Operator: Welcome to the SoftBank Group Corp, earnings results announcement for the Q2 of the fiscal year ending March 2027. Before we start, please turn your mobile phone to the silent mode. First of all, I would like to introduce today's participants. From left, we have Yoshimitsu Goto, Board Director and CFO. Kazuko Kimiwada, Corporate Officer, Senior Vice President, and CAO. Mark Agne, Executive Partner, SB Investment Advisers and SB Global Advisers. Ian Thornton, Vice President, Investor Relations. Today's announcement is live broadcast over internet. Now, I would like to invite Yoshimitsu Goto to present you the earnings results and business overview.
Yoshimitsu Goto: My name is Goto, CFO, SoftBank Group. Thank you very much for your time today. First, I would like to touch on the environments and today's circumstances. In United States, after the midterm election, there are several movements going on, and we would like to stay focused how that is going to change or move. Understand there are some concerns over corporates, AI corporates, hyperscalers. Some people talked about credit risk for those hyperscalers. Some people talk about data center construction in United States. Some people also worries about the job security or working environment security.
Yoshimitsu Goto: Each one of them is something that we also need to focus on and understand the situations, as we are also the company which has a relation to AI business. But when you look at the human history that we've been experiencing, the industrial revolution, and trying to achieve the next level after developing technologies and so on, and with that we've been able to see the great changes in the working styles or lifestyles of the people. I believe that we are at the kind of transitions to the next levels with AI, and we believe we are making a progress one step ahead towards AI, but that doesn't mean that there is any ends of these steps.
Yoshimitsu Goto: It is like a space universe that we've been keep moving up or we are making a step forward. Technologies, I believe it's irreversible. Not refusing such developments of the technology, but rather, we need to learn and understand how can we live together with those technology so that that can be changing to our lifestyles and working styles. One can be a challenge for us, and that also relates to our corporate philosophy, which is Information revolution, happiness for everyone. I believe that this corporate philosophy has been there for our company in the past 43, 44, 45 years, I believe.
Yoshimitsu Goto: I think it is also a need to understand how can we benefit from such technology or AI. I think that we also need to play a role to actually communicate and share such benefits. As I showing you this slide, this is just a part of the benefit, and this is the kind of benefits that we learn and hear from the OpenAI example. For example, ChatGPT users save an average of 40 to 60 minutes per day at work, or 300 million or over people each week use ChatGPT for health-related advices. Or 75% of ChatGPT users at business can perform tasks they previously could not. These are the kind of one of the examples of the benefits that we've been receiving from AI. There are several positive effects to your daily life.
Yoshimitsu Goto: Masa, our Chairman and CEO, often says that back in year 2000, when we called the internet bubble, if we look back that moment, one measurement can be the number of users of internet, for example. In the past history of human being, the number of users has been growing, keep growing, and such AI services are adding even more users on top of such number of users of internet. How are we going to behave? Not really betting around just with ideas, but because we believe that the benefit that we will be able to receive from such technology development is unlimitedly growing, and with the understanding of the advancement of AI and technology, we would like to steadily make forward steps. For example, we have been always sharing our focus in four pillars.
Yoshimitsu Goto: One is the AI model, which is represented by the investment in OpenAI that we completed JPY 30 billion in follow-on investments. The second pillar is AI chip. Arm actually has been making great progress in the past 3 months, and Arm's share among major hyperscalers is exceeding 60% by the end of 2026. AI infrastructure, data center, or data center-related companies investment is the action that we have been making. For example, SB Energy, which is our subsidiary, has signed the 20-year lease agreement with OpenAI for data centers with 8 gigawatt IT capacity.
Yoshimitsu Goto: Also, we have completed the acquisition of DigitalBridge. We are very glad to understand that they have great expertise, know-hows, and experiences in the digital infrastructure investments, and hopefully we will be able to collaborate together going forward. The fourth pillar is Physical AI. There are several progress. For example, Gravis Robotics, which is the company that developed the autonomous driving AI for the construction machines in Switzerland. Also, AutoStore in Norway. This is a company that developed the automation system for the warehouse, and they have signed the strategic supply agreement with Amazon this time. Zipline provides the distribution service using the autonomous flying drone in United States.
Yoshimitsu Goto: With this financial round, they have raised $1.5 billion at $21.5 billion valuation. I will be touching on this company later, but they have been actually making it happen to provide several services in United States. These are the four pillars that we have been focusing in this quarter. These are the results as a result of us making progress forward. Because it is a quarterly result, it is only for the 3 months period of the results, which is a little short. I would like to share the progress in the past 3 months, and also financial highlights that I would like to touch on for today's opportunity. There are four highlights.
