Kingsoft Cloud Holdings Limited

Good day and thank you for standing by. Welcome to the Kingsoft Cloud's fourth quarter and full year 2022 earnings conference call. At this time all participants are in a listen-only mode. After the speaker's presentation there will be a question and answer session. To ask a question during the session you'll need to press star 1 and 1 on your telephone and you will then hear an automated message advising your hand is raised. To withdraw your question please press star 1 and 1 again. please be advised today's conference is being recorded. I'd now like to hand the conference over to your first speaker today, Nicole Shan, IR Manager. Please go ahead.

Thank you, operator. Hello, everyone, and thank you for joining us today. Prince of Wales post-quarter and full year 2022 earnings release was distributed earlier today and is available on our IR website at ir.ksyun.com as well as on the global newswire services. On the call today from Kingsoft Club, we have our Vice Chairman and the CEO, Mr. Tao Zou, and the CFO, Mr. Henry He. Mr. Zou will review our business strategy, operations, and company highlights, followed by Mr. He, who will discuss the financials and the guidance. They will be available to answer your questions during the Q&A session that follows. There will be consecutive interperations for your convenience and reference purpose only. In case of any discrepancies, management statement in the original language, will for will. Before we begin, I would like to remind you that this conference call contains forward-looking statements within the meaning of Section 21E of the Security Exchange Act of 1934 as amended and as defined in the U.S. Priority Security Litigation Reform Act of 1995. These forward-looking statements are based upon management's current expectations and current market and operating conditions. And they relate to UN study about known risk, uncertainties, and other factors, of which are difficult to predict, and many of which are beyond the company's control, which may cause the company's actual results, performance, or achievements to differ materially from those in the forward-looking statements. Further information regarding this and other risk, uncertainties, or factors are included in the company's findings with the U.S. and D.C., The company does not undertake any obligation to update any forward-looking statements as a result of new information, future events, or otherwise, except as required under equitable law. Finally, please note that unless otherwise stated, all financial figures mentioned during this conference call are denominated in RMB. It's now my pleasure to introduce our Vice Chairman and CEO, Mr. Zhou. Please go ahead. Hello.

Welcome to the 4th quarter of 2022 of Xinshan Cloud, which is the year-round performance change conference. The past year of 2022 has been extremely busy. In the face of a complex and diverse management environment, Xinshan Cloud has successfully passed the test. Since becoming the CEO of the company, we have maintained a high-quality and sustainable strategy. Hello, everyone, and thank you all for joining Kingsoft Cloud's fourth quarter and fiscal year 2022 earnings call.

2022 was an extraordinary year in many ways, and we're pleased to have successfully navigated the various challenges we faced in the complex and dynamic environment. As I took on the CEO role, we have remained committed to our strategy for high-quality sustainable growth while continuing to building success based on technology. We have also implemented cost reduction and efficiency initiatives, which have resulted in steady improvements to our profitability.

Specifically, this quarter, after adjustment, the net profit reached RMB1.69 billion, which is four times higher than before. After adjustment, the net profit reached 7.9%, which is 6.7% higher than before. The operating cash flow reached RMB3.7 billion, Since the second quarter of 2022, we have obtained cash flow for the third consecutive quarter. For the first time, we have implemented a free cash flow transfer in the third quarter. These initial results show the strong willpower of Jinxiaoyun, which makes us able to raise funds again.

I am pleased to highlight some of our notable results with you. In Q4, we saw a remarkable increase in our adjusted gross profit, reaching RMB 169 million, representing a fourfold increase year over year. Adjusted gross margin increased to 7.9%, rising by a significant 6.7% from the same period last year. Furthermore, our net operating cash inflow amounted to RMB 317 million, marking the third consecutive quarter of positive net operating cash flow since Q2 of last year. We also recorded a quarterly free cash inflow for the first time, which is an important milestone for us. These impressive results demonstrated our strong business resilience and provided a solid foundation for us to stabilize, restart, and emerge even stronger.

In terms of supply and demand, this quarter's revenue is RMB 3.5 billion, which is a steady rise. We have identified the appropriate strategy for the head and waist clients respectively. For the head and waist clients, we have selected high-quality clients. In order to balance income and profit, we create the best customer experience and full-fledged reputation. As for the clients, we use the reputation and energy-saving core capabilities of our top K clients to drive the clients' business opportunities. continue to reduce the dependence on the top K customers. At the same time, drive the increase in the income and profit level of Gongyouyun. In the past six months, we have systematically sorted out 11 segmented industries, and evaluated more than 200 companies. In the end, we selected and newly signed more than 30 high-quality subcontractor customers. With the success of these strategies, the income ratio of the top three customers in Gongyouyun continues to decline. Now, I would like to provide some updates on our progress in three areas, namely public cloud, enterprise cloud, and research and development. I'll start with public cloud services.

