Thank you very much for waiting. We will now begin SoftBank Corporation Earnings Results Presentation for the Fiscal Year-Ended March 31, 2025. We would like to introduce today's attendees, President and CEO, Miyakawa; Representative Director and COO, Shimba; Board Director, Executive Vice President and CFO, Fujihara. Today's presentation will be broadcast via the Internet. President and CEO, Miyakawa will give an overview of SoftBank's consolidated financial results and business overview.

I'm Miyakawa of SoftBank. Thank you so much for your attendance today. I will now explain our financial results for the last fiscal year. First, revenue was JPY6,544.3 billion, up 8%. We achieved a goal of JPY6.5 trillion set in the mid-term Management Plan one year ahead of schedule. The breakdown by segment is as follows

all businesses reported an increase in revenues. Financial distribution and enterprise, all these businesses achieved double-digit growth. Operating income was JPY989.0 billion, up 13%. The target of JPY970 billion set in the mid-term plan was achieved one year ahead of schedule. Operating income also increased in all businesses. In addition to double-digit gains in media, commerce, and distribution, financial returned to profitability, recording over JPY30 billion for the full year. Net income was JPY526.1 billion, up 8%. Since both revenue and operating income were on upward trend, we had expected to achieve net income one year ahead of schedule, but it will, unfortunately, it will be carried over to FY2025 due to a valuation loss on the portfolio company at the end of the fiscal year. The main source of dividends, which is primary free cash flow, was JPY603.3 billion, which remained at a high level. The consolidated financial summary is as here both revenue and income increased. The dividend was set at JPY8.6 per share as committed at the beginning of the fiscal year. From here, I will talk about the upward revision of the financial targets of the mid-term plan and our approach to the financial forecast for FY2025. Since we already achieved our revenue target of JPY6.5 trillion last fiscal year, we have decided to revise our forecast upward by JPY200 billion. If the increase in revenues continues this year, it will mark the 10th consecutive fiscal year since the merger of the four telecommunications companies and will be a new record high. We are aiming to increase revenues in all segments this fiscal year as well. One difficult thing to forecast is the sales of electricity in the consumer business, which is affected by weather conditions, but we are determined to increase revenues in all business segments without any excuses. Next, let me explain about operating income. As you can see, when we announced our mid-term plan in May 2023, we committed to achieve the target of JPY970 billion in FY2025. We were able to overcome the impact of price reductions and achieve an upswing for the second consecutive fiscal year. As a result, we achieved the goals of the mid-term plan one year ahead of schedule while making growth investments in generative AI. In the current fiscal year, while continuing to invest aggressively in AI, we intend to revise our financial targets upward to achieve JPY1 trillion. The breakdown of the JPY1 trillion is as shown here, aiming for continuous profit growth in all segments. We are aiming for more than JPY1 trillion while strengthening upfront investments for growth in the next mid-term plan from next year onward. In the Consumer business, which is the backbone of our business, we aim to increase profit by approximately JPY20 billion despite the severe competitive environment. In Enterprise business, we aim for double-digit growth. Media, EC, and Financial are combined for the purpose of preparing for PayPay's IPO, but we are aiming for a 10% increase in both businesses altogether. So we are now positioned to achieve an operating income of JPY1.1 trillion from these businesses combined. With this, we are finally in position to generate R&D expenditures at the JPY100 billion level per year, which was my goal when I became President. We would like to conduct R&D with an eye toward commercialization, not just basic research, and we would like to focus on the development of our own services. This includes investments in AI as in the past, which we hope will lead to growth in the next mid-term management plan. Also, this is not all of it to be used. So if it remains, it will be an addition to what is stated. We have also revised upward our net income target to JPY540 billion. We will continue to achieve record high profits. We expect to pay a dividend of JPY8.6 per share in FY2025. In this inflationary environment, isn't a flat dividend practically the same as a dividend cut? I'm asking this to myself, but there are many opportunities to invest in rapidly growing AI, and we believe that aggressive investments now will lead to higher corporate value in the future. So we are forecasting the same amount for the current fiscal year at JPY8.6. After making this kind of upfront investment, we would like to consider share buybacks and dividend increases. As for the roadmap for the realization of our long-term vision, we achieved our revenues and operating income targets one year ahead of schedule, and we are making better progress than initially expected. Towards the third phase, we will manage the company from a mid to long-term perspective. From here onwards, I will give an update by business segment. First, let's look at the Consumer business. In the mid-term plan, we