Yoshimitsu Goto: First, investment gain, JPY 1.7 trillion. Net gain or loss, this time is net loss of JPY 0.8 trillion for the H1 of the fiscal 2026. Net asset value, which is an important measure for us, JPY 59.8 trillion. This is the JPY 20 trillion increase compared year-on-year basis. But the market has changed and declined largely, so our asset also follows suit. This JPY 20 trillion increase actually comes down, and I will come to the numbers later. Also completed JPY 30 billion follow-on investment in OpenAI. The last, for this Q2, I believe, as long as I understand, quarterly financing-wise, this was the largest financing quarter in the history.
Yoshimitsu Goto: About JPY 9 trillion or $56.5 billion has been raised in this Q2. Utilizing the financing proceeds, we have repaid $40 billion bridge facility ahead of schedule, which was originally structured early this year. The consolidated results, H1, JPY 2.2 trillion declines in the investment gain or loss resulted in JPY 1.7 trillion. Last quarter was a record high, so compared to the last quarter, it has actually a big decline from the previous quarter, JPY 2.2 trillion. JPY 1.7 trillion itself is something that I believe is a good enough number. Also, income before income tax or net income, respectively, you see the loss of JPY 438 billion or JPY 768 billion, respectively. On your far right of this graph, this is from the evaluation loss of Intel. That was one of the largest factors.
Yoshimitsu Goto: As a result, -JPY 133 billion loss. This is a quarter number, but we are not very much chasing the short-term ups and downs. In full year, we would like to see results as a business. For the net income, -JPY 1.1 trillion. There are two major reasons. First of all, we recorded loss on power generation-related equipment for AI data centers in the United States. It is about -$3.7 billion. Another one is that we have raised $7.4 billion in August 2026 through prepaid forward contracts backed by Intel shares.
Yoshimitsu Goto: After then, Intel share price declines. This records the derivative loss. If the share price increases after the derivative transaction, that also records the same way due to the increase in Intel share price following the contract. In the United States, the impact to P&L from the AI data center equipment, I would like to break it down. First, equipment, because data center is going to be a rush in terms of the constructions and competition is going to be fierce going forward. There is a relatively long lead time equipment there for the data centers. For those that we have procured in advance. For the data center project, so the power project, before they are finalized, we have placed an order and also procured.
Yoshimitsu Goto: Later on, that is going to be allocated and assigned to each respective project. We made bridge investments in advance to secure power generation equipment in anticipation of shortage. $9.5 billion here is the cumulative amount, total amount. This $9.5 relates to how we have been procuring the equipment. As I mentioned earlier, while we start launching the project one after another, those equipment is going to be allocated and assigned, and then we collect the investment. That was a flow that we have been expecting.
Yoshimitsu Goto: Later on, because such project has been witnessing the rapid development of the technology, and we also need to aim for even better results of the project, we had the review of the project portfolio. Based on that, we have revised certain power generation equipment procurement strategy. As a result, we recorded a loss related to non-refundable advanced payment and modifications to signed contract, which resulted in $3.7 billion losses. Also, while we are discussing with the vendors, we will be making purchases. At the time of the purchase, we have been entitled to rebates on future equipment purchases up to $2.1 billion.
Yoshimitsu Goto: This data center project is just the beginning, and it will take next 5 years, next 10 years, and so on, and the size is going to be growing. Based on that assumption, for the mid-term or mid to long-term, we believe it is positive to be entitled with such rebates. This 2.9 billion loss has been recorded, and the balance as of 30 September is $6.6 billion. This $6.6 billion equivalent is going to be utilized for PORTS Technology Campus power generation projects and so on.
Yoshimitsu Goto: About 9 billion the equipment has been reviewed, and followed by the review of the project portfolio, we have revised the strategy, but the remaining is going to be utilized for the future projects. Also, we have entitled to rebates on future equipment purchases. Along with the new challenge that we are making, this accounting treatment has occurred. From the accounting perspective, as of today, what do we need to record? What do we need to value? These are the numbers that we need to recognize as of today, and that has been fully reflected to our accountings.
Kazuko Kimiwada: Let me talk about key indicators. Net asset value, loan-to-value, and cash position. Those are, to us, most important key indicators. As for net asset value, it was about JPY 60 trillion as of the end of September. Loan-to-value, 15.9%. Cash position, JPY 5 trillion. From a safety perspective, those numbers are what we are pleased with. Looking back, in terms of net asset value, for example, second from the right, JPY 59.8 trillion was as of the end of September. Since then, to date, taking into account of the latest share price, the NAV as of today is JPY 44.7 trillion. It is down from JPY 59.8 trillion. However, from global perspective, this is not something that we are uncomfortable with. This chart starts the Q1 of FY 2024 and shows quarterly trend.