In Q4, revenues from this business remained stable at RMB $1.35 billion, representing a slight increase from the third quarter. We fine-tuned our positioning and implemented differentiated strategies for key account customers and mid-size customers. With Key Account customers, we strive to maintain a balance between revenue and profitability while delivering the ultimate service experience to establish a superior word-of-mouth reputation for our full-stack solutions. We leverage this strong reputation and the scalable core capabilities we have developed to serve Key Account customers to expand our business opportunities with mid-size customers. This approach allowed us to gradually reduce our dependence on key account customers while driving revenue growth and profit enhancement in our public cloud business. During the past half year, we carried out a systematic review of more than 200 companies from about 10 industries and find more than 30 new mid-size customers with high growth potential. Thanks to these strategies, the revenue contribution from our top three public cloud customers has been non-defined, while the revenue contribution from mid-size customers continues to increase steadily. As a result, we have improved our customer mix while maintaining stable revenue growth in the public cloud business.

We are determined to implement high-quality, sustainable strategic ideas to clarify and standardize industry cloud project management requirements, and fundamentally improve industry cloud business quality. Specifically, we have four points to point out. First, focus on the accumulation of core free product components and solution capabilities. Second, focus on free output as the core, to ensure that there is a higher ratio of free product solutions in the process. Third, to select industry and customers with a higher value of life cycle and grow together with customers. Fourth, to improve project delivery management, improve customer experience, and reduce delivery costs. Moving on to enterprise cloud services, revenues increased by 26.4% quarter-over-quarter to RMB 790 million in Q4.

We remained steadfast in executing our high-quality and sustainable growth strategy and further clarified and institutionalized our project management best practices to fundamentally enhance our enterprise cloud business quality. This effort centered on four key initiatives. First, we focused on accumulating and enriching our core offering of proprietary products and solution capabilities. Second, We continued to enhance the revenue share of our proprietary products and solutions across our projects. Third, we targeted industries and customers with high potential lifetime value catering to their evolving needs and grow with them. Lastly, we further enhanced our project execution to improve customer experience and reduce costs. These initiatives not only generated a relatively high estimated margin for enterprise cloud projects in the current financial period, but will also drive sustainable margin expansion in the long run.

to complete its urban smart transformation and development projects, and to deliver the core free energy products of the company, Yinghe Yun. We are also delivering projects such as Beijing Water Resources Bureau Zhenwu Yun and Shibi Zhenwu Yun, to continue to expand our advantages and business scale in the field of Zhenwu Yun.

Taking the public services sector, for example, we focus on the public services cloud model and developed a benchmark system consisting of cloud products, services, and operations. In Q4, we completed a smart city upgrade project for the public services and big data management bureau of Cherokee municipality in the province, in which we deployed our core proprietary enterprise cloud solution, Galaxy Cloud, We are also carrying out a number of other projects, including the Beijing Water Authority Public Services Cloud and the Qibi Public Cloud, further sharpening our competitive advantages and business scale in public services.

In this quarter, we completed the second phase of expansion projects for Sichuan Tianfu Health Cloud and Chongqing Weizhenwei Shadow Cloud. This is our service to continue to provide value to industry cloud customers through leading product solution solutions and excellent services. In the financial sector, we continue to deepen cooperation with state-owned banks and large shareholder banks, and focus on the operation capability of financial big data support services. In addition, while CREIT is stabilizing and storing customers, we will deepen cooperation with companies in customer resources and payment,

In the healthcare sector, we continue to enhance our five major models, namely the Regional Healthcare Cloud Model, Medical Image Cloud Model, Integrated Healthcare Organization Model, Regional Integrated Model, and Smart Hospital Model. During the quarter, we completed the second phase capacity expansion project for the healthcare cloud in the Sichuan Tianfu new area and the medical image cloud of the Chongqing Health Commission. This showcased our ability to provide continuous ongoing support for enterprise cloud customers using our market-leading products, solutions, and services. In the financial services sector, we further deepen cooperation with state-owned banks and major commercial banks by focusing on providing financial big data support and operational capabilities. In addition, while retaining a stable existing customer base, we strengthen our project deployment partnership with Camelot, particularly enhancing our synergies and cross-sellings in the banking sector.

Technology is the most core development strategy of J&J. In the third quarter of 2022, we developed the development strategy of the Beijing-Wuhan Double Research and Development Center and accelerated implementation in the fourth quarter. We aim to effectively balance the cost of R&D while improving sustainable R&D capabilities. Since December 2022, In terms of technology,

we continued to advance our core strategy of building success based on technology. In third quarter 2022, we developed the Beijing-Wuhan dual research center strategy and we executed well in the fourth quarter. We aim to sustainably enhance our R&D capabilities while maintaining a disciplined R&D budget. We doubled the number of R&D staff in the Wuhan Research Center within just three months of its launch in December 2022, and we expect to grow the headcount there to more than 1,000 over the next three years, injecting momentum into our R&D initiatives and helping cement our industry leadership.