had originally announced that mobile sales would increase from FY2024, but we achieved the plan one year ahead of schedule and we aim to increase revenues for three consecutive years starting in FY2023. As for operating income, we would increase in FY2023 as planned and aim to increase it for three consecutive years. In addition, our target for the number of smartphone subscribers was to maintain an annual net increase of more than 1 million. And we achieved a net increase of 1.04 million in FY2024. We will strive to continue to expand our customer base. We've been making efforts in terms of both quantity and quality and the balance of conversion to the SoftBank brand was positive in both the first and second half of the last fiscal year. In addition to expanding our customer base, we are also working to improve ARPU, aiming to increase mobile sales and profits. As it is the end of the fiscal year, we would like to summarize our efforts in mobile networking -- network. The percentage of 5G compatible handsets has increased to 76%. As you can see, 5G connectivity is now better than at any other companies. Mobile traffic, which has been doubling every year, increased by about 1.8x last year. In spite of this, to ensure that quality does not deteriorate, we have continued to maintain throughput at the same level as last year by adding more base stations and expanding the bandwidth. We are also actively working on AI, but as a telecommunications company, we are focusing our efforts on telecommunication domain as well as our first and foremost priority. In addition, we are rapidly expanding our 5G stand-alone area. The area in green is last year's expanded area. After expanding nationwide, we will introduce new services that take advantage of stand-alone slicing and low latency. Next, I will provide an update on our Enterprise business. First, with regards to revenue, at the time of our mid-term plan announcement, we stated our goal of achieving double-digit growth in solution sales. As of fiscal year 2024, we have achieved 27% year-on-year increase. Looking ahead to FY2025, we expect it to reach JPY970 billion in total enterprise-related revenue, and we are beginning to see a clear path towards JPY1 trillion in revenue from the Enterprise business alone. As for operating profit, we also stated our goal of achieving double-digit growth annually. Over the past two years, we have averaged 12% growth, and we are projecting double-digit growth again this fiscal year. Our key focus area for this year is to strengthen our foundation as an AX integrator or AI transformation integrator. This means supporting companies in building their own AI by organizing their proprietary data and using it to train AI models tailored to their specific needs. As part of this initiative, in January, we launched Cross Boost, a proprietary SaaS product that leverages generative AI to streamline customer support operations. As you can see, Cross Boost is a corporate service that utilizes generative AI to respond to customer inquiries by referencing manuals, customer databases, and other resources to generate appropriate answers. Cross Boost act as an entry point into AI services. As companies use it daily, it accumulates customer support-related data and enhances its response accuracy. This, in turn, enables the development of AI models specifically tuned for each company, leading to automation of call centers. Through such phased implementation, we aim to support companies in adopting AI. To summarize our enterprise business growth strategy, we have grown through digital transformation so far. Moving forward, we will accelerate growth through AI transformation as well. Next, let me provide an update on our Media and E-commerce business and Finance business. We are presenting these together today because PayPay has initiated preparations for its IPO. To avoid any risk of prelisting solicitation issues, we will consolidate disclosures until the IPO. Afterward, if IPO is successful, we plan to resume disclosing each business separately for greater clarity. For FY2025, we will disclose combined forecast for Media & E-commerce and Finance. As shown here, we aim to achieve double-digit profit growth for the combined Media & EC and Finance segment. PayPay continues to perform well with gross merchandise value or GMV, rising 23% to JPY15.4 trillion. Despite its scale, it continues to grow at over 20%. Its consolidated EBITDA reached JPY45.6 billion, marking its second consecutive year of profitability. PayPay also achieved a full year operating profit of over JPY30 billion. As we mentioned in third quarter, we completed the restructuring of our financial businesses in April. PayPay Card, which becomes a subsidiary in 2022, has now been fully integrated into PayPay and is experiencing significant growth. Under PayPay's leadership, we will accelerate the growth of banking and securities service offerings. Additionally, GMV for SB Payment Service is progressing steadily with an aim of 15% growth annually. We will continue to target growth, particularly in non-telecommunications segment. Next, I will update you on next-generation social infrastructure efforts. The key areas of focus will be AI data centers, homegrown LLM, sovereign cloud, next-generation memory technology and Cristal intelligence. First, as for the AI data center, on March 14, we announced the successful completion of a contract to acquire the land and buildings of Sharp Sakai plant for JPY100 billion. This data center will be used as an important operation base