 52.6 billion in the first half, we expect to exceed HKD 100 billion for the full year. Let's look at our bottom line of net profit. We have maintained a stable growth of net profit as well.
Indeed, similarly, in 2021, due to the rapid growth in consumer electronics and the subsidies of the state, which boosted net profit of the company to HKD 1.3 billion in 2021. After that, similarly, we returned to normal growth trajectory, and since listing, the average growth rate of profit reached 28%. Let's look at our breakdown. Enterprise business has reached HKD 28.2 billion, up by 9.8%. Consumer electronics reached HKD 21.5 billion with a year-on-year growth of 25.3%. Cloud computing and related services reached HKD 2.2 billion, up by 8%. By regions, North Asia has recorded HKD 33.5 billion, while in Southeast Asia, our market sales has hit HKD 19 billion, with a year-on-year increase of 43.7%. In terms of asset and liabilities, we have reached cash balance of HKD 4.26 billion, with our total asset standing at HKD 50.5 billion.
The ratio of net debt to total ratio, 0.18, is a very healthy ratio. In terms of operations, our inventory in the first half of 2026 was recorded at 60 days. It is an increase compared to 2025, mainly due to the fact of price increase in the market and the shortages of supply chains. We proactively stock up raw materials for downstream partners so that the standing inventory has increased. This will enhance our cooperations with downstream partners and also has a positive impact of our group gross profit margins. Our AR days were slightly lower than that of last year because we have stepped up our effort to recover AR. The total days of operations is 68 days, which is a healthy level. In Chinese region, we seized the prosperity cycle of the industry and achieved four-line growth.
In terms of domestic computing power ecosystems, we have recorded high growth across the board for core brands. We have been working with all equipment ecosystems for many years, and also in terms of xMacron and Huawei, [Zhenhu], we have long-term close cooperations, which has exceeded our rapid business expansions. At the same time, driven by hardware upgrade and improved computing power, we have recorded rapid growth in server, PC, and storage. In terms of core brand partners, our growth has been impressive as well. In terms of Lenovo, it's a growth of 45.3%. SanDisk even achieved a growth rate of 245%. In terms of HP, our longest-standing hardware partners, we have a growth of 39.6%. Gradually, we have been shifting away from hardware distribution to AI scenario-based deliveries. We have the support of technological certification and exclusive general distribution certification.
Therefore, we built a complete AI chain. In terms of Huawei Cloud, we have obtained the highest level of HCD expert in the world. We have also passed the certification of China Academy of Information and Communications Technology. We became the first batch of advanced certification of TokenHub platforms in AI cloud, Alibaba Cloud, Microsoft Cloud, Amazon Cloud. We won multiple awards from AI ecosystem, and in terms of Microsoft Cloud, we won the world's only award for growth rate of AI ecosystem partners. We also gained exclusive distribution right of VMware, SaaS and Veeam in mainland China. When Broadcom acquired VMware two years ago, VMware had four general distributors for mainland China, and we created a brand-new cooperation model together with VMware and became the exclusive general distributor in Southeast Asia and mainland China for VMware.
We have received most of their sales team, and at the same time, we perform most of the function on behalf of the manufacturers. In 2025, our business has reached growth rate of 30%. Adopting such a model, we also signed exclusive right with SaaS and VMware Live Recovery. In the future, we will continue to promote such a new cooperation model. In Southeast Asia, we have stabilized with a solid market-leading position with rising profit contributions. In Malaysia, our net profit growth rate has reached 55.2%. In Thailand, we reached 45.2%, Singapore 39.7%, and Indonesia 35.4%. At the same time, Chinese brands export has experienced strong growth. The year-on-year revenue increased by 48%, and the contributions of Chinese brands to our profit has reached more than 50%. At the same time, we have developed multiple new business growth drivers with solid compute-driven business.
We are Nvidia's core network partners in Southeast Asia. We are also top partners of HPE and Dell. So we have taken a large number of AI computing power demand. While consolidating corporations with manufacturers, we developed new businesses, including Starlink, Satellite Internet, PV, energy storage, new energy, and AGIBOT embodied intelligence, realizing a diversified growth curve. In terms of cloud start, we connect compute resources and models to build an AI implementation surface platforms. We have the delivery capabilities of FDE and built a differentiated computing competitive advantage. We also have upstream computing power supply. We have the adaptation of multiple mainstream large models and local deliveries. Therefore, we will provide a large number of enterprise customers with integrated delivery capabilities. In general, our business strategy is build compute foundation aggregate model, deliver industry-specific set scenario, and drive scale-up development as technology platform in AI era.
That's all from me.
Thank you, Mr. Gu. Up next, we would like to invite Mr. [Lo Shang], our Chief Technology Officer, to talk about our AI development platforms.
Thank you. Good afternoon. Today, I will update you on VSTECS technology platform, including our result and also our plans for the second half. Over the past year, the AI market and customer needs has changed rapidly. In response, we have integrated resources, in-house product, and localized services to connect computing resources with foundation models. We have leverage on our channel networks across China and Southeast Asia. Through these platforms, we aim to provide more complete and sustainable AI solutions, and therefore, we will be able to further deepen customer engagement and strengthen customer loyalty. Next, I will introduce our overall platform plans. First, let's take a look at AI technology platforms.
The market and the AI market, our judgment is that foundation model technology matures. The market is moving from training era to inference era and application era. At this juncture, our client are not just buying computing power, graphic cards, or accessing certain large marketable. Our client cares more about how AI can enter new business operations and to drive their business growth. They no longer need resources. They need to connect application services and computing power end to end, and they require a complete solution. They are no longer buying from computing power. They are buying for solutions. Based on this view, starting from 2026, we have planned the strategy of AI era technology platform. At the bottom of our strategy, we have compute ecosystems. This helps resolve and address fundamental resources issues, enabling clients to access resources conveniently and stably.