We focus on some core product technologies, such as to create a consistent user experience. For example, we have released the new version of the 7th generation cloud host, container service, and cloud function. The performance has been significantly improved. The core products of the industry cloud base, the new version of the Galaxy Cloud, has added 79 ISE, PaaS, and operating core functions. The product power has been greatly improved. The data collection and collection capabilities of large data cloud platforms and large data engine core products, as well as the combination of architecture deployment capabilities, have all been improved. We have also actively improved the support for different operating systems, databases, and chips. The compatibility has been significantly improved.

We strive to deliver the ultimate user experience across our core products and technology categories, including cloud hosts, cloud native, enterprise cloud infrastructure, enterprise storage, and big data cloud platforms. And we benchmark ourselves against the top tier players in the cloud industry. For example, we recently launched our seventh generation cloud host, as well as new versions of our container services and serverless cloud functions, delivering significant performance improvements. We added 79 key operating features, including various IaaS, PaaS solutions, and operation management functions to the upgraded version of Galaxy Cloud, significantly enhancing the competitiveness of our flagship enterprise cloud products. We also upgraded the data collection, data integration, and hybrid architecture deployment capabilities of our big data cloud platform and engine solutions. In addition, we significantly enhanced our product compatibility with various operating systems, databases, and chips.

In addition, since the release of GPT-3.5, We maintain a high level of attention and actively expand our business opportunities. First of all, PPT requires a huge amount of computing and data. Therefore, we use cloud to calculate natural high-end coincidences. Whether it is model training or scene application, we all need cloud to calculate negative energy. The third is cloud. We cooperate with many leading AI companies. There are solutions verified by the market. can be quickly deployed when there is a need. Second, due to the fact that Internet cloud manufacturers usually have their own GPT-AI development business, Jinshan Cloud, as a medium-sized cloud service provider, has the natural advantage of serving the widest number of customer groups and has begun to gain recognition in the market. Thirdly, GPT's ground application, especially the traditional industry that is relatively weak in ground application and IT capabilities, needs to carry out systematic implementation and planning according to the actual situation of each enterprise, and to develop, install, deploy, adjust, and implement a series of tasks. Tiananmen's strong IT comprehensive service and delivery capabilities will be able to effectively capture the huge business opportunities in this field.

Since the debut of GPT-3.5, we have been closely following its development and actively exploring relevant business opportunities. First, GPT models require massive computing power and vast troves of data, making cloud computing a natural fit for this technology and an essential enabler for use cases, including both training AI models and applying them to various scenarios. Through years of collaboration with leading AI companies, we have developed a market-tested solution that can be rapidly deployed on demand. Second, major internet cloud service providers generally are developing their own GPT businesses, whereas we remain a neutral player. This means we can serve a wider range of customers, a natural advantage that the market is beginning to recognize. Third, the application of GPT models, especially in traditional industries with relatively underdeveloped IT capabilities, will require extensive preparation work unique to each company, including consulting and planning, process re-engineering, customized development, installation and deployment, and ongoing maintenance. Our strong and wide-ranging IT support and deployment capabilities will enable us to capitalize on such huge opportunities in this market.

Looking forward to the future, facing new opportunities and challenges, no matter how the environment changes, Jinshan Yun will always insist on following this high-quality, sustainable path, self-sustaining, high-end, solid ground, and continue to create value for customers, shareholders, employees, and society. Next, let's welcome our CFO, Henry, to introduce the performance of the fourth quarter and the whole year. Thank you.

Overall, looking back at the challenges we faced in 2022, we are gratified that our proactive strategic adjustments enabled us to achieve positive initial results and strengthening our conviction that we are on the right track. Looking ahead, in the face of new challenges and opportunities, we will pursue high-quality and sustainable development, no matter how the environment changes, and roll up our sleeves to create sustainable value for our customers, shareholders, employees, and the society. I will now pass the call over to our CFO Henry to go over our financials for the quarter and the full year 2022. Thank you.