for AI agents. Next, regarding Sarashina, a homegrown AI, we recently completed a commercial model called Sarashina mini. So after final confirmation through in-house trial starting in June, we will target commercial use of Sarashina mini by this autumn. The Sarashina mini was built using a data distillation technique with 70 billion parameters and a large LLM with 460 billion parameters was built. So using this training data, we were able to create an optimized high-performance model called Sarashina mini. In the future, by combining multiple Sarashina mini models to create a large-scale model, and continuously training through a method called dense up cycling, we will build an even higher performance LLM in a short period of time. I want to show you some demonstrations today. Here is the comparison of the Sarashina mini's performance with the GPT-4o and 4o mini. GPT-4o is set to have 200 billion parameters and 4o mini has 8 billion parameters. So the Sarashina mini falls in between these two. So I would like to compare this with the 70 billion parameters of the Sarashina mini, how they answer to a single question. I prepared three questions. The first one, according to Japan's Road Traffic Act, are helmets required for side cars specified two-wheeled vehicles and three-wheeled vehicles? So please pay attention to the screen in front of you. As you can see, it's quite speedy and Sarashina mini is showing the answer. So I just put each answer in summary. Sarashina mini has a good understanding of the Japanese legal system and correctly responded by distinguishing between a tricycle classified as a car, which does not require a helmet and the motorcycle classified as a two-wheeled, which does require a helmet. It also correctly states that helmets are mandatory for passengers inside cars. On the other hand, GPT-4o gave incorrect answers in some cases. So I have two more questions to show you. The next one, let's say, I move to a new house. What are some things to keep in mind when introducing yourself to neighbors? So as you can see how each one provides your answers. So 4o mini has just completed answering. So Sarashina mini, who understands Japanese culture and customs very well, advises us to bring a sort of gift like snacks and towels and also even advises us to bring our own local trash bags and what to do when our neighbors are absent. GPT-4o also advises taking snacks and towels with you, but this does not cover the customary Japanese practices. No advice was given on what to do in the event of an absence of the neighbor either. The third one, here is the answer to the question. Could you tell me the background behind the implementation of the New Freelance Law? It's really fast. How it provides answers is very specific to Japanese language. So this Sarashina mini correctly answered everything from the official name of the New Freelance Law to when it would take effect? To how large the freelance population is, while GPT-4 only provides general background information. So as you have just seen Sarashina will be updated daily and differentiated by a model that understands regional characteristics and cultures. We help to create industry-specific models that can be used to advance all industries. And eventually, it will become an AI that can be used in the physical world, like robot, automate, mobile and all this in the physical world. Now let me touch on our cloud business. Our vision for the cloud business emphasizes sovereignty in both cloud and AI, meaning control over data, technology, and operations. Within our AI data centers, we are installing GPUs, CPUs, and storage and developing the necessary software applications and SAR Solutions in-house. Our platform will also feature, Sarashina, a homegrown generative AI model as part of our sovereign AI strategy. This architecture will become the foundation for our next major growth strategy. Next, let me talk about next-generation memory development, which I have not talked about publicly in the past. As AI shifts toward inference rather than training, our AI data centers need to maximize GPU performance and handle larger volume of data. This increases the importance of high-performance memory capable of transferring massive data to GPUs. However, current high-bandwidth memory or HBM has challenges. It consumes significant power and has processing limitations. Together with our partners, we aim to develop a core memory sample within two years and assesses its feasibility. Our budget will approximately JPY3 billion for this initiative. Details on our partners and technologies will be announced in due course. This new memory will differ structurally from 3D stacked HBM and expected to enable ultrafast processing. We are committed to developing it with high hopes that it will support next-generation infrastructure. Now let me update on our Cristal intelligence. We are continuing discussions with OpenAI, and our aim is not merely to improve operational efficiency but to create transformative products that can reshape corporate management. Let me elaborate on that. Current AI adoption focuses on optimizing individual tasks without impacting company-wide operations, in my opinion. However, Cristal intelligence seeks to interconnect AI systems across departments to optimize and transform entire corporate management. We will begin by implementing it internally at SoftBank as a first step toward defining what a next-generation company should look like. To drive this strategy, we have established two new divisions directly under my leadership