Kazuko Kimiwada: From that long-term perspective, you may say that you are coming back to where you were in 2024 in the only last 3 months, but that is necessarily not true. SoftBank's net asset value in the long run is going through like this in this chart. Far left, FY 1998, there was internet bubble era, and the share price increased sharply, but the market went through ups and downs, and some startups may have gone bankrupt, and the net bubble got bursted in the end. But looking back, what was the net bubble, you may ask? Because some companies grew since the net bubble burst, and a few of them has become what now we call hyperscalers. Again, my point is, we are not swayed by short-term up and down, and rather, we want to keep focus on a long time horizon. Now, long-term value.
Kazuko Kimiwada: It was 15.9% as of end of Q2, but the net asset value went down by JPY 20 trillion, like I said earlier. Compared to net asset value, it is hard to say how much it is as of now, but it is suffice to say it should be higher than 15.9%. I should have said earlier when I talked about NAV or net asset value. OpenAI is going through a new finance funding round, which I am sure you are aware of by looking at media. But until official announcement is made, we have not factored in that numbers in our report. So once official release is there, and then OpenAI's valuation is finalized and confirmed, that will be reflected in our long-term value and net asset value, which is likely in Q3. So we should expect good representation of OpenAI's good valuation.
Kazuko Kimiwada: Then it should be positive to LTV as well as net asset value. Nonetheless, as of today, we are well under 25%, which is suffice to say we are operating the business in a very safe manner. Cash position, JPY 5 trillion as of end of September. But on 1 October, we invested in OpenAI. That said, still, even as of today, we have plenty of cash position. Let me talk about Arm. In Q2, they hit the record high of revenue, increased by 27% year-on-year. In terms of EPS, fully diluted, again, they had a record Q2 EPS or 33% increase year-on-year. Here is Arm's guidance for Q3 in terms of revenue and fully diluted EPS. For revenue, $1.54 billion or 24% up.
Kazuko Kimiwada: Of course, there is a range of ±$50 million, but well above analyst consensus. Likewise, for EPS, above analyst consensus. So the quarter continues to grow in a very steady manner. About Arm's business, adoption of Arm technology in data center is growing rapidly. As you know, data centers are power-hungry, and that might be one of the reasons why some people are not happy with data center. But Arm's CPUs are highly energy efficient, so the total shipment of Arm's technology has grown rapidly, and it has been growing exponentially since September 2018. As announced, Arm AGI CPU, which is driving Arm's data center share, it is designed for agentic AI, and quality is well recognized. Compared to x86-based CPU, Arm's has two times better performance with the same power. Like I said earlier, Arm's expanding share at top hyperscalers over 60%.
Kazuko Kimiwada: Guys like Amazon and Microsoft are adopting Arm's technology more and more. It was only 15% in 2024, but now well above 60%. Let me shift focus to OpenAI. Cumulative investment amount, JPY 64.6 billion. We completed a follow-on investment, and we have about 13% ownership in OpenAI. OpenAI's growth is supported by those integrated three layers. Those layers are enhancing OpenAI's value each other. Layers are products and platforms, AI models, and compute. They have got the full stack approach.
Kazuko Kimiwada: Again, those layers are enhancing OpenAI's value. Feedback from the customers are reflected in enhancement of the model development, and that should help developing better and more enhanced models going forward. So how good open model is, it is shown here on this chart. This is an assessment by a third party comparing different models by different companies. The vertical shows performance, horizontal shows cost. There are Anthropic and Google models. As you can see, top left, OpenAI, it is great performance and lower cost. OpenAI is very distinctive compared to other models in terms of performance and cost.
Kazuko Kimiwada: The higher it goes, the better from cost per performance per cost perspective. That is how the graph should be interpreted. ChatGPT has been continuously a weapon of OpenAI, especially for consumers, and they are building a massive user base. ChatGPT weekly active users hit 1.2 billion as of September 2026. It has been only four years when this was launched, and it is very unique in terms of such a high growth of a customer base. This slide shows rapidly built a massive user base, and compared to other products, ChatGPT reached 1 billion users only in 3 years, which is the fastest app in history to reach that level.
Kazuko Kimiwada: Again, in terms of consumer segment, ChatGPT is one of the growth drivers. Their ARR reached $1 billion in August, or 10 times just in 5 months. I have been talking about consumer aspects of OpenAI's product, but in a competitive environment of AI, enterprise business should be better in terms of contribution to model companies' bottom line. In fact, OpenAI achieved original target 6 months ahead of schedule. Currently, enterprise share of total revenue account for 50%.