Thank you, Dongdong, and I welcome everyone for joining the call. Before diving into the financial details, I will walk you through a quick summary for the fourth quarter of 2022. First of all, with our strong commitment to improve profitability, we have taken comprehensive measures from all perspectives. including proactive adjustments to our CDS services, strategic restructuring of custom mix, prudent enterprise cloud selection, and strict control for fixed assets and operational expenses. Since the third quarter of 2022, our adjusted growth margin has been increased for five consecutive quarters. Increasing from 1.2% in the fourth quarter in 2022 to 3.6% in the second quarter last year. 6.3% third quarter and further to 7.9% this quarter. Adjusted growth profit increased by 408% year-over-year to 168.5 million RMB this quarter. Workgear's poll certainly underutilized the service based on the current customer demands and record loss on disposal of properties and equipment. However, we believe this is helpful for long-term developments and growth margin expansion. Non-GAAP EBITDA margin profit was negative 245.1 million RMB, impacted by non-recurring Hong Kong IPO listing expenses of 94.9 million RMB. and a loss of disposal of properties and equipment of 28.8 million RMB. Non-GAAP EBITDA margin was negative 11.5%. However, if excluding the IPO expenses and the loss of disposal of properties and equipment expenses, our non-GAAP EBITDA margin could have been negative 5.7%. compared with negative 10.3% last quarter and negative 4.7% in the same period of 2021. Second, our operating cash flow has been positive for the past three quarters consecutively, and we have achieved 370.4 million RMB net operating cash inflow this quarter. Thanks to our prudent control over capital expenditures, Free cash flow, as measured by net cash generated from operating activities minus capital expenditures, was 259.6 million RMB, marking the first quarter of the positive free cash flow, demonstrating our strong commitment and a successful execution of our cash management. Third, our cash and cash equivalents and short-term investments was 4.7 billion RMB by December end 2022. Considering the improvement of our profitability, our scaling down of capital expenditures and cash inflow of operating cash flow, our cash reserve is well positioned and sufficient to support us walking through the challenging year and provide flexibility to further business development. Lastly, our total revenue was 2,131 million RMB this quarter. Revenue from public cloud services was 1.34 billion RMB, remained stable compared with last quarter. Revenue from animals cloud was 785.9 million RMB, increased by 26.4% quarter over quarter. With a more balanced and healthy business mix, we believe we are well positioned to start a new journey for our sustainable long-term development, being able to allocate more resources to expand our mid- to long-tail customer bases and high-quality non-internet companies. Now I will go through our financial in detail. Our total cost of revenue decreased by 25.2% year-over-year to 1,969.1 million RMB. IDC costs decreased significantly by 20% a year-over-year from 1,321.9 million RMB to 1,057.6 million RMB this quarter. Depreciation and amortization costs increased by 6.4% from 227.2 million RMB in the same period last year to 241.7 million RMB this quarter. Solution development and services cost decreased from 497.2 million RMB to 465.8 million RMB this quarter. The decrease was mainly due to the synergies from overlapping headcount reduction within Camelot and InfoCloud on the COVID-19 impact and other synergy initiatives on the demand side in last December. Fulfillment costs and other costs were 155.6 million RMB. 48.3 million RMB this quarter. The adjusted growth profit of this quarter increased by 408% to 168.5 million RMB, representing adjusted growth margin of 7.9%. The significant growth margin improvement was mainly due to the impact of cost control measures and the strategic adjustments of our revenue mix. In terms of expenses, Excluding share-based compensation, our total adjusted operational expenses was $729.7 million RMB. However, still impacted by Hong Kong IP electricity expenses of $94.4 million RMB. The disposal of fixed assets of $28.8 million RMB. Within that, adjusted R&D expenses was $239.4 million RMB. remained relatively stable compared with 231.6 million RMB from last quarter. Adjusted selling and marketing expenses was 118.4 million RMB compared with 125.5 million RMB last quarter. Excluding the listing expenses and disposal of fixed assets of 123.7 million RMB, adjusted G&A expenses increased slightly from 219.9 million RMB last quarter to 248.2 million RMB. We have taken various measures to cut down expenses, including but not limited to the following aspects. First of all, we review weekly the variable operational expenses, especially in marketing and administrative expenses. Second, we streamline headcount management within the firm and review our cost strategy and adjusting employee fees and other structural personnel. Third, along with the scaling down of certain customer CDN services, we accordingly improve the efficiency of online resources. We also dispose certain fixed assets and scaling down benefits costs as well. Net loss margin was negative 24.5% this quarter, and adjusting net loss margin was negative 25.9%. The adjustment was mainly due to the foreign exchange gain of 132.3 million RMB caused by a fluctuation of U.S. dollar RMB exchange rates, which is a non-cash impact. As of December end 2022, our cash and cash equivalents and the short-term investments was 4.7 billion RMB, providing us sufficient liquidity for operations. During the fourth quarter, We have repaid certain loans within the group and the banks to reduce our interest cost. The capital expenditure for this quarter was 110.8 million RMB, which primarily consists of payments for servers. In terms of share repurchase, regarding our 100 million US dollar share repurchase program within a 12-month period as approved by the Board of Directors announced in March 2022, Since the release of our second quarter results up to the year end of 2022, we bought a total of 12.3 million ADR shares for the cost of roughly 29.2 million US dollars. Going forward, we still have authorization from the Board of Directors and the flexibility to continue execution from time to time as weight to the mandated repurchase program. These efforts fully demonstrate our board and management's strong commitments and confidence in the long-term business development of the company. Finally, we have successfully finished the due primary listing on the main board of Hong Kong Stock Exchange by the way of introduction on December 30, 2022. In March 2023, we have been selected and included into Hong Kong Composite Index, Shanghai Hong Kong Stock Connect, and Shenzhen Hong Kong Stock Connect at the same time. Primarily in Hong Kong has helped us broaden our investor base and open up new investment channels. Our teams have been communicating more frequently and broadly with domestic investors, and we have seen more active trading patterns since we joined Hong Kong Stock Connect. Looking ahead, although we are still implementing our strategic initiatives, including business repositioning and cost control efforts on an ongoing basis, such adjustments have already yielded positive preliminary results as reflected in a clear improvement of profit margin in Q3 and Q4. We expect our total revenue to be between 1.85 billion RMB and 2.05 billion RMB for the fourth quarter of 2023. While the forecast and the comments above are based on our current and preliminary views on the market and operational conditions, which are subject to change, we firmly believe that given time, our potential positive impact of our ongoing strategic initiatives will continue to amplify and reflect our financial in the mid to long term. Thank you.