We calculate based on the region capacity and cost requirement. In the middle layer is self-developed technology and built by our ecosystem platforms. The platform aims to lower barriers for clients. They no longer need to repeatedly interface with different models. On the top layer, we have AI agent marketplace. We turned model capabilities into business applications that clients can feel and can use directly. We also have localized FDE service, which means that we will send teams to the customers, and the team will ensure that both the model, the agents, can integrate stably into the client's business. We have planned for the AI platform value chain. Computing resources provide a foundation. Model platform lowers access barrier. AI agent deliver business outcome. FDE services ensure continuous implementation. This is the plan starting from the early 2026.
In the first half, our platforms has made progress. Let's look at our resources layer at the bottom. Our compute management and scheduling software now covers major cloud providers, including computing resources in China and overseas. This create a unified management foundation across clouds, chips, and region. In terms of ecosystem cooperation, we receive Alibaba Cloud's AI Pioneer Partners and High Growth Partner awards. Also we receive industry award from AWS. We have customers include Ceres Group, Shanghai Fengshan, and Jiangling Motors Group. Through this project and customers, we have accumulated resources. On the level, this is model service platforms. We have worked with providers, including Zhipu, Kimi, and others to conduct in-depth collaborations. We have also utilized self-development model platforms for acceleration.
Starting from the early beginning of the year, we have already been promoting and launching such model service platforms, and we also worked with Huawei to co-develop a model gateway solution for intelligent computing hybrid cloud environment. The solution is being introduced in three cities and several national-level intelligent computing center. Our in-house model has made a milestone change. We have upgraded from selling computing power to selling solutions. We have worked with and other partners to gain certifications from China Academy of Information and Communications Technology. This certification has included security management, operation, to comprehensively test and verify stability. At the model service platform level, we have already accumulated implementation services. In terms of AI agent and FDE services, in the first half, through our effort, we have already built our enterprise-grade foundations. We worked with multiple partners, including China Mobile and other agencies to go abroad.
We have also established FDE resources pool across China and Southeast Asia. This has included the services for AI models, for customizations, and others. At the back, we have in-house R&D system and programming system. This is also our foundation for in-house-developed technology and risk control management system. In the first half, most important work is to fully involve AI in our R&D. Now we have realized AI coding for entire R&D personnel, starting from demand analysis, to design, to coding, and testings, and every link of it. Secondly, we embedded LLM in self-developed business processes. We have now reached a milestone. Our in-house developed systems has driven growth. First, our delivery cycle was reduced by 30%. Our efficiency has increased by 40%. With this improvement and efficiency, we have reduced overall burden, and also it enhanced our capability in delivery.
Internally, we have been accumulating our experiences, therefore, we can build product on top of it. Lastly, we have a MaaSLink services on the top level. In the first half, we have completed the constructions of MaaSLink, which has covered three mainstream models. It covers China and Southeast Asia. We have established a Chinese station and overseas stations. Now, the current model aggregation platforms from business development perspective, we have three directions and covering three types of client. Number one, we help Chinese models to go abroad. We enable our Chinese models to connect with overseas customers and relying on channels accumulated by our groups in Southeast Asia. Therefore, we help the models to enter into Southeast Asian market, providing local companies with large language model capability. Second, we help domestic enterprise to go abroad.
We help the Chinese companies going abroad to connect with large language model with equal point access. Therefore, the Chinese companies can use the leading models more easily and reliably. The third directions of our group is to serve domestic market. We have aggregated Chinese models, including Zhipu, and others. Also, we leverage our group's channel and distributor partners and procurement scale, and allowed us to provide secure, reliable, and cost-effective MaaS services. We have four core capabilities, including intelligent routing, distributor coordination, security and compliance, and channel support. The MaaSLink is now built Enterprise grade. This is unmatched by other service provider. We want to utilize MaaSLink to help model providers to gain channel access, operating support, and others. We will give customers unified, neutral, and trusted access to models. You have received a exclusive computing power card.
We have prepared small tokens for you. In total, it is 5 million tokens. I will briefly demonstrate how to use the quota. If you are interested, you can log into the platform and try it out yourself after the meeting. We will switch the screen. This is our developed platforms. You can see we have a Chinese station as well as overseas station. We have GP and cloud version. Here you have a place for models. We have accumulated more than 40 models. After the meeting, if you want to enjoy the tokens, you can register on our plan. Type in your name and email address. After clicking Next, you can input in the box below with the code in the cart. Then you can redeem the benefit. Please. Here, in addition to LLM platforms, we also have an agent development platforms in our plan.
It plays a critical role because it turns computing and model capabilities into usable applications that deliver business result. Computing resources and models are the foundation, but the customers ultimately expect high efficiency, lower cost, and business growth. AI agents are therefore the main vehicle for delivering business result. In the first half of 2026, we follow two paths for agent development. First, we work with ecosystem partners. We bring proven industry agent and scenario solutions into AI agent marketplace. With the group's channel and customer reach, we are bringing these solutions to more enterprises. Second, we also continue to develop our own agent. We can develop customized agents based on customer's workflow, knowledge, and management requirement. Both paths are supported by our AI agent development platforms.