the next-generation strategy division and the next-generation technology development division. These teams will accelerate our efforts toward long-term growth. We envision a future of coexistence with AI and aim to build the necessary infrastructure for that society, develop the essential software and connect everything through AI agents, positioning ourself as an indispensable company for the future. Now shift the gear to a different subject. First, let me provide a brief update on our individual shareholders. In October last year, we conducted a 10

1 stock split. Since then, the number of individual shareholders has increased significantly, rising by 500,000 from 860,000 to 1.36 million as of the end of March. Notably, the number of shareholders is 40 or younger, those who think about business on a 10 to 20-year time horizon has increased by 2.4 times and the proportion rose from 27% to 41%. This is a clear indication that we are beginning to see the results we aimed for. We believe that cultivating this kind of shareholder base will enhance long-term corporate value. Last but not the least, an update on our ESG initiatives. First, governance. We are nominating two new outside directors. If approved at the shareholders' meeting, our female Board representation will rise from 27% to 36%. We will continue to maintain a structure where more than half of our directors are external in order to fulfill our responsibilities to society, including protection of minority shareholders' interest. On renewable energy, our mid-term goal is to increase our effective renewable energy usage ratio to 50% by FY2025 and 100% by FY2030. For FY2024, we are projecting 54%, meaning we will reach our FY2025 goal one year ahead of schedule. Accordingly, we have revised our FY2025 target upwards to 60%, and we'll continue working toward the 2030 goal. To summarize today's presentation, all business segments posted revenue and profit growth in FY2024 achieving the target of our mid-term plan one year ahead of schedule. We have revised our financial target upwards to over JPY1 trillion, while keep investing for the future to support next mid-term business plan. PayPay has commenced IPO preparation. Lastly, we shared an update on our progress indulging next-generation infrastructure. We remain firmly committed to our vision of information revolution, happiness for everyone. Thank you very much.

Now we'll open it up to a Q&A session. [Operator Instructions].

Yomiuri Newspapers [indiscernible] I have two questions. One, regarding the next-generation memory development, it has -- I think it's a big dream. So who's going to actually make and how it's going to be sold and who's going to hold the inventory and so on? So among the semiconductors, I'm sure that you will be sharing some roles with your partners, but how are you going to allocate the roles of handling that? How much do you think you can make this into business? If you could share that as much as possible, it would be appreciated. The second one is about other companies. The NTT announced that to make NTT data to a wholly-owned subsidiary. So how this reorganization of NTT would impact the entire industry, you were concerned about that from before. So what is your opinion about this announcement of subsidiaries?

First, regarding the next-generation memory, it must have been a surprise to you. So now memory in Japan, there is a great patent held by a University in Japan. And if we were to realize this, I believe that this memory structure is as competent as what is available now as GPU. So we have been -- I have been discussing with notable experts that you may be aware of the names of, but if we were to establish the memory factory, then it would cost a lot. So as for SoftBank, I had thought that it's not an area that we should be investing. But as we develop AI speaking of other companies like deep research kind of structure, what we call Sarashina search and we are trying to providing new and -- answers. So that requires speed and so HBM needs to evolve more and more. Otherwise, the research functionality would not be useful in the future. So therefore, we started thinking about trying the memory area. So we have been actually working on drawing the design. But -- so in two years, the core samples should be completed and among some billion yen and we would just invest JPY3 billion or so. As SoftBank whether this is going to be our one of the core businesses, but for now, I'm thinking about just making it to like on IP side. So as far as we could use that, it would be great. So that's my expectation. So therefore, it is not that we are going to step into this memory industry. Number two, about NTT data, I had expected that this question would be raised. But once I start talking about it, I would get -- too get excited. So since we are the first right after this announcement, so let me give you a comment briefly. Looking back, the past DOCOMO NTT data, in order to remain the fair competition they separated, but now they would like to become one. I wonder there must be someone who want to make a giant NTT once again. So -- and also, when it comes -- if NTT wants to make overseas expansion of their business and if so, this -- that is understandable. But making NTT data a wholly-owned subsidiary, if that is the best thing to do or not, if when I look at the NTT, I don't think that would be the best way to do so. So because I know of President, Sasaki, very well. And if this would happen, then the great part of President, Sasaki, would be just halved. So I'm a bit concerned. So if for transmission lines and those base stations are used only for NTT Group, then it's not going to be a fair competition so therefore, I believe that NTT law did not -- NTT Act was not repealed. So that's my opinion.

