We'll now begin Fujitsu Limited IR Day 2026. Thank you very much for taking the time out of your busy schedules to join us today. This event is being held in a hybrid format, both at Fujitsu's Marunouchi office and via live online streaming. As for today's agenda, we'll begin with approximately 90 minutes of presentations, followed by approximately 60 minutes of Q and A for a total duration of two and a half hours. Now, let me introduce today's speakers. Takeshi Isobe, Representative Director and Corporate Executive Officer, Corporate Vice President, and CFO. Vivek Mahajan, Corporate Executive Officer, Corporate Vice President, and CTO. Shunsuke Onishi, Corporate Executive Officer, Corporate Vice President, and CRO. Yoshinami Takahashi, Corporate Executive Officer, Corporate Vice President, and COO. Megumi Shimazu, Corporate Executive Officer, Corporate Vice President, and COO.
Those are our five speakers for today. Now, let us begin. Mr. Isobe, please take the stage.
Good morning. I am Isobe, the CFO. Thank you very much for joining us today for Fujitsu Limited IR Day. Here at Fujitsu, we announced our mid to long-term Management Vision 2035 in May. Today, the corporate vice presidents in charge of each area will explain the individual business strategies and growth drivers that underpin this management vision, with the aim of clarifying the roadmap for realizing the vision. First, I will start by briefly reviewing the outline of our mid to long-term management vision as a recap, and then I will briefly introduce the topics I'd like to cover today and the agenda. This is the Fujitsu Group's purpose. Since I discuss this regularly, I won't go into details, but our mid to long-term management vision is nothing other than our efforts to realize this purpose.
Now, to recap, I'll briefly discuss the background and direction of mid to long-term management vision. This slide shows the positioning of the mid to long-term Management Vision 2035. Building on initiatives and transformations from past mid-term plan periods, we have defined the next decade as a period for technology-driven value creation. Precisely because we are in rapidly changing times, we decided to first firmly establish our mid to long-term direction and then move forward step by step each fiscal year toward that goal. We aim for sustainable growth by providing trusted technology and implementing AI-driven business. We set the timeline for 2035 because we believe that the next 10 years will mark a major turning point for society. Be it society, the environment, industry, or national security, the world is facing a variety of challenges, and rapid technological evolution has the potential to fundamentally transform existing rules and values.
To address these various social challenges, we will take a technology-driven approach to create value. Specifically, these are the three areas: trusted, energy-efficient, and safe and secure computing infrastructure, the Sovereign Platform. The human-robot collaboration with autonomous evolution or physical AI. And advancing decision-making by analyzing large-scale data through digital twins or intelligent society. These are the broad strategic pillars, creating new business through technological innovation and, of course, the further evolution of our service solutions business, which is the core of our current growth. Underpinning both are Fujitsu's unique strengths, the customer base we have cultivated over many years, the industry-specific expertise gained from there, and our proprietary technology foundation. This illustrates the drivers that will enable us to execute these strategies and achieve sustainable growth. The foundation is our highly reliable Sovereign Platform, with the AI platform deployed on top of it.
Then there is the application layer, represented by Uvance, and the professional services layer, including consulting and FDE, which engage directly with customers to deliver value. By integrating the business and the operational expertise we have accumulated, as well as our own and partners' technologies into this four-tier vertically integrated model and orchestrating the entire system, we will implement and deliver trusted optimum AI solutions for our customers. Today's presentations will be organized according to the strategies and growth drivers that are discussed earlier. First, on the creation of technology-powered new businesses, we will discuss the differentiated technologies centered on the Sovereign Platform and AI platform mentioned earlier. Next, on the evolution of our service solutions business at the professional service layer, we will discuss market demand and pricing strategies.
At the application layer, regarding Uvance's next phase of growth and the transformation of our delivery, the respective Corporate Vice Presidents in charge will explain specific business strategies. Finally, I will explain the framework for expanding cash generation capacity and optimal capital allocation. Now, this is about it for the opening remarks. Now, regarding Fujitsu's technology strategy, the source of our competitiveness and the foundation supporting all our growth strategies, Mr. Mahajan, please take the floor.
Good morning, everyone. This is Mahajan, I am CTO. So let me talk about leading innovation for AI sovereignty, and so let me talk about the technology. Under the current environment, AI sovereignty are the actual words that are catching a lot of attention. We started about five years ago to take on this technology strategy called AI sovereignty. To realize AI sovereignty, we have caused the innovation in how we evolve the innovation, how we will utilize those innovation. That is what I want to talk about today. This is one-page summary of our strategy. As mentioned by President Isobe, CFO Isobe, in May, we announced Vision 2035 to accomplish JPY 3 trillion in 2035, and we talked about technology to support them. There are three technologies. One is a sovereign platform, physical app, and AI and intelligence society. AI technology, Fujitsu Kozuchi, and computing for AI.