Kazuko Kimiwada: Before, people criticized OpenAI that OpenAI was doing business only for consumers or they are not strong in enterprise area, but OpenAI is expanding enterprise business like this. Codex has been adopted much wider than before. Initially, Codex was designed for software engineers, but now Codex is integrated in ChatGPT Work, which is for enterprise use cases.
Kazuko Kimiwada: Since the integration, the adoption has been growing, and in fact, combined weekly active users reached over 40 million. It has been, again, only 8 months since the launch in February, and it grew by 40 times, which is amazing. OpenAI's business model and financial soundness have been much, much more stronger in the last few months. OpenAI's ARR as of September was about $50 billion, or it grew 25 times in the last 3 years, but especially in the immediate last month, it was up 70% from July to September.
Mark Agne: Next is AI infrastructure. Growing AI usage actually drives data center demand. Last year, installed capacity was 82 gigawatt. Based on the numbers, 5 years later, 2030, demand forecast is about 274 gigawatt, which actually exceeding the demands largely. The gap of the demand and supply, and this is actually required for the latest edge information services. It is necessary to make an effort to satisfy such demands. These are the latest portfolio of SoftBank Group for the data center. In the middle is the PORTS Technology Campus, 8 gigawatt. Off-taker is OpenAI for 20-year contract. On your left, Milam County AI Data Center, 753 megawatt.
Mark Agne: Again, OpenAI as an off-taker for 15-year contract. Far right, AI data center in France, which was announced recently. In total, 5 gigawatt is aimed. Off-taker is to be determined. We expect there are many demands from corporates, and lots will require such capacity. From the financial point of view, it is important that we make sure that we have a safety and healthy financial management while we operate this developed data center project. SBG is expected to be founded through project finance, non-recourse to SoftBank Group, referring to the high credit offtaker. On your right, you see the scope of delivery. Data center building with power infrastructure SoftBank Group will be working on.
Mark Agne: There, it is going to require servers, GPUs. Such IT equipment will be prepared by offtaker. Therefore, while securing our financial discipline, we will be developing our data center projects. That is how we are proceeding our projects. The progress or the status of PORTS Technology Campus project, which is the joint project with government. This is located in Ohio. Back in August, SB Energy, our subsidiary, signed 20-year lease agreement with OpenAI. NVIDIA agreed to provide the credit support to OpenAI for 4.25 gigawatt portion with IT load basis. NVIDIA is going to supply the compute infra of the data center, and also committed to invest $3 billion in SB Energy. Out of $3 billion, $1.5 billion has already been funded, which is very positive in terms of enhancement of their capitals.
Mark Agne: For French project, total 5 gigawatt targeted. First, by 2030, we would like to achieve 3.1 gigawatt. 1.3 gigawatt in Dunkirk has also been assigned as a French government fast-track designation. 1.0 gigawatt in Bosquel, this has already been assigned, designated as the French government fast-track designation. We would like to well collaborate with French people to make this project in a good progress. As you are aware, there are several concerns been expressed over data center. Therefore, we would like to make sure to work in partnership with local communities. By doing so, hopefully that we will be able to have a good understanding from the local communities about our data center project.
Mark Agne: For example, Milam County in Texas. They have a track record of developing 900 megawatt of power generation capacity in Milam County. There are some concern that you will be out of water or you will be out of power, but actually, there are already the track record of developing such megawatt of the power generation. As for water, understand there are some concerns over it. We may be overusing water for the data centers and so on, but we are having initiative to minimize water use with closed-loop air-cooled system. In addition, we will be having such initiative to support the local economy by hiring and sourcing locally.
Mark Agne: Through such initiative that we would like to proceed our projects. In Malaysia, we also making the advancement of AI infrastructure. We made an announcement yesterday. Together with Grab and also PETROS, the state-owned energy company of Sarawak, Malaysia, that we announced together to plan to build AI infrastructure. It is going to be the first project. The planned initial investment is $1.85 billion, and we are taking about $1 billion, Grab $750 million, and PETROS $100 million. In total, with those three players, that we would like to aim for the success launch of this and the success progress of this project to develop 5 gigawatt or more of AI compute capacity over the long term.
Mark Agne: In infrastructure business, we made an investment in DigitalBridge. Acquisition transaction has been completed. Regulatory approvals all provided. They focus on digital infrastructure, and the asset under management is about $121 billion, which is a quite a large global asset manager. Investment about $3.1 billion, which is already completed the end of September. They do have a very attractive portfolio underneath of them. On the very top is the data center-related portfolio companies.