The platform integrates model access, enterprise knowledge, workflow, tool use, permissions, and evaluations, so that our clients can customize their AI agents more efficiently. Finally, it needs to connect with the production environment of the customers. Deployment and integration is very important. These tasks require localized support. For this purpose, we built FDE service networks across China and Southeast Asia. Our FDE works directly with customers to bring platforms and AI agent capabilities to the customer's environment. We now have our FDE resources pool of more than 500 engineers covering six countries and regions and more than 30 cities, while 200 of them are based across in Southeast Asia. In China, we have teams across different cities, and about 35% of our Chinese-based engineers hold AI-related certifications. We have FDE services, including private model deployment, environment adaptation, agent and application development, and AI-enabled operations.
This team connects models and platforms with customers' business ecosystems and to provide continuous implementation protection. We have already accumulated technology platforms and initial capabilities. In the second half, we also have three pathways. First, we will expand the size and regional coverage of FDE teams, and we also increase professional AI certification coverage. We improve delivery tools, standard processes, and service systems. Second, we will continue to build AI agent marketplace. We will combine enhanced development with ecosystem cooperations, therefore expanding scenario-based AI agents, and the proven industrial models and applications will be standardized in the marketplace for rapid deployment and replication. Thirdly, we will continue to scale MaaSLink operation. MaaSLink will continue to serve China and Southeast Asia. We will increase the language models, the types and the quantities of them, to improve our services. We will also work on our operations capabilities.
All these three priorities support one common goal. That is to further connect resources, platform, AI agent, localized services, to help us to build our capabilities. That's all from me. Thank you.
Thank you, Mr. Lau, for your presentation. We will now begin the Q&A session. We welcome questions from you. Please identify yourself, including your name and organization, before asking questions. May we have the first question, please?
First, thank you for the management for giving me the opportunity. Congratulations on the company for your outstanding achievement. My question is, for Southeast Asian market, what will be your future priorities, and which segment will contribute to your profit more?
Thank you. Please. Thank you. Over to the management.
In Southeast Asia, except for Vietnam, we have a full coverage of all other countries. We worked with hundreds of manufacturers, so as long as they are leading companies in the industries, we will work with them. In this sense, there are no key account. We are a platform, so the more clients we have, the better. We are a platform that connect and serve the upstream and downstreams. In this sense, there are no priorities, but in the business focus, it will be on AI computing power. As you can see from our presentations, starting from last year, enterprise-grade computing power are actually categorized in our enterprise level. The growth rate seems low for two reasons. First, the growth rate in the last year was too high. Second, very importantly, the supply was actually very tight from manufacturers, resulting in back order. They do not have supply to us, and sometimes we rejected some orders because we could not sign the contract.
If we sign the contract, the products will keep increasing in prices, including the memory chips and GPUs. In Southeast Asia, we can deliver Nvidia server. However, there are many, many factors. If, for example, if the contract conditions are too strict, if it is not open, then it will be very difficult for us to sign the contract. Therefore, last year in Southeast Asia, there are two reasons. First, lack of supply. Therefore, there are no deliveries. Second, some are afraid of the risk of price increase, resulting in the lack of delivery. This is our focus. Any manufacturer, so long as they provide high-quality services and products, we will work with you.
Second question, please.
Dear management, Mr. Li. I am a fund manager from [Haizhu Jin], [Haizhu Yuan]. I am glad to see the outstanding performance of the company. I have got three questions.
Number one, you have talked about implementation of AI, including MaaSLink. I would like to understand your future strategic positions in the next two to three years. Is it geographical expansion and product upgrade? The second question, we know that AI demand has been quite robust. The market expect that the company will grow in enterprise system. However, in the first half, we can see that consumer electronics grow even faster. However, in our understanding, consumer electronics in China and Southeast Asia seems to be sluggish. How do you achieve that? The third question, cash flow and inventory. We can see that there is an increase of net debt, and also our inventory has increased by HKD 5 billion. I would like to understand your analysis. I do not understand your first questions on AI and future strategy.
I will hand over to our engineers to answer the question.
The second questions, enterprise level growth seems to be slower. As I mentioned, partly due to manufacturers. The top internet giants have absorbed a lot of production capabilities of the market. Our major partners, including HPE or Dell, they cannot keep up with the supply. Indeed, the market demand has been robust. However, they deliver the orders to maybe some larger customers. Also you ask, why does consumer electronics grow a lot? Of course, actually, the overall market of consumer electronics are not so good. How can we beat market average? Because we have expanded product lines. Second, Apple sales in Southeast Asia has been very strong. That is to your second question. To your third questions, inventory. Inventory increase is a strategic choice.
It is not that we have overstocked our inventories. The price increase has been very aggressive, so it is difficult to get stock. It is actually lucky to get inventory because we are buying as much as we can. There are some advantages because we have inventory and the price keeps increasing. So you buy low, sell high. For our corporations, for our partners, we are also creating opportunities for them. We are not stocking up inventories because we cannot sell them. Majority of our stock, say, for example, laptops, they are more like consumer goods. Actually, the CNY 10,000 laptops can sell for CNY 20,000. So, the price has been rising. And of course, consumer electronics like printers are declining. Hard drive, we have stocked up a lot of hard drive as well. Consumer end has risen several fold, so now it is not easy to sell.
Ultimately, the demand is there. So you need to differentiate between different inventories. This is a proactive choice. It is not a wrong prediction resulting in inventory backlog. That is basically the situations. You have another question that seems to mix. I will follow up with AI strategy. There are several levels, including computing power levels. We worked with domestic chips in China, say, for example, Huawei and Qualcomm, to expand our coverage in chips. On the model level, we worked with CSP and model service provider, as well as cloud lenders like Alibaba and Huawei. We promote model aggregation platforms, as we presented before. We aggregate models, including domestic models and global ones. Those models are all from leading manufacturers, and we will promote them in our own channels. That is how we expand into models business in the AI era.