Thank you. So to ask a question, you will need to press star 1 and 1 on your telephone and wait for your name to be announced. And to withdraw your question, you can press star 1 and 1 again. Please stand by while we compile the Q&A roster.

Thank you. We'll now take our first question.

Please stand by. This is from the line of Brian Gong from Citi. Please go ahead.

Thank you, Director-General, Director-General, Clark, Nicole.

I have two questions. The first question is about industry language. How do we look at the needs of three verticals this year? Last year, the industry might have been affected by some restrictions. Will this make the increase faster this year? How do we balance the income increase and margin? These two relationships. The second question is about CHAP GPT. Director Zhou mentioned that we, as independent operators, have an advantage in China. But I would like to ask, if China has the ability to develop a large-scale model of the Internet, it basically has its own cloud services. If companies use their models, will they be more inclined to use their cloud services? Thank you for these two questions. I will translate myself. Thanks, Max, for taking my question. I have two questions. First is regarding the enterprise cloud, regarding the three verticals. How should we see the demand this year? Last year, the delivery was impacted by, you know, COVID. So this year, should we see faster growth? And how should we balance the growth, revenue growth and the margin? The second question is about China's GBT. Zou Zong just mentioned, as an independent cloud provider, we have some advantages in China. But, you know, in China, you know, for those internet giants who have capability to develop large language models, they all have their own cloud, you know, service. So if the enterprise users, they use their large language model, will those users more, you know, have more intention to use those internet giants cloud services? Thank you. Let me answer. There are still some places that need to be added. The first question is about the South China Sea. With the spread of the epidemic, the whole country is also recovering from economic development. We are aware of the impact of the epidemic last year. Thank you. China China China China China China China China China China China China This is the first point. The second point is... I understand what you mean. How can we balance profit and market? I mentioned it repeatedly in my speech. We have a high quality of sustainable development. It doesn't mean that we don't pursue income growth. But we put the focus on every client. We will evaluate whether this project is a well-rounded and open project. Is this project a profitable project? If we don't do it, we definitely won't do it. Especially if we extend a life cycle. If it's a long-term, it's a non-profit project. Then we think it doesn't belong to a high-end client. This is called sustainable development. What does it mean? The country is also mentioning it. You can't say that it will be gone this year and it will be gone next year. To be honest, this kind of business has no strategic meaning for us. Maybe it will be of some help in the current financial recovery. But in fact, all companies are clear. No matter how much you say this year, this piece will be gone next year. This actually has no strategic meaning for us. So, you may have misunderstood that we are going to increase our income. Of course, you also see that our income level seems to have slowed down compared to the past. But in fact, it is all about high-quality and sustainable adjustment. So, you can also see that while the income is steadily rising, our profit capacity is greatly improved. In fact, it is because of some of our past ups and downs. low-quality, non-continuous-development projects. And then it leads to the replacement of high-quality, sustainable development. So don't misunderstand the market. If we misunderstand the market, we won't pursue this growth. No, we shouldn't. We should put the past, this single pursuit of growth, into a high-quality, sustainable growth. Because we think this is more from the long-term point of view. It has more strategic value.

So, as you rightly pointed out, so we're noticing that right now we are going through the opening up phase after the COVID period, and the country is also reverting back to the model of business development. And admittedly, last year, the COVID situation and relevant restrictions did have impact to our deployment and delivery of our enterprise cloud business. And some of the planned deployment that was originally planned to be completed in Q4 last year were actually delayed to this year. But if you look at the situation now, although we haven't disclosed the particulars and the concrete numbers of the specific three verticals that you mentioned, we do remain highly confident about our operating metrics in enterprise cloud, including revenue, including gross profit, and including operating margin. And we expect to have significant improvement in those metrics. And secondly, I would like to clarify our pursuit of high quality and sustainable growth does not necessarily mean that we do not pursue growth. For every customer and for every project, what we do is to evaluate whether that customer and that project is centered around the core cloud business, and whether that brings about profitability to us. In other words, if some of the customers and projects, even if we evaluate them from a longer time period, and it will not bring about profitability to the company, then this is actually not a high-quality project and not high-quality development. So the point I would like to mention and I would like to highlight is that in the past, we have been overly emphasizing the growth. And now what we need to do is to replace that overly emphasis on growth to high quality and sustainable growth, which is of strategic value to us.