Mark Agne: They have invested, for example, Vantage, Switch, both of them are very well-known player in the market, as well as DataBank and other cell towers investments. For example, JTOWER in Japan has been invested by them as well. Next is Vision Fund that I would like to share with you the current status. Investment gain or loss in the past 3 months, not much big movement this time for this quarter, slightly increased. The breakdown, you see Vision Fund 1 and Vision Fund 2.
Mark Agne: For the case of Vision Fund 1, there are several positive factors such as ByteDance, but at the same time, as a public company, Coupang has been experiencing the decline in share price, so net-net, it is about flat. For Vision Fund 2, there are several portfolio companies actually showing good results, but at the same time, Silver Bands 7, which I referred to earlier for the energy-related investment, was done together with SBG, and for that, we have loss. Still, the result was +558 million. The cumulative investment return, I believe that it is good to see for the long term and for the Vision Fund 1 showing a very good return, JPY 113 billion cumulative investment return. Vision Fund 2 is actually also showing the good results, JPY 144 billion cumulative investment return.
Mark Agne: Of course, we have been experiencing ups and downs, but overall, I believe that they are making a good direction. Investment monetization is actually going smoothly. Steady progress has been made, like Lenskart been monetized in total about $1.5 billion investments sales. Not only public securities, but for the private companies, there are some buyers who has interest that we have been monetizing, which leads to this result. Listing pipelines, those which we have finished the filing for OpenAI, IPO is OpenAI or SB Energy. Those two are one of the material pipelines which has already filed for IPO. Prism, which formerly known as OYO, is another one for listing pipeline. On your right-hand side, ByteDance, Revolut, Fanatics, Nuro, those are also the pipelines or the portfolio companies that we have a good expectation in the future. The total fair value of those is $147 billion.
Yoshimitsu Goto: Now, let me talk about Physical AI. We have three pillars here: autonomous mobility, automation infrastructure, and robotics. Robotics, it is going to be leading to humanoid in the future, and robotics might be most relevant to general public like us. We invested in 30-plus companies in those key areas. Let me pick up some points. First, Gravis Robotics. What they do is they retrofit existing construction equipment with sensors and AI control systems to enable autonomous excavation and loading. This solution should help address labor shortage and improve productivity on construction sites. Now, AutoStore. We invested in this business a long time ago.
Yoshimitsu Goto: This is Norwegian developer of warehouse automation system. AI-powered robots automatically transports storage bins, and AI-powered robots automatically pick items in a warehouse, and they are well recognized. In fact, in August, they entered into a strategic global supply agreement with Amazon. That was one of the biggest news back in summer. The third company that I wanted to introduce to you is Zipline. This is a US autonomous drone delivery service provider. They raised $1.5 billion at a $21.5 billion valuation. They work with Walmart, for example, to launch 300 additional sites through its partnership.
Yoshimitsu Goto: Uber is targeting 1 million deliveries per day with Zipline by 2029. That should be a huge step toward the time when many people enjoy Zipline. Let me share with you the video now. It is kind of amazing, and I hope that they can address the concerns around thieves of delivered items. Uber sometimes struggle to hire workforce, and there was a dispute between the management and the unions. But for the users and for the operators like Uber, this is a great technological advancement that is useful and helpful. Last but not least, let me share with you once again the financial policy, which remains unchanged.
Yoshimitsu Goto: As you know, we maintain LTV below 25% in normal circumstances and maintain at least two years' worth of bond redemption in cash position. This is very simple, which is good for not only professional institutional investor, but also retail investors. In fact, we have a lot of retail investors helping us and supporting us. Also, we have retail creditors. We want to represent easy-to-understand financial policy as our fundamental position. In Q2 FY26, like I said earlier, in Q2 this year, I think that we had the biggest financing quarter in our history.
Yoshimitsu Goto: Part of the finance money was for investment and other for repayment or redemptions. In fact, we fully repaid bridge facility before maturity. As for investment, this chart compares FY 2025 and 2026. Since the beginning of this fiscal year, we have invested in OpenAI in three tranches, and also we invested in DigitalBridge, ABB Robotics. Those are big investment activities this year. So maybe for the full year, we might end up investing more than last year. About repayment of bridge facility before maturity. In February and March, we agreed with OpenAI to make an investment of JPY 30 billion. Back then, we signed bridge facility of JPY 40 billion. So we told OpenAI that we already had done financing to invest into OpenAI of JPY 30 billion. How much has been used?