On the upper level, we have agents, application agents. On the agents level, we are more about ecosystem collaborations because there are many verticals and industry-specific scenarios. Every industry comes with their own capabilities, strengths, and weaknesses. So on the agents level, we primarily engage in ecosystem collaborations. We will also integrate various industries and verticals to provide cost-effective applications of agents. We also have applications, models, and computing powers and integrate them into our AI strategy. This is the strategy level, but of course, in implementations, we require on-site teams for implementation, maintenance, and operation, so to ensure the success of our strategy. Because AI is a large language model, it is a probability model, not 100% correct, so it still requires manual debugging, approval, and maintenance. We have been building our service teams to ensure the operation. Basically, these four layers.
Thank you, Mr. [Long], for the explanations. Next question, please. Thank you very much for your patience. Please go ahead with your question.
Thank you. I come from Hong Kong, from a European retirement pension funds. We have a view and three questions. We have been a shareholder before. We wanted to buy, but we are hesitating now, so we have several questions for you. In 2025, your revenue is HKD 97.6 billion, out of HKD 100 billion. It is an increase of 10%, net profit HKD 1.39 billion, an increase of 28%. If you break it down, you will discover three aspects. Number one, the business is a low-margin distribution. It is squeezed from both ends. You say both ends, but I am not sure if it is a correct understanding in your business. You have upstream, so for example, Huawei, Nvidia, Dell, AWS. They have the pricing power, and downstream, there are large customers, like government and enterprise.
The payment cycle may be a little bit passive. Secondly, your operating, your working capital is large, and it seems to see some interest-bearing debt pressure. It started to rise. Then, GP margin, 4.5%. Net profit is -1.4%. This is a distribution business. Its abilities to counter-fluctuation is weak. The last question, which is also our primary concern, it seems that cloud and AI is your second growth curve, which seems to take up a large percentage of your revenue. However, it is not so significant because it has not reached revaluation standard yet. Though the revenue is HKD 5 billion, it has increased, the total proportion is still low at 5.2%, so it is relatively small. It is very difficult to hedge your distribution business. In my conclusions, you do have revenue, and you do have institutional support.
You have many institutional investors, but the point is how to turn the billion-dollar business into free cash flow, double-digit growth, profit rate, it still requires a leap forward.
Maybe this is a small industry and not so significant, so there are no specific industry analysts to develop business analyst models to evaluate our business. Therefore, you all use a unified model. Say, for example, high GP margin represent growth potential in the future. Low GP margin is considered to be future. I believe it requires different teams to study different types of companies. You need a specific coverage and analyze with a new perspective. In our industries, we have been in the business for many years. High GP margin does not necessarily equals to a good industry. For example, in China, software industry with a 60%-70% of GP margin, however, few companies make the money.
Therefore, high GP margin does not necessarily mean high return. For simple example, cybersecurity industry in China, basically no company is making money. Maybe only one of them makes money, but they make money from side business, not their main business. Therefore, I do not think that high GP margin represents no risk or high potential. Yes, we do have GP margins. However, this is normal for the same industry. If some company have high margins and other companies have low margin, of course, then you have high risk. But then we are looking at different industries. Different industries come with different situations. For our industry, our GP margin is normal. It is not too high and too low compared to market average. So you need to compare apple to apples, because for different types of company, it is very difficult to benchmark.
For example, we have different business models, different turnover rate, and it is not comparable. Say, for example, if the GP margins, let us say, for example, you take two months to make the sales, then you have six times of turnover a year. Then you need to do the compound because your profit is calculating based on annual basis. So you just multiply by six, then you have six turnovers. It is not high or low. If you are in real estate, you make one sales in 30 years, then if you have 30% of GP margin, it is still very low. In this sense, I would like to give you a view. Indeed, from absolute point of view, this is a low GP margin, so you may think that it represent poor risk resistance. That could be intuitive. However, we have been listed since 2002.
As Mr. Gu mentions that our ROE has never been below 10%. It is actually 18% in the first half of the year. Facts have proven that there is no significant risk. I believe I need to elaborate with details to your questions. Number one, there are different stakeholders in the upstreams. Indeed, we have corporations, but of course there could be some conflict of interest. Anyway, 90% of our upstream manufacturers, they are not just in a buyer-seller relationship with us, they are our partners. We are all partners in the ecosystems. They do not squeeze us because many manufacturers have given us exclusive distributor right. We have exclusive agency for them. Of course, at being a general distributor, we act as their external partners. We help them with market coverage, price management, and market research.
If we go down, it would be very harmful for the manufacturers as well. Therefore, of course, the pricing in this year has been quite different. It has been reasoned. Before that, customized products have been declining in pricing. Of course, in the past, I do not even dare to hold inventories at all, because if we buy too much in advance, then it could easily lead to losses. For example, if you buy a phone three months after launch, the price of that will drop. Of course, some manufacturers provide us with price protection policies. Yes, of course, we can import, we can buy, we can trust the manufacturers. They will give us various price protections. In some extreme cases, they will give us price protection for two to three years. It is not to say that they love us. In fact, they need us.
It is like choosing to get married. Divorce-ish. It is for the both side of it. It is the way we work with manufacturers. Indeed, some companies will hold different business styles. They have their own values. Some may think that you earn too much, you need to give away. But some manufacturers are very fair, so it varies. In general, we are partners in the same ecosystem. Put it this way, the manufacturers, they themselves make a lot. Nvidia, they have 80% of GP margins, but that has nothing to do with us. When they lose money, they still have to pay for us because we have served them, we worked for them. When they make monies, when they make a lot, we also become fortunate because, for the same thing, we work with them. This is a cycle.