Okay. The second question is a very good question. I will also reply to it later. Let's see if Mr. Cao can answer it. As you said, we have also announced to the outside world that whether it is the cloud or the modern office, we ourselves will definitely not do this large-scale model. This does not mean that we cannot invest or that the threshold is very high. In fact, we think that um um I mean, But we have also studied and analyzed it seriously. The problem you just asked is that if these big companies have done it, and everyone uses their language model, who will use it? This is how we think about it, in several levels. The first one is that the model of this language is not necessarily eight layers. Of course, the big companies will have a higher success rate. At least, based on the capital investment of Tengda, it has an advantage. But it doesn't necessarily mean that big companies can do it. TechGP, OpenAI, is also a startup company. It's not from Google or big companies. Of course, we just talked about big companies. It has its advantages. So, we also see that there are a number of companies in the market that have been doing this for the past few years. Actually, most of them will choose us as the cloud service provider provided by them. The reason is also very simple. It's your big companies. You did it. Why should I go find you and listen to you? The business secrets are completely gone. So this is why I talked about neutrality in the conversation. This has already been proven by the market. This is one. The second, our point of view is that we actually repeat the discussion, including with a lot of This this investment is also good. This is good for doing technology. Our point of view is that China should really do it in the short term. With the United States or to reach the United States level of this 4.0 is even higher. I don't think I can do it in the short term. But but We also saw that actually that uh uh uh I think everyone can see it. This is also a great inspiration for us. So I also talked about it in the conversation. In fact, we think it is more likely that with this technology direction, this new generation of AI technology direction, this small parameter industry application model may have more real value in China. 对,就虽然我们短时间之内很难达到美国的这个最高端的这个AI能力,但是我们完全可以把现有已经具备的这个,现有的AI能力赋能到这个海灯去,这个是非常有现实意义的。 In fact, in some areas, I won't talk about it. In fact, it has already been very effective. So we believe that it's not only about the large-scale model. It has value and future. It's not. So we believe, especially in this wave of startup companies, the big probability in the future will gradually narrow down to do a lot of specific industries in this small and medium-sized model. Or you can understand that it is in the industry model. Yes, so these will also be our potential customer sources, or even become a possibility for us to have a right direction in the future. So I just talked about this kind of new AI ability. This is how to give companies a negative energy. This is actually a new technology era. This is a new demand. Yes, so what I just said is that maybe especially for some non-IT traditional industries, they may know that GPD and new AI technology can really be implemented and can be used, but they won't. They don't have this ability. This is actually another opportunity for us. So to put it simply, as I said before, these small and medium-sized startup companies that do AI will choose us. In addition, with this new AI capability, there will be a lot of small and medium-sized models in various industries. I think these are all our potential customers. Okay? Okay, so I'll take a brief hint from CEO.

So we have announced, in the sense of Kingsoft and its affiliated company, that we will not be sending a lot of energy and resources or working on the big language model ourselves. Because we have very clear value proposition, which is we're providing solution to our 2B customers and therefore working on such a model is not a significant strategic value to us. However, the platform companies, the platform internet companies, they have the core capability and they have the relevant numbers and therefore are more suitable to develop such big models. However, in relation to your question of who would use our service and products, I have two main things. One is that although the larger technology companies, the technology giants, have their relative advantages, as we mentioned just now, but it does not necessarily mean that those models need to be developed, research and developed by such a giant. OpenAI poses a good example, which is not a technology giant, however, generated the best in class such a model. And then the second is that also as we discussed in the prepared remarks, because of our neutrality, all these venture teams that are in smaller companies, because of the potential conflict of interest with major internet giants, we also choose the service of us rather than those internet-based those internet giants based across service providers. And then the second proposition I would like to say is that although in the short period of time, we do think that it is unlikely that the big language model, the GPT in China, will develop to a level similar to that of the GPT 4.0 in the United States. However, it does not necessarily mean that smaller models do not present real values and applications in the wide industries in China. For example, we're already seeing a lot of small models with parameter number amounting to 6 to 10 billion having very vivid and concrete applications in various industries. And for those application scenarios, are basically also our potential business opportunities. So to sum up, two of our potential business opportunities the current wave of TPT process for us. One is the venture teams from smaller companies will work on those models. And then secondly is the application of such models and small models, which we might also call industry models, their application into the traditional companies with relatively underdeveloped IT capabilities, which will help them to apply such models in their day-to-day operations.