Yoshimitsu Goto: In fact, this facility allows us to invest not only OpenAI, but also other businesses like DigitalBridge and ABB. So we drew down from this facility to invest in DigitalBridge and ABB. The last tranche of OpenAI is supposed to be done on 1 October. As of September, we had undrawn commitment. That JPY 10 billion for OpenAI's final tranche was financed separately, so that is why we canceled this undrawn commitment. In the end, we were able to repay fully ahead of schedule, which is planned March 2027, but again, 2 months ahead of schedule, we were able to repay bridge facility. I think from bank's perspective, the early repayment would be appreciated. Because if repayment is not completed closer to the maturity date, banks are getting worried and agitated. I think from relationship with bank's perspective, that early repayment was good.
Yoshimitsu Goto: The permanent finance or long-term loan. Bond issuance, including issuance in the US, and asset monetization. Those are the sources for repayment of a facility. This is a very traditional method for us to tap into, and those are the sources for repayment, as you can see here. In Q2, if you go down deeper, those are the financing highlights. By combining 7, 8 methods, each and every method is challenging, and we have very busy time sometime. But we have an access to the retail market, retail bond market in Japan and in the US. We issued 144A basis bond in the US. From asset-backed finance perspective, we utilized margin loan backed by Arm shares. We had loan using OpenAI shares. With combining those financing methods, we raised funds in Q2 FY 2026.
Yoshimitsu Goto: In August, we raised $7.4 billion through prepaid forward contract or collared contract. The timing was good. For credit ratings, we obtained a new issuer credit rating from Fitch. Fitch is now with us as a credit rating agency. Fitch rating, it is BB+, outlook stable. So we will make sure that we stay communicating with rating agencies as usual. Before closing, just so you know, Fukuoka SoftBank Hawks won the Japan Series Championship once again this year, which was great news. Masa has sent us a lot of messages. Well, he said at an interview the other day, "This is the AI era starting. History is going to be created." He also mentioned that what is going to happen is too big, so big that what I have done so far looks so small.
Yoshimitsu Goto: Of course, there are some concerns about security, and I think now is the time to take a look at AI once again to be ready for AI's growth. We do not want to compete against AI. We people have different lifestyles, and that is the matter of fact. The people's lives are completely different from AI's world. We ask ourselves who we are, and AI, as they grow, is going to be a part of our lives naturally and blended with us naturally. Eventually, our life will be built around AI in the future. That is how people have been evolving. So again, our corporate philosophy, information revolution, happiness for everyone. That is all for myself. Thank you very much.
Operator: Now we would like to take questions. First, we would like to take questions from the floor. Please wait for the microphone and start with your name and affiliation. For those participating online, please access Zoom using the instructions provided in the email sent to you in advance. To avoid audio feedback or looping, we kindly ask that anyone asking questions refrain from connecting to any live streams other than the Zoom webinar. If you have questions, please click the raise hand button and wait until you are called on. If you wish to withdraw your question, please click the lower hand button. If you are joining the Japanese Zoom, please ask your question in Japanese. We would like to take up to two questions per person so that we can take questions from as many people as possible. Now we would like to take questions from the floor.
Ian Thornton: The very front line with the brown jacket, gentleman, please.
[Analyst] (Bloomberg): MJ from Bloomberg. I have two questions, please. First, you are reviewing the project profile, you mentioned earlier, but you procured equipment in advance, and that has also been reviewed, I understand, from your presentation. Reviewing the profile, can you share with us more details about how you have been reviewing your project profile, for example, Ohio project? Are you expecting the downsizing of Ohio project, for example?
Yoshimitsu Goto: As for Ohio project, we are not discussing to downsize anything. Project itself is making good progress so far. The reviewing is the mid-to-long-term data center project strategy. What kind of strategy is most appropriate at the time is something that we always like to revisit. If it is the project that we already have a track record of 10 years or 20 years, it is not really the case. But if it is a traditional power project, because we have a track record that we do not have much cases like we have been doing this time by purchasing or procuring the equipment in advance to a project itself. But this time, it is the AI data center project, which is very new.
Yoshimitsu Goto: A year ago, we thought that this is good invest. Two years ago, we thought this would be the best. But later on, there are things that we believe that is most appropriate now with doing this way. That is something that the company needs to kind of reflect those situations. But at the same time, when you look at the demand and supply, the demand is very large, so we do believe that there are competitions. Therefore, sometimes we believe we need to procure in advance.
Yoshimitsu Goto: But at the same time, there are cases where we need to revisit and revise based on the most appropriate level for the projects. So, we would like to take an action as fast as possible, and that should be reflected to the accountings.
[Analyst] (Bloomberg): Thank you. To confirm, the recording loss, has that been canceled, the orders? The recorded loss is only at the level of paying the reservation fee, and that reservation fee will not be coming back. To revisit and review the contract, there are costs related to that. So, there are some additional costs onto it. So, adding all that, we are recording as a loss this time. Thank you. My second question, which is extension of my first question, and sorry about this is the other company's things that Elon Musk tweeted saying that SoftBank is defaulting on turbine orders.