For Apples, we can clear out the stock in one week. Sometimes we have a shortage of inventories. We are waiting for the stocks to arrive, and then Apple gave me a two-month billing cycles. We have not received the goods yet, they give us money. Then the GP margin is very large. It is infinite. In our business, we invest very little in terms of equipment. We do not have a high depreciation, we do not have a high personnel wages, and the sales volume is relatively large. Therefore, our GP margin is reasonable in this sense. Whether the GP margin is reasonable or not, it will finally reflected in ROE of our capital investment. We have an ROE standing at 15%. Of course, our net profit margin is only 1% or so, but it will double for any company.
If their ROE is 30% or 50%, can they make that? Is it sustainable in the long term? We have maintained at 15% for a long time. It is even increasing, so it does not matter. We have a net profit margin of 1%. In summary, our GP margin, it is not necessarily low. If it is too high, the manufacturers will not be happy. Secondly, we are a platform company. Our core value is to reduce operating costs for manufacturers. If our profit is too thick, the manufacturer's product will have a very poor performance. The most important thing is to improve operational efficiency, and our long-term direction is to reduce the margin. While at a lower GP margin, we can achieve higher returns by improving operational efficiency. In this way, when competitors enter into the market, the manufacturers will even love us more.
They will favor us, and they need us. This is our direction. Also to the second aspect of questions, higher gearing ratio. Actually, it will decrease. It is only normal for us to use more capital because we have increased inventory, and our scale gets bigger and bigger. Say, for example, in China, our interest rate is quite similar to that of SOE, and we do not require any personal guarantee because banks have been working with us for many years, and they have been treating us as a reliable partner, so we have low interest rate. For cloud and AI business, yes, they are new business. The cost of development is high, so it is not profitable, but high valuation is another thing. AI business makes a lot of business, but of course, software growth rate, growth margin is high.
However, we still need to work for delivery and development. If we do not develop new products, actually, we will make money. We will make even more money. It is not sustainable in the long-term future. As we sold our mature products, we developed newer products that are not necessarily profitable now. Why do we have to do it? First of all, the technology industry is fast-changing. We need to be able, be capable to introduce new services and product to the society, and we need to keep pace with the latest trends in the industry. That only serve our long-term development future. It help us to identify market trends. It is beneficial. Actually, it is quite rare in China. We are not loss-making for IT development. This is quite rare in China.
On the other hand, for our company's development strategy in the long term, we believe IT industry is a fast-growing industry, and really is the largest industry in the world. Human, in the future, will become more informational-driven and become more intelligent. This is a huge sector. Large customer manufacturers may be able to do it by themselves, but there are also many long-tail needs. The manufacturers will develop their product, and ultimately, they will be able to sell for the same sector, say for example, hard disk. If you have a same performance, of course, the lower your price, the better. Actually, there are actually some long-tail channels. We also have product. We have been developing many customers. We have accumulated many customers over the years. Manufacturers, or at least a majority of the customers of our manufacturers, do need us. This industry is booming.
It is large scale. Our main business is distribution. If nothing new developed, it is still worth engaging in distribution. We are number eight in the world, and the largest distributor is four to five times larger than us. I firmly believe that one day, China will become the most powerful technology power, and by then, we should be even bigger than the existing international top distributors, and there is a huge room in the future. It has been reflected in our paper as well. We have been growing by 20% annually in recent years. In fact, we will upgrade ourselves at the same time, while not giving up on our main business. As we know that the road to the unknown future, the exploration is arduous. Just like Alibaba and Tencent, they invested a lot in large models.
However, we believe that the future of mankind, we are willing to give resources and investment for them. It is the same for us. We will continue to invest in AI. It is a long march, so we do not expect much profit in the short term. We are not radical, aggressive in our transformation. We are strategic. We are being strategic. We acknowledge that there is a need for transformation. We need to improve our capabilities, and we need to expand our boundaries by tapping into new sectors, and that is within our capability. I am not sure if it answers your questions.
Thank you.
Thank you, Chairman, Mr. Li. Next questions, please.
Hi, Mr. Li. I am Xia Xueling, reporter from The Times Weekly. Congratulations for achieving a good result. My questions. Could you please talk about the opportunities and challenges brought by AI to distribution industry?
I will answer the questions briefly. First, for challenges. Customers' demand and the technology complexity has been increasing, so we need to gain more technological certifications. We need to develop new delivery solutions and adaptations, and to reach higher thresholds and requirements. So for our personnel, for qualifications, we encounter challenges. Also, large language models has been squeezing software markets, say for example, Adobe and so on. Previously, they require software, but now they are relying on large language models. So some sector will become smaller and smaller, while LLM has gained larger market share. So here, in us, our value lies in actively exploring how to build aggregated service platform for LLMs and FDE services. We also developed agents to create more values. For opportunities, as we mentioned before, we believe in large language model aggregation. This is our opportunity.
We also help our clients to finish the last mile of AI lending in implementations. We have FDE services. Also, in terms of hardware, for many companies, our clients, when they are implementing and lending AIs, they use large language models in the cloud. But with increasing maturities of inference, some companies may use end devices out of security concerns, and because they have their financial data, they have personal data, and it is not so convenient to upload to the cloud to the LLMs. So they may require private model On-prem models, they use API, PCs, AI workbench to cover more end devices. They have this kind of upgrade and replacement demand. I would like to supplement with AI. Of course, hardware combining with AI, this is a new thing. For any company, after adoptions of AI, your efficiencies will increase.