Next question from floor, please.

Yamashita from Fuji Television. Thank you, Miyakawa-san for your presentation. Let me ask you about the trend in mobile as a sector. NTT DOCOMO and KDDI au announced price hike the other day. What's your view? What are you going to address going forward in that trend? And those two competitors with zone and satellite mobile service, those added services will be utilized for their price hike ambition, what's your view?

About the pricing, the inflation is becoming a challenge in the society in general. In that circumstance, the telecommunication industry is against the trend, meaning pricing is decreasing. This is industry structural issue or challenges. And one of the top carriers DOCOMO and AU, they started a price hike, which could be an opportunity for the industry as a whole because our industry, like I mentioned, has been deflationary industry as pricing has been going down, we try to address the situation by reducing cost, but we are beginning to see the limitation of how much we can do by ourselves. For the whole industry to grow in a healthy manner, maybe this could be a timing to revisit the pricing. As for SoftBank, we'd like to consider when and how it would be the best for us to act. Talking about added value, so we have been making pricing complicating by adding added values, if you will. So as far as SoftBank is concerned, we need to have more time to talk publicly about what kind of service offerings that we should add. We have some ideas in our mind, but it's too early to tell in the public what kind of added service that we have in our mind.

Next question.

Yamamoto from Niho [ph] Economic Newspapers. I have two questions. One, AI agent. So after fiscal year 2025, how you would make it profitable? And NTT has become the resale agent of the open AI enterprise product. So how would you position that? So this is my first question.

Okay, then I will answer one by one. About the AI agent, how we would make it profitable. That is something we need to spend a little more time in the next mid to long management plan announcement, I want to make it to happen there at that time. So about generative AI and others, whether AI is profitable or not or can be monetized? That is a discussion and this is just at the entrance. So it's not matured yet at all. So AI is center. So our society is going to move forward. So we do not have to hurry at this moment. And NTT data with OpenAI, so there is ChatGPT Enterprise standard version. So I believe that they made a contract over that, so which does not affect Cristal intelligence at all. So we are aiming to develop and provide even higher-performance AI agent product.

Thank you. The second question is again about NTT data. Under this reorganization of NTT, so how do you think this reorganization of NTT effect impact the market structure?

I don't know if I'm a right person to answer to this, but NTT's advantage or the interest is the overseas expansion and NTT data has an overseas expansion base as well. So we were also thinking, and we were having a discussion with NTT data for the further overseas expansion opportunities. Looking at domestic market of East and West of NTT and the data would just work for NTT only then the transmission lines and basis are unique assets. So those racks can be used through data or transmission lines, can be used through NTT data. If that thing would happen, then it will not be competitive at all. So therefore, under NTT Act, there should be a decent environment to remain for us to be able to have a fair competitive market. So it is understandable that if NTT wants to make overseas expansion, they would like to utilize NTT data.

Next question from the floor again.

Sana, Freelance Journalist. First question is about pricing of DOCOMO and KDDI. You also mentioned that additional values are added, but for the sub-brand products, it seems like they are defocusing on sub-brands as opposed to they are refocusing on standard or high value pricing. So what's your view on that?