Yoshimitsu Goto: Any comments on that?
Yoshimitsu Goto: We have no idea what kind of intention was that, but this is simply not correct at all. There is no such default. As I explained earlier, we remain focused and on track to execute our AI infrastructure plans.
[Analyst] (Bloomberg): Thank you.
Ian Thornton: Thank you. Next question from the floor. Yagi from Nikkei. First question about the midterm election in the US. Democrats won, and states like Texas and Ohio, the governors who were against data center buildings were elected. SB Energy and SB Neo, what do you think the result of the midterm election would have on those businesses? About the result of the election, we are not in a position to make any comment. There are professionals to analyze the result of elections. Regardless, results of elections, data center businesses and a lot of different opinions about data center business in the US have been what we have been aware of even before the election, and we stay focused on building good relationship with the communities.
Yoshimitsu Goto: Related to the question, data center project run by SB Energy, according to the prospectus, there were five projects, and it looks most of them have been not started yet. About the progress of SB Energy's projects, do you think we should anticipate some delay? The new US administration or leadership may require more time to discuss or review data center projects. But when or if such a thing happens, we will try make sure that our anticipated projects won't be delayed.
Ian Thornton: Next question is about ABB Robotics acquisition. The Chinese authorities issued on 4 November that they would not accept the acquisition, but you originally planned to complete acquisition by the end of this fiscal year. I wonder if that kind of a US-China, sorry, China's government objection would affect, and how are you going to complete that acquisition regardless?
Yoshimitsu Goto: As of today, I don't have any extra comment other than already publicly announced comments. But we work closely with ABB, and whenever an event happens that we should disclose, we'll definitely do that immediately. Nonetheless, we remain work closely with ABB.
Ian Thornton: Thank you. Any other questions from the floor?
Tatsuya Goto: Goto from TV Tokyo. Thank you very much. I have two questions. First, you talked about some concerns over AI, but 3 months before, there are some changes in the risk assessment. What kind of risks have been focused more compared to 3 months ago? Also, the data center project review, is there any big changes compared to 3 months ago?
Yoshimitsu Goto: The risk as an investment company is a question of how we are going to look at the risk. For example, in the case of the data center project, assume that there are some delays in schedule. That itself is not a material risk from the investment company's point of view. When it comes to more direct, for example, OpenAI investments, as an investment company, this is something that we are very much interested in. When it comes to fundamentals of OpenAI, they are doing quite steadily growing, as I mentioned in earlier slides. For the mid-to-long-term AI development and visions, we are continuously expecting their future growth. What is happening in the short-term period, we do not recognize as a material risk as an investment company.
Yoshimitsu Goto: When it comes to financial risk, for example, net asset value, which is one of the most important KPI for the company, if it is extremely declining, that will be the risk for the loan-to-value. That is something that we always need to pay attention to. As far as we see that KPI, we are seeing a very safe level. Compared to 3 months ago or 6 months ago, we have been able to achieve the upside, exceeding the risk or the negatives. 3 months later, we do not believe that we are facing the material risk since then. When it comes to a market crash or happened, that will be when there is any sign of those market crash, there are things that we do need to consider. As of today, I believe that we are making a safety drive of the business.
Tatsuya Goto: For the mid-to-long-term period, risk-wise, you do not really see the changes from 3 months ago?
Yoshimitsu Goto: Yes, you are correct.
Tatsuya Goto: The political landscape change in United States, I believe you have been having a good relationship with Mr. President Trump, which is quite good in terms of making an investment in United States. But we have been seeing the big loss in Republicans, but the top dialogue may have changed the scene, and that is going to impact your business itself.
Yoshimitsu Goto: Actually, the project itself has been already launched or scheduled to be launched one by one. As somebody asked earlier, the result of the midterm election itself does not really have an impact on the data center project overall. But of course, overall, there are some voices over the conservative view on the data center constructions, which may impact the overall data centers, not limited to ours project. But as you know, Masa.
Yoshimitsu Goto: The president's relationship does not really impact those itself. Thank you.
Ian Thornton: Any other question from the floor?
Masanori Yamashita: Yamashita from Yomiuri Newspaper. Related to that question, hyperscalers and AI developers are often criticized that it is a kind of cash is circulating among those supply chain. For example, NVIDIA and OpenAI and SoftBank, it is like structure is such way that monies are just circulating around those stakeholders. There are some external risks, like a market condition. IMF, in fact, said that there is a risk of going backward. Or risk around money circulation is often covered on the media, but is that a risk or not?