Brian, now, Hello, Brian. For the first question I'm asking about Enterprise Cloud, probably some information. I think, first of all, if you look at our Q&Q growth, if you remember last quarter, we created about 600 million RMB revenue for Enterprise Cloud from the three verticals in total. But I probably encourage you to look at it in two perspectives, because in Q4, we deliver around $784 million in revenue, which is actually on net-to-net and Q-on-Q basis is about $180 million R&D increase. And if you remember, in December last year in Beijing, everyone probably remembers, people are stationed at home, and the city was basically affected by the COVID pandemic in December last year. So if you put that into a contact, you will see that given that we have a solid relationship with our customers, and even we have a lot of constraints from an operational perspective, but we're still increasing on a quarterly basis, that actually proven our capability in a difficult time of execute, deliver, and the booking on the revenue on that enterprise cloud, which actually demonstrate our technology and the client relationship. The second point I also want to mention is while on this quarter we didn't provide a color on the backlog and I appreciate you didn't ask them that question but we're hoping that going forward we'll disclose more information especially the backlog and the new signing on the contracts on the enterprise cloud revenue but I'm happy to provide some colors if we see the backlog at this moment at this time the backlog we have today will fully cover our potential budget for the enterprise cloud in the year of 2023 for this year. And we're still in Q1, so we're hoping we're increasing the backlog this year, and we do have the confidence that we're going to move into the more balancing of the growth and the quality model, but the potential and the capability of the growth, we do have that confidence at hand. And the third point, we also didn't mention that in the prepared remarks, is given we focus on the vertical and the customers, the percentage of the repeating customers in Q4 and Q3 last year has been increasing quite a lot, which means that even though you see about $700 million revenue on a quarterly basis, you know the percentage of that number coming from the same customer but other different phases of project has been increasing quite a lot that also give us a base that for next year and this year next year we can have the potential incremental revenue build on a solid basis and hopefully for the next next time in a positive earning course we can help to provide some color on the backlog as well as the percentage on the regime customers going forward we have that plan for the budgeting and disclosure process going forward hopefully will be helpful official right

Thank you. We'll now take our next question. Please stand by.

This is from the line of Timothy Sowell from Goldman Sachs.

Please go ahead.

Thank you, Mr. Guan. This is my question. I have two questions for Mr. Guan. The first one is about the public sector. Mr. Guan, can you help us look at it? In 2023, we are going to look at the needs of major Internet companies and small and medium-sized companies. What kind of company needs will be the strongest? The second one is that we see that in 2022, the company has a very strong strategic position. The overall net profit and EBITDA are all very positive changes. Thank you, Benjamin, for taking my question. I have two questions. First, could Benjamin share your outlook on the public health demand in 2023? And between the big enterprise, big internal companies and the small SMEs, which kind of companies do you think have a bigger demand on public health for this year? And secondly, we saw that last year, our company has very good execution on profitability in both growth margin and EBITDA margin. Could you make sure you updated our look for the growth margin and EBITDA margin for the United States? Thank you.

The first one, let me answer it. If Liu Changyong adds more questions. I understand it myself. Because we... Right? uh, uh, uh, uh,

Just quickly translate for CEO. So in relation to a first question, I think that public cloud services and products are really just centered around some of the core components, including computing, storage, and network. So I do not necessarily think that demand has to do, differs, and have to do with the size of the enterprise that we serve. So everything really depends on the particular application scenario and the business situation of our customer enterprise. For example, we would have our video company customers with their core demand coming from our CDN and storage business. And we would also have customers, for example, like in the AI industry, that would really demand the computing at their core demand. So that is the general response I have for your first question.

Thank you. I'll take on the second one. So before I go in there about the 2023 target, I just want to lay out the three major reasons for 2022. First of all, for the PP&E, we book on the balance sheets. We did try our efforts to reduce redundancy and other misallocation of the the revenue and the resources. So that's actually the effort has been taken for our two or three quarters since the middle of 2022. And I don't think those are going to be the one time of impact because once you increase the utilization ratio for the assets and resources, those benefits on the growth margin will be gradually relieved over time. So that will be the first layer of the margin expansion for this year. Second of all is, given we, you know, as a team, we're together to change the combination of customer base. So right now, as our CEO mentioned, our mid-size client base has been increasing sequentially as a total value of total revenue. And I would say that the pricing and the profitability from those buckets of the customer are definitely much better than we're serving on a single client basis, right? So that's going to be two benefits, which layer for the course margin for 2022. And the third, obviously, is the variable expenses control, as I mentioned in the prepared remarks. The very basic things, the travel expenses, accommodation, how we pay people, and how we change the mix of the incentives, the cash plus the stocks, and a lot of things we're doing in the past year. And I would say that only part of the benefits have been reflected in the Q4 currently in Q4, and hopefully those change our policy of expenses, including the share-based compensation, including cash follows and so forth will be reflected in q1 q2 going forward in 2023 in the new year so as a result our operating expenses and operating margin will be better than last year the reason this growth margin is lifted and the variable cost is reduced so while i'm sorry i cannot give you a very clear numerical target of the ebitda margin and operating profits but i would say two things first of all we are hoping to increasing our growth margin on the sequential basis quarter on quarter, so hopefully we can see the expansion very stable, I would say stable relatively in the next two quarters, point number one. Point number two is given the controlled expenses and the variable cost, so hopefully we can be more visible and can be faster to reach the break-even and profitability of the EBITDA margin for this year. I think these are the two major objectives for us. But if you're asking about operational margin and a break even on that margin side, I think obviously we'll see the balance of the growth. And we also need to be very nimble to make investments on our capex and also on the growth side to make money and put our investment to serve our high-quality customers. So we do not prevent that opportunity. So it gives us a little bit of flexibility to balancing on the operational margin and the EBITDA margin. But hopefully we can be more visible to getting to the point that our EBITDA margin Thank you.