Whether sub-brand is not profitable, I don't think so. Y!mobile, for example, we have a brand power of Y!mobile and Y!mobile should be an entry point for generic potential customers and then like we do Pay-toku want to upsell our products to those Y!mobile customers. And that kind of effort has been making fruit and I think it's better for customers to have many options and some users, we do not mind paying more for the services and other users may mind paying more. So if they want to refocus or de-prioritize sub-brands, they are different from us. So do you take it as an opportunity for us -- for you? We never know until we try whether it's an opportunity for us or not. But as far as SoftBank is concerned, we have no intention to defocus a sub-brand or making the brand smaller.

Next question, KDDI is a satellite connectivity and service offering. And they have a partnership with the SpaceX and Rakuten announced they want to offer service in 2026. For you, what's your view on satellite mobile communication and service offering?

It's different satellite configuration from the one we have been studying with OneWeb. Well, actually, next year, we have a plan to launch satellite mobile direct service as a SoftBank service next year, but we are still not sure whether the service should be free of charge or paid. At the moment, I cannot tell exactly who is the satellite operator that we are working with to launch a service next year. But again, we have a plan on our own.

Next question. So this will be the last question from the floor.

I'm Ishino, Freelancer. First question is about the churn rate. So as far as I see the churn rate is getting higher, what is the reason behind the -- Q4 is the seasonal timing. But compared to the previous year, it seems a bit higher?

Regarding churn rate, so there is a contract called SIM-only with Y!mobile. So therefore, there are a certain number of customers who convert to others in a short period of time. So it is not a surprise for us.

Thank you. Second question is about other carriers' price increase. So you were -- I believe the first -- I would say the first one to actually mention about price increase from before. So are you thinking about following them or do you think you can still bear with -- on the cost reduction? So it could be a great opportunity for you to acquire new subscribers or so new users. So what's your opinion on that?

Yes, I have a few things in my mind. Of course, price increase for our industry -- in order for our industry to survive moving forward, so it has to be a fair enough environment. And so considering the inflation or social issues, so I think it is the right direction to increase prices but we are competing each other. And we are doing well with revenues and also the ARPU is now almost at JPY10, so when we look at in mid to long-term perspective, we don't want to compromise ourselves in our tactics. So when will be the right timing for us to increase prices? And we would like to well consider the right timing and our strategy. So we are not talking about other carriers' products price structure with unnecessary things added. So by doing so, if the price was higher, it's not a healthy way or healthy structure in terms of price hike. So what would be a satisfactory approach for our customers or users? So of course, we can make the same services and products, but I do not believe that would satisfy our customers. We would like to spend more time thoroughly and then to come up with the -- a thorough strategy. Thank you.

Is there any other question you want to ask from the floor, maybe you're the last one from the floor.

Matsumoto from Business Insider. Thank you for taking my questions. First, about Sarashina and the timing of commercializing Sarashina. I believe you mentioned summer or autumn 2025. Is it like ChatGPT like chatting service, like a dialogue kind of service for consumers? That's the first question. Thank you.

Thank you. Maybe let me answer to the first question first, whether enterprise or consumer, we plan to offer the service to both segments. Obviously, percentage-wise, consumer could be bigger customer base, whether dialogue or chat type service or not, maybe that will be the start. But maybe video creation, image creation, we have some capabilities in our mind, but probably we take step-by-step approach, starting from tech space and expand service offerings or capabilities.

Thank you. The second question, there was a report that OpenAI gave up turning itself into a profitable entity. I wonder if this could have any impact on Cristal intelligence? Thank you.

I confirm that with OpenAI that at this time, our decision should not affect or impact Cristal intelligence at all. AI is evolving very fast and I understand their view of trying to make AI as safe as possible, and we will see how it goes going forward.

So that's all from the floor. Now we would like to take questions from Zoom. [Operator Instructions]. Bloomberg News, Hugo-san. Please unmute and ask your question, please.

I'm Hugo from Bloomberg News. Thank you so much for your presentation. One of the main topics today is PayPay -- preparation of PayPay's IPO. What is the purpose of that? I would like to hear once again. And you said that you started the preparation of IPO. What do you mean -- how do you define the start? What did you do for that? And also you -- they will be thinking about the list on the U.S. Stock Exchange and so on. So please share that kind of information in details.