In the era of AI, IT industries will develop rapidly because we have enriched product and services. In this sense, as channels, we have opportunities to capture. Also, on the other hand, AI helps us to be more professional. For a proper use of AI, it requires stronger professional capabilities. So we also need to have relevant reserves. If a company does not learn and iterate, and to keep up with the progress, then it will easily be eliminated in the new era. Of course, all humans, and the foundation of organizational management, it's down to information. With AI, the way we access information is more democratized. So if a company does not change its organizations, it may easily fall behind in the short terms. In summary, I believe AI reminds us to keep learning. Do not stand still.
Indeed, opportunities abound, but if we do not progress, then we're definitely eliminated. AI has been advancing rapidly, so we need to learn and to make progress.
Next question.
Thank you. I'm glad to see such a good result from the interim report. I'm from Huaxi Securities. Actually, this is the best interim result among many years. In conventional manner, for analyst, I actually think that it is abnormal. Personally, maybe I think that there are many reasons, maybe due to the overall price increase in the market, or there are timing issues due to your inventories. You have made good preparations in the eras, but that could be double-edged. The fluctuations of your result may also be heightened. So my question for the management in this situation, do you see a double-edged sword? Do you have a better means to turn it into a more smoothened result?
I actually agree. I agree to you. I don't believe in a 50% year-on-year growth can be sustainable, because it's like a child, from one to 10, then you have a high growth rate. From 18 years old and onward, can you grow rapidly every year? I doubt it. We were established back in 1991, so it's been 35 years now. We have been on track. It is impossible for us to enter into another high growth length. But we manage our business very well with a stable profit growth. 20% of growth, it is already satisfactory to the management. So for this year, we have a 50%. Is it purely due to shortage of goods or price increase? I believe partially, yes, but it is not all the story. We believe in our operations. Our industry, one of the core advantage is operations.
Our operations is the best, in our understanding, among all listed companies in our industry. You cannot find another one to be able to maintain such a stable growth spanning a 30-year cycle. If you look at companies from the U.S., from Chinese mainland, and Taiwan, you can see that they may fail to respond in some years, resulting in fluctuations, but we have been performing well. We have built an ecosystem, and we are very quickly in our response. Without such a capability, it will be impossible to maintain such a stable growth. Our product line is very big. It is not reliant on one person. It requires a system and a set of rules. This is a joint effort of the entire companies. We have so many product, and none of them set the price because of me.
We have a team that is very mature in the industry, and our Company culture has been very good when it comes to trust. Trust is the biggest factor in reducing operational cost, because if there is no trust in the company, you will be hesitant to authorize meetings, delegate your powers. But in our company, we trust. People trust the boss, and I trust them, too. It would be hard to imagine the growth rates to be this high. Do we need to set higher KPIs? No, we never set KPIs. I simply tell people, "Just do your best." There are many external factors affecting our business, so it would be difficult to set KPIs. If we set it too low, then it will be no growth. If we set it too high, then it cannot be completed.
I tell them, "Just do your best and work hard." I trust them. I trust them to deliver. Because if they cannot complete their KPI, their reward for the year will be gone, and they will be afraid of us. Many people think that culture is abstract, but think about that. During the civil war, Kuomintang against Communist Party of China. Why Kuomintang lost? Their capability should be very good. They are definitely stronger. Many of their leaders graduated from Whampoa Military Academy. They trained. However, the soldiers of the Communist Party are mainly from the uneducated. They are farmers. But why do they win? Because they have a different culture. They have an organization that wins. Another example, let us think about diamond. The diamond happened because all the atoms are lined up in a very regular way. However, if it is lined differently, it is not diamond, it is carbon.
I believe this kind of connections are culture, they are organization, and it is important. It is important in an environment. Last year, we have, I believe, about 100 million HKD or so of revenue from hard drives. It is caused by price increase of SanDisk. Actually, the revenue in China has been high and the risk control capabilities has improved. Another thing, we have better capability, and we gain share from competitors. In Southeast Asia, in Thailand, we ranked third three years ago, but now it is number one. In the era of the lack of inventories, the top companies tends to get more stocks. Actually, three factors combined, operations, which keeps increasing, and also culture, and also the head effect, which means that the manufacturers support more when they have a short of supply. They supply better to the top companies.
So all these three combined. It is, of course, the price increase, as you mentioned, it helps as well. But in the second half, can we deliver another 50% of year-on-year growth? It is hard to predict because we do not set KPI. So we cannot give you a concrete data. But based on calculations, more than 20% is definite. Possibly, we can reach 30% for the full year. 50%, it could be down to a series of luck coming together, so it is hard. It's a moonshot. But in the long term, you may have a limited understanding of our industry. For our industry, 20%-25% growth rate, it is already very good. So we cannot have higher or too high of an expectation. This year may be an exception. So thank you very much.
Thank you. Another question, please.
Hi, I am from Huaxiong Group. My name is Wanya. I've got two questions. First, distribution business of storage. Could you please elaborate? Second, how do you view the impact of price fluctuations on your performance? Also, I've got another question about cloud business. In the first half, cloud business grew by 8%. However, the general business grew by 8%, and the segment profit increased by 33% in the number. Gross margin of cloud business increased by 4%. So I would like to compare with your growth rates to other leading cloud vendors. It seems to be out of line. Why? Also, the improvement of margins. Where does it come from? Does it come from increase of rebate from cloud manufacturers, or does it come from the proportions of cloud software business?