Yoshimitsu Goto: We need to assess, evaluate more in detail before coming to conclusion, because it ultimately depends on who provides credit. For example, if it is a bank, they look at how the money they offer is going to be repaid, and what kind of credit supports are provided to the borrowers. That is an aspect that money is provided by a bank. From a credit perspective, such a round-tripping transaction is negative or risk, because stock transactions does not matter to debt providers. Hyperscalers credit, the question is how the hyperscalers credit is viewed. Just an example, there are a lot of hyperscalers out there, Microsoft, Amazon, maybe not NVIDIA, but there are a lot of hyperscalers.
Yoshimitsu Goto: From our perspective, they are enjoying much higher credit than ours. Should we anticipate risk around the round-tripping transaction? If you look at loan transactions one by one, I am sure that as a money provider, they have concluded there should be no risk. Money provider, you mentioned, does that include SoftBank? We are not providing debt, like data center project or power generation project. We are equity provider. Banks and creditors are debt providers, and to them, the most important thing is where the money is coming back from. We are small equity provider, but the debt provider. I think you are talking about loan providers or risk for loan providers that might be brought by round-trip transactions.
Ian Thornton: Any other questions, please?
Yuta Asomura: Asomura from Nikkei Business. I would like to ask you about the financial strategy. The share price is declining, therefore that the net asset value has been declining compared to September. LTV has been widening, and based on that understanding, investment strategy may be defense mode instead of offense. Still, having said that, you have been able to securing the safetiness and healthiness of your financial policy, therefore that you are still in the offensive mode for the investment activities?
Yoshimitsu Goto: That is a very good point. In short term, financial action and the mid to long term financial action may different. Short-term period, we have been seeing some widenings. Under such circumstance, we cannot be the same behavior. This is not the environment that we can be all offense in terms of investment. We do need to take a look. It's not like a completely stop in terms of the activities, but I think that we should have some discussion that should take a look at the moment. I think that we need to kind of balance between gas and brake so that we'll be able to manage our loan-to-value less than 25%. I think that you are correct that we do need to have a balance and manage the gas and brake, and that's something that we should be doing for the short-term period.
Yuta Asomura: When it comes to long-term period, you still focusing all AI?
Yoshimitsu Goto: All in AI is not something that I can answer yes or no, but we've been preparing for the AI investment for the mid to long-term period, yes. Actually, for the mid to long-term wise, I think that it's a good time to buy. Usually, the best timing for the investment is when you see the price decline, but at the same time, that's the time that you need to take a look. So it's a little bit difficult to take a balance in between those two, though.
Ian Thornton: That's all from questions from the floor. Now, let us take question from online participants. Again, please turn off access to live streaming other than Zoom webinar. First, Mr. Chayama from Toyo Keizai magazine. I have two questions. First, about revision of program and plan of US power generation equipment. I think my question was, what was the original plan and what was the revised plan? Could you share with us as much as possible since it's going to be a long-term project?
Yoshimitsu Goto: So, I suspect you may need to revise the strategy. So again, how are you going to control risk of losses going forward? You are talking about our strategy, and our decisions around our strategy is something that I want to refrain from commenting on. But I think there are different aspects. For example, or maybe different types of discussion.
Yoshimitsu Goto: For example, what kind of equipment would be better suited going forward? Other discussion might be the equipment that we thought would be needed turn out to be challenging to acquire, more so than we thought before. So, we did that kind of revision or assessment in a very conservative manner, and we recorded impairment loss, for example. But I cannot promise it's not going to happen again. Going forward, we need to discuss how to evaluate assets that are not in use anymore. So next question is about AI infrastructure development in Malaysia.
Yoshimitsu Goto: In the US, because of the midterm election, I think it's more challenging to build data centers in the US. But investment in Malaysia is for the purpose of diversifying investment geographically, or in other words, how are you going to build a portfolio geographically in terms of AI data centers? Yes, AI data centers build, or product or services, or what we want to do to share good things with the people in the community. Talking about projects in Malaysia, we, as a part of Asia, we are pleased that we announced this kind of big project in Malaysia, and we hope that we have these kind of opportunities going forward again. It is not for the purpose of diversifying the risk or de-risking concentration. Not necessarily. We made this decision from global strategy perspective.
Ian Thornton: Thank you. For the interest of time, this was the last question.
Operator: Thank you very much. This concludes the SoftBank Group Corp. earnings results announcement for the Q2 of the fiscal year ending March 2007. This video footage of this meeting will be distributed on demand from our corporate website. Thank you very much again for joining the SoftBank Group Corp. earnings results announcement for the Q2 of the fiscal year ending March 2027.