Well, regarding PayPay's IPO, personally, I believe that it was a little too early. But when I hear what they say, considering the motivation of the employees, also the further competent talent acquisition. So from those perspectives, they had a strong will to go public, so which I understood. So with this opportunity, we also went through the same path. And we also had opportunity to grow even further with IPO. So after a thorough consideration, I decided to agree. So I believe that it would also lead to improve the corporate value as well. So I also mentioned that to consider going and considering overseas expansion, but we are not in a position at SoftBank to make any comments on this. And so far, the governance enhancement of PayPay, and we have been preparing for IPO in different perspectives. And also we were talking about the very specific more concrete item is now in discussion or a consideration. So now the discussion is more in detail. That's why we said that they started preparation.

So you said that you expect overseas expansion of PayPay. So it means that PayPay in the future would consider overseas expansion? Is that right? And in that case, the listing overseas and that will be on the U.S. Stock Exchange. Is that a correct way to understand?

Well, I'm very, very sorry, but it could lead to the pre-solicitation. I don't -- I cannot say any details, and especially the first part of your question, I cannot touch upon much. So why I said that I expect overseas expansion is my expectation to them. So whether PayPay would do so or not is not something that I can say right now. The second point and which area, which market is that -- so where PayPay's corporate value can be highly evaluated, then I think the U.S. is quite reasonable as an option. So if U.S. is one of the options, then I believe that, that is really reasonable to think.

Next, from the Zoom, Ishka-san, a Freelance Journalist. Please unmute and ask a question.

You mentioned that you plan to launch the satellite mobile direct service next year; whereas NTT Group announced that they will launch HAPS service next year. So what's your status on your HAPS initiative in SoftBank?

Thank you for your question. HAPS, we're having final discussion with regulatory body, and we'd like to launch HAPS service as soon as possible. Once we finalize discussion with regulator, we want to make announcement accordingly.

And according to the Opensignal, your network quality is rated poorer than KDDI, what's your view?

I think competition is good because it should drive improvement of the industry. We will definitely win back the position next time. And I believe that my engineers are listening to my message. And also, we want to be positioned as a threat to President Matsuda.

Thank you. Next question is from [indiscernible] from Toyo Keizai. Please unmute and ask question, please.

I'm Maki instead of Tayama. I have a question about data center. I have two questions around that. First, what bit collaboration, you are one of the members of that. So how do you evaluate the purpose and the objective of this initiative? And also you have an idea to expand and rollout the data center areas distributed throughout nationwide. And the second question is about with the foreign affiliated fund, I think data centers are being built, but this also causes the conflict with the local people, residents. So regarding data centers, any regulations and rules, what do you think about those establishing the regulations around the data center construction?

First, about what bit collaboration, yes, I'm also a member of this discussion. And what bit is to keep the balance between the demand and the supply by pulling the power cables and communication cables together. So to make the more effective collaboration between the power and information communications. So this is a very important discussion in order to proceed the realization of core existing society with AI. So this discussion has to be actively made time to time and also to be able to be used by the industry as a whole. So we -- so even though there was area that we wanted to build the data center, but due to the lack of power, it was not feasible. But this collaboration should be able to solve such challenges in the future.

And the second question, so in Inzai City in Chiba Prefecture, the Shima and Mie Prefecture. Now the submarine cables are -- were there. And now data centers are concentrated in that area. And if more data centers are built in inside city alone, that will be overwhelmed.

So when we look at local areas, it's more obvious and tens of megawatts, and there are some blacklist. So the balance between demand and the supply is really important. Of course, we do not believe that we can do anything we want to do in terms of increasing the number of data centers. So we need to solve the social issues. We also -- we believe that the government also should set some certain level of degrees of rules and regulations around that.

Now we'd like to take questions from analysts and institutional investors. First, Kikuchi-san from SMBC Nikko Securities.

I have two questions. Miyakawa-san mentioned that return to shareholders and potentially more return to shareholders, which was a positive surprise for me. More specifically, under what conditions you may consider share buyback or increased dividend? Talking about return to shareholders. Profit of LINE Yahoo, which is a listed company is not going to be used for you as a dividend source. So as SBKK stand-alone, do you have plenty of capacity to add more return to shareholders? So again, the question is, under what condition you consider adding more return to shareholders.