I'd like to get your questions. As a general distributor, we do not have the initiatives to change our strategies here. The GP margin and the growth of cloud business are affected by several factors. Say, for example, Alibaba Cloud have a frequent change of their strategies of their business lately, and their policy has changed significantly as well. So to be prudent, we need to be conservative in such a context. So there's not much growth. But Huawei has grown significantly. However, for our company, we started to do Huawei Cloud business relatively late. We adopted expansionary strategy, which means that we are willing to accept lower margins. Now, previously we did not have any assessment, but now our position is relatively stable. So now we have been assessing the profitability in this business. So we give up orders with lower margins. In short, the growth rate of cloud business has declined. However, we are already the largest distributor of Alibaba Cloud for Huawei.
They adopted other distributors, including China Software and Runton. We believe our position is relatively stable, so we care more about the profit. We care less about market share. For storage, we have big brands, three big ones: Western Digital, Seagate, and SanDisk. Actually, there will be a shortage of inventories for another two to three years. Actually, only less than 60% of market demand has been met, so definitely there is a shortage of supply. Large-scale hard drive, for even consumer levels, well, the price has been too high, so the willingness to purchase has been declining in the first half. We have this data here. Our revenue, our sales of hard drive has increased. It's increased by about 100%. However, the number of sales has decreased, indicating fewer individuals are willing to buy, despite the fact that the turnover has increased. How about in the future?
In the future, this is a structural issue. Large language model. On the consumer side, everyone wants large language model. There is a shortage of computing power, so they want to seize the land, and grab the opportunities regardless of cost. There are sales despite the price increase. On the other hand, global economy is not so good. After such a big increase, the purchase intentions will definitely decline. All these two combined, the final balance will be reached. There is still a shortage of supply. There are issues with the supply chain. In the next two to three years, it will still be very painful, but we still will be able to make money.
Thank you, Chairman Li. Any other questions from this side? Please pass the microphone.
Hi, management. I am [Xintong Jian] from Hong Kong Caihua Media. First, congratulations from the outstanding performance.
Just now, Chairman Li analyzed from the industry perspective, you talked about market growth and corporate governance. My question is, compared with other ICT distributor, what are your core differentiated barriers?
We do not have any differentiated barrier, but the market barrier is high. For example, Nvidia is there, but it is very difficult to build the second Nvidia. This indicates that the market has a very high barrier to entry. If there are no barriers, then definitely 80% of the models will go bankrupt. In Internet business, the one can raise funds will survive. For example, Jack Ma raised the fund. It is like chosen by God. The one ultimately selected by the capital markets will survive. For example, Meituan are lucky, they succeeded. History does require such companies. For our industry, the barrier to entry, actually the window to gain exclusive distributorship has passed.
It was about 30 years ago. It took a lot to build sales network. In the past, we had a very high GP margin, say for example, 30% to 50%, but now it falls to a lower level. If you sell only a single product, you will definitely lose money. Relying on a single point of entry, then of course it will be unrealistic. That is the first part. The first barrier is time window. Second, we have a lot of accumulation in upstreams and downstreams. We have upstream manufacturers and downstream customers. They trust us. We have established long-term cooperations. This is a very high barrier to entry for our unique ones, as I said. In the world, it will be very difficult to find another company that within so many years since listed, will be able to maintain such a sustainable growth.
We experience bubbles of tech stocks, we have experienced financial crisis and all kinds of turmoil, including the crisis in Southeast Asia, and product iterations. In the past, notebooks were the only mobile terminals, but now it is replaced by smartphones. The technology has changed. Without a flexible enough of procurement plan, then you definitely lose money. It is not because you are stupid. It is because of a rigid system. Say, for example, if each department growth by 15% and the industry is growing by 20%, then you are falling behind by 5%. For our company, we have a unique advantage that is to really accumulate and build an ecosystems, and we built a trusting organization. How to put it? Actually, we have been delegating power to the frontline soldiers. As [Yuan Zhongfei] said, actually, we have been demonstrating that many years ago.
Otherwise, we would not be able to maintain such a good profitability every year. We are highly decentralized. With a decentralized organization, our supervisions at the back ends will be better. Otherwise, it will be out of control. We have a high degree of decentralization. We do not manage by KPI. To your questions, what kind of unique barriers to entry or competition edge? We cover different industry. We have networks that connect with different industries and different countries across China and Southeast Asia. We worked with 700 world-class companies across different industries and different countries. We have good organizational capabilities and corporate culture. We are humble. If I make a mistake, for example, I will admit to that. Many companies won't do this. If the boss is too proud to admit mistakes, then the companies will not be able to make progress.
All progress comes from admitting one's own shortcomings. We are humble. We were engaged in P2P. We tried that before. We also applied for an internet lending companies license in Chongqing. We spent CNY 70 million developing measurement models, and we loaned money. However, we shut it down in March of 2018. I approved the project, and I also made the decisions to close that. We spent seven months on it and ultimately decided to abandon it. We invested CNY 70 million. It was a hard call, but I made the decisions. I know that, but it still cost us CNY 70 million. If you don't have money, then it will be done. We built a platform for everyone, allowing employees to learn and grow. We allow people to grow. For other competitive advantage, I believe it is still down to organizational capabilities.
We are definitely able to learn and continue to improve.
Thank you very much. Next question, please.
Hi, management. I am [Oying Ying] from Wilson Securities. I am glad to hear from the management. I can feel that the company is very pragmatic. My question is, capital allocation priorities of the future. Is it on the expansion of domestic computing power channel? Second, M&A in Southeast Asia. Third one, dividend of shareholders. For these three, how do you put your priorities? Could you please do a ranking?
I have been increasing my holdings of the company stocks. I had never sold it, so I need money myself. Definitely, in my best interest, we will make dividends. I am aligned with everyone else. The dividend payout ratio is between 35% to 50%. In the past, we acquired ECS, and back in the days, our gearing ratio was too high to pay dividends.