Thank you, Madeline.

Thank you. We'll now take our next question. Please stand by. And this is from Alan Lee from JP Morgan. Please go ahead.

Thank you for accepting my question. I have a question about capital spending. In the past few years, we have seen that our capital spending has actually maintained a certain strength and hit a little bit of the pre-money. Later, because of the change in the macro environment, it may affect our overall use rate, including the influence of profit rate. Now, in the face of AI, which is a relatively large industry opportunity, what is our idea on capital spending this year? How do we choose? Let me quickly translate my question. My question is regarding the CapEx. We pre-invest in CapEx in the past few years, but due to all kinds of macro headwinds, our overall utilization rate seems not very high, and it also negatively impacts our margins. So looking into this year, given the potentially huge opportunity generated by AI industries, So how should we think about our CapEx plan in 2023? And we will take a more proactive approach or reactive approach in terms of AI-related CapEx investing. Thank you.

As you said, we can simply put it this way, the new generation of AI technology is exploding. Whether it's a forward-looking or a strategic plan, we've already taken action. Compared to last year, at least in this field of development, from my point of view, it's basically a green light. But we're not blind to it. How do I put it? We're all on it. We still have an optimistic attitude. I think it's because of some historical patterns that we developed in the previous generation. I think it's too hot now. I think it will still be in a more normal stable state in the future. So our overall attitude is to be optimistic. Yes, we will definitely China China China uh, uh, uh, uh, uh, uh, So as you really pointed out with the new generation AI technology, mainly the GPT eruption or disruption, we have had actually already preemptive actions in terms of reserving both technology and resources.

However, I would summarize our attitude as cautiously optimistic rather than blindly pursuing this potential opportunity because based on our experience with the larger generation of AI industry trajectory, we do think that currently the market is tends to be or seems to be overheated, and we expect it to actually enter into a more sustainable but slower growth kind of mode. What we will continue to do in this, we will adopt a kind of approach that's run with small steps and with quick iterations. I also want to point out that in terms of the hardware, there's also the more Yes, which essentially means that the development and upgrade of such hardware will actually go into the past. So we do not think it makes a lot of sense. to massively awarding the last generation of such chips and servers.

Thank you. Thank you, Alan. I'll probably put some additional color on this question as well. So if you remember, on the year of IPO 2020, we spent about 1 billion RMB on the CapEx. So first of all, I'll mention the point. As a company, we do have capacity and experience to managing spending 1 billion RMB on whatever with purpose, and we do have a very experienced human team, and we do know the supply chain, and also we have a good partnership with our ecosystem partner, including our shareholders, so Kingsoft Group as well as Xiaomi, to managing the complexity of buying those big chunk of assets at a reasonable price, given the past experience. So that's the first point. The second point is, I think the question is coming from more about investments, not only about CapEx. If you look at investment, I'll put it into two buckets. The team, the technology R&D team, which our senior people as a team managing today, we have quite a lot of very experienced engineers and programmers. And those expenses is not a capex. And expenses are booked on our P&L income statement. And as you mentioned, we spend over $1 billion R&D every year on those expenses and R&D investment. So I think we do not hesitate to invest into R&D products. So I think that's very clear. However, if you are looking into the cash flow item on the capital expenditures, which I want to say, also going to have a different model because as our CEO Zozo mentioned, we see the great potential opportunities will come in from both internet clients as well as implementation and application of the user cases from a lot of diversified and non-internet clients, especially from traditional enterprises. And those business models will not consume Kingsoft Cloud's own capital. because they will build their own environment and providing their own computing capabilities. And we call it an OPEX model. So I think we're not going to wait to see those good business opportunities, but we will balance out. We need to spend money ourselves, putting our own IDC, or we build a computing environment and help our enterprise cloud clients to use the AI capability and using their money to create a revenue block. I think those are the two choices we both have on hand, and it's not clustering ourselves on one side. and the last point i would put away is i think the the question is not really about whether or not is about who is spent for so i think selecting the right client is also very important i think it's our ecosystem partner including our shareholders for example xiaomi and the kings of cloud against office to build for the office software xiaomi for the ed cars we will go 100% to make investment and spend capex. However, for the company and the customers, which do not fall into the criteria that our CEO mentioned, getting customers cloud-native technology will probably be set back a little bit. So I think this question about who we spend money for is also very important. And hopefully it's helpful for you. Thank you.

Yeah, sure. Thanks, Emmanuel, for the answer. Yeah, it's very helpful. Thank you.

Thank you. And at this point, I would now hand back to Nicole Shan for closing remarks.

Thank you, Aubrey. Thanks, everyone. This concludes our earnings call. Again, for joining us today, if you have any further questions, please feel free to contact us. Look forward to speaking with you again next quarter. Have a nice day. Bye-bye.

Thank you. This does conclude the conference for today. Thank you for participating, and you may now disconnect.