Like I described in the presentation under certain inflationary society, keeping your dividend as is effectively lower dividend. At the same time, in order for us to increase corporate value in the long run, we need to keep investing for the future. So if we could make a profit, even investing for the future, definitely, the capacity should be used for adding more return to shareholders. In the next mid-term plan, whether it be increased dividend or share buyback, we want to consider as one of the commitments in mid-term plan next time. But I just have an idea, but I don't have a concrete plan. But like Kikuchi-san mentioned, SoftBank KK's stand-alone profit is a dividend -- source for a dividend. So if we make more profit as SBKK stand-alone, we should be able to use that capacity to add more return to shareholders. Proactive investment in AI as much as about JPY100 billion. And if we could save half, then it could help us to make a stand-alone profit, then we will consider how we can return to shareholders.

Thank you. So aside from next mid-term plan, for the particular fiscal year, you may consider doing something as more return to shareholders?

Yes, we are open.

Second question is about PayPay's IPO. I understand we cannot ask you detailed questions. But generally speaking, PayPay or finance has the biggest potential for growth in our -- in your segments and if PayPay goes public or financial business goes public, then SoftBank KK cannot control the company or business anymore. For example, LINE Yahoo, it looks like it's not really reflected in your share price because someone like LINE Yahoo is out of your control, as a listed company. If an important segment goes out of your control, especially if or when the business goes public abroad, it's for array from your control, which could be a waste of your value or assets? I understand you have -- you're mindful of corporate value and I'm a little bit concerned about your group business management, if your important segment goes public.

Personally, I think it's too early, and I wanted to keep PayPay as long as possible or a little bit longer. But LY or PayPay, we were involved -- we involved in establishing a business. And eventually, we expect them to go public anyway. So it's not necessarily out of our plan. From our strategic perspective, we don't see LY as outside of our group. They are in our family. I'm not talking about numbers or our equity interest, but rather services that we want to offer and the direction that the humankind should go to. So we have the same vision with those group companies. Whether PayPay goes public or not, we remain on the same page with regards to the vision, we are looking ahead to. And hopefully, I don't know if I say this or not, I forget about this. Anyway, hopefully, we want to deepen our relationship with them going forward. So I hope that you think we are looking far ahead.

Next, Masuno-san from Nomura Securities. Please unmute and ask your questions.

This is Masuno from Nomura Securities. I have one question about JPY100 billion of negative due to the upfront investment. I don't know -- I cannot understand the background of this negative figure. And the Sakai data center with this impact, and I believe that there will be more cost on the development and the research and development cost. So as you explained about large LLM, and it would take another year or so even it's been taking a year-and-a-half already, so which is taking more cost so -- but still, you need to spend the cost on the research and development. Is that the reason of this JPY100 billion. So please elaborate on the background of JPY100 billion.

This is just a rough number. We do not have the detailed plan on this. So JPY100 billion, we wanted to be able to spend this much -- this amount of cost on the research and development. So that's what I had planned when I became President. So even with the accumulation of JPY100 billion of R&D, still the company would be on the stable operation. This is what I wanted to do. So that's why this rough JPY100 billion, I said that it would be for R&D. It's not that we are going to spend all of this. For now, the last year and a little bit more, this is what we expect for this fiscal year. So if anything remains out of JPY100 billion, then there will be an upswing of the operating income, which is a good thing. So we just want to become a company with the abundant R&D cost -- R&D budget.

Thank you. We'd like to take the last question. And in the interest of time, please keep the question one. So Tokunaga-san from Daiwa Securities. Please ask us one question.

About next mid-term business plan, so competitors are raising prices and enterprise where the Cristal intelligence is coming up, then enterprise business could grow further. Should we expect continuous growth? Or in the third phase, would you like to invest more like JPY100 billion? So what kind of consideration we should have in our mind when we are looking at your third phase of mid-term plan?

I cannot disclose a lot of numbers now, but the things that we are working on, for example, data centers, our own GenAI, like Sarashina, we want to show in numbers in those initiatives that we are working on in the next mid-term plan. So we can show grown -- even further grown SoftBank in the mid-term plan.

That's the end of Q&A. Thank you very much for joining us, SoftBank Corporation's earnings results for the fiscal year ended March 31, 2025.