But I don't think money is a problem for us for acquisitions. In 2008, when we acquired ECS, we had CNY 1 billion of our funds, but we spent CNY 1.4 billion. The bank trusted us. They gave us loans. We are grateful for the bank support. If there is a very good target in the markets and we are willing to, we are open to it. We ourselves were worth, we evaluated at CNY 1 billion, but we acquired a company with a valuation of CNY 1.4 billion. So, if we believe we find the right targets, then we will do it. But of course, we need to assess the risk. But we don't make provisions of money just for acquisitions. There is no such a plan. If we encounter a good opportunity, then we will seek financing. The company has been growing very fast.
Our industry, starting from distribution system, then we developed tech products ourselves. Apple happened in 2008. Apple has a distribution network abroad. They trusted us. They gave us credit. They supported us to develop channels. It is not a spot sales. Giving credit comes with risks. It is just like banks. What if our downstreams do not pay us back? Therefore, it challenged our risk control. We have been growing very fast. If we grow by 50% this year, I do not believe it will be able to be sustainable, because you need to have good clients, you need to accumulate clients, and the market is about survival of the fittest. Of course, if you bring in new clients, you can resolve part of the problems. Of course, the industry has a ceiling. It is not so easy to grow fast.
Second, we do not want to take too many risks. Therefore, the bottleneck of development does not lie in capital. It is not about capital. It is a natural growth. We do not need to consider too much of capital allocations. We do not have a capital strategy.
Another question from this audience. Thank you. Please.
I am Guotai Haitong Overseas Technology. How do you see MaaSLink business? What about the long-term market potentials and also strategic positionings of your development? Also about the reasonable GP margins, and also core competitiveness development advantage. In the long run, will you consider disclosing the ARR, and the annualized return and GP margin of MaaSLink?
Of course, if it is promising, it would definitely be beneficial to us. I will not say that any company, if they go bankrupt, say for example, if Huawei goes bankrupt, if Nvidia goes bankrupt, I would hope that there is no impact to our company. We want to be the channels of the whole industry, not one single enterprise. If they go bankrupt, we will continue to serve others. Therefore, we do not want to make any product the only main thing of our company. Of course, we will follow the trend of technological development. For example, now AI is our direction. That is a correct one. We worked with D-Link mainly in Indonesia and Malaysia, while Vietnam, Hong Kong, and Thailand are not willing to give us the space. First, the country should be open. D-Link now accounts for a relatively low proportion, and it is not expected to be too high in the future. D-Link itself is doing well.
Whether it incurs loss or enjoys high profit, our GP margin remains fixed. Our GP margin is relatively high at the time being. Due to the shortage of supply, we do not invest too much, and the profit is good. I believe the true value of our company in the future is to connect more with upstream and downstream. Scale is our value. Revenue is our value. The larger the scale, the more needed we are. As I mentioned before, what we pursue is ROE, and now our ROE is at a reasonable level with higher ROE than ours. Our ROE may increase because of scale increase, application of AI technology. With AI, we have lowered our costs. Our labor cost has not increased, but the scale has gone up. Starlink will not be disclosed separately in the future.
In the future, we will continue to disclose by sector. We want to be a channel for the industry, for a sector. Our value is to reduce transaction cost for the industry. With a lower and lower GP margin, I hope that our ROEs will still be able to increase, and that indicates a more valuable company.
Any other questions from the audience?
Good afternoon. I am from [Guosheng Jihua] Fifth Road of Guosheng Securities. Your revenue is good, but net profit margin is low, less than 2%. Investment return, it seems to be zero. I was encouraged in view this is ourselves. My question for you, do you aim to expand investment income? Say, for example, this month, ChangXin Memory Technologies hit IPO. Do you participate in IPO subscription? This is safe and fast. Your company has a capital. Why don't you prefer to try?
Mass memory is preparing for IPO as well. You're so familiar with your industry, so why don't you try it to enhance your profit?
To be the lucky ones and to get rich overnight, but it will not be sustainable. Getting rich overnight, it is scheming. It is based on unfair exchanges. Some companies don't actually create much value by the society. They hype it up only to gain interest. They do not provide a deep insight, but still it leads to high valuations. We are not Zhuge Liang, but we act like ones. I believe with time going by, their pricing will fluctuate, and they will come back to where it should be. I don't understand you, Xu. First, I strongly oppose humanoid robot. Society progress is driven by division of labor. With division of labor, specialization occurs. However, humanoid robots goes against the logics of industrial development.
Actually, if you invest trillions of dollars in humanoid robots, they will not excel compared to human. So it happens in scientific explorations. Human explore all the times, it fails 99%. There won't be many successes, but of course, humans needs dreams. We need to develop some special love for new things, otherwise we cannot progress. I'm very rational. I will not indulge myself in selecting those startups.
Any other questions?
Hi, management. I'm [Chen Zhongying] from Guohai Overseas. The company has equity incentive. What about the CAGR?
Everyone said it. We said about increase by more than 20%. Now we have reached 50% across different sector. We have been implementing that. We have been implementing and allocating the incentive based on their contributions. I remember the threshold was up. There are different thresholds and different threshold for different market in different countries.
In your questions, actually, people ask, is it profit growth due to SanDisk or price increase? Actually, we don't sell them in Southeast Asia. To another questions, supply chain. If you go for hunger marketing, then of course you may gain more profit in Southeast Asia. AI server has a lack of supply because they are supplied to the U.S. Thank you.
Any other questions? Thank you very much for your questions, and we would like to thank the management for answer. We have come to the end of today's conference. Thank you again for your participation, and thank you for your interest and support to the company. We look forward to seeing you next time. Thank you