And one more thing, to realize AI is a network for AI. Those are three technologies, and I want to elaborate more on these three. Not just technologies, but also application and professional services are available to make sure we offer a high value to the customer. It is not just our technology. We are also working with partners to accomplish these technology. Our IP is the main focus to offer, but we also have global partners, NVIDIA, AMD, Yaskawa, and Broadcom, and there are so many other players as you see on the slide, and we are working closely with them. AI platform, Fujitsu Kozuchi. Let me start with this platform. Why we built such AI platform? How are we trying to win using this AI platform? That is what I want to explain first.
The most important part is for enterprise business, we are offering Fujitsu Kozuchi to be able to take the role of such a core business, first using multi-AI agent framework, and the other thing is Takane. This is an enterprise GenAI and domain-specific AI, and then vision. Our IP Amalgamation AI for vision is available, and Kozuchi Physical OS is supporting this. Security, this is a quite hot topic. A multi-agent AI framework, we offer the value in this framework is because that automatically generate AI agent, so we can adapt to different changes in environment, and then it can actually continuously evolve autonomously. According to the responses from customers, we offer 85% response accuracy, so it is quite high. Takane, this is the technology to support. We have four models. One is our own IP, 10 billion- 30 billion parameters are utilized.
And the other thing is more than 30 million different models on the field are produced and we are working. We also have 200 million parameters available, and there is also a large size. We are using 5 million parameters are utilized. So we do have a lineup to be able to cover a wide range of different customers. Vision, Amalgamation AI for vision, only a few pages of images are only needed. You do not really need to have a specific knowledge. Especially if there is the defense that requires a very high demand from sovereignty. This solution is highly appreciated by such users as a solution. Kozuchi Physical AI and Security. How do we connect them over to businesses? This is related to the vision earlier. Kozuchi is for sovereign platform and intelligence society and physical AI. So use this Kozuchi for all of them.
By 2035 we are looking to accomplish JPY 1.5 trillion in business out of Kozuchi. In physical AI, Fujitsu is making robots, so we will collaborate with the robot AI. We will have an open platform AI technology. Well, robot is representing one of the physical AI, but also there will be many different devices such as drones, and we are making physical OS. What is this physical OS? Robot, human being can work together on this software. Especially the major challenge for robot is the memory in the short and long term. There has been a lot of studies conducted, and it has been a major challenge. The robot, once they take an action, they tend to forget the previous action. So we needed to make sure the robot can be also utilized adapting the changes in the environment on the actual field.
Also, the interaction taking place to robots is something that we need. There was a major experiment the other day as well. In the real world, AIs are having conversation each other, and same thing with the robots. Having conversation between robot and make a judgment to run the business is something that we are trying to accomplish. In security, on a manufacturing field, a retail field, in the defense area, security matters a lot now, and we have a strength and advantage in the security area. This has already been announced with FANUC, NVIDIA, Kawasaki, Yaskawa, we are working with them, to form a software circle. It is not just with robot vendor, someone like NVIDIA and Qualcomm, we are also working with them at the same time, not just in Japan, but also trying to cover globally to build our Kozuchi Physical AI.
Of course, we are also working at the universities as well. So we are trying to accomplish growth and create a large business to get to JPY 750 billion.
Customer Zero for physical AI, we have this Kasajima plant. We use this as a dark factory. Fujitsu digital twin technology is utilized here, and it will cover the production generation and the planning. We also combine with manufacturing, testing, and the logistics utilizing physical OS. As you know, a robot is really good at repeating fixed behaviors, but if there is a change in environment in lighting, distance, the robot tends to get lost. That is where we utilize physical AI, and we are able to utilize our advantage there. Using Customer Zero, we have already introduced this, and we are evolving this technology for sure. Kozuchi AI stack is supported by security, such physical AI is supported by security. To secure Kozuchi and Takane, there are two AI security technologies. One is AI scanner, the other one is AI guardrails.
We have a quite high level of technologies available around the world. We also have the detailed numbers on the screen, which I will not go over too much in details, but need to understand it, like with the Claude Fable 5, GPT-6. We are quite strong for such high level of information. But they are not really strong at the information being stolen situation. Same thing for Takane itself on standalone. It is not something they have stronger performance than Claude Fable or GPT-6, but if we have guardrail, if we have scanner, then we can actually accomplish a quite high level of security. Security itself, this is the security for physical AI. I hope you want your understanding on this point. The security technology is quite important to accomplish sovereign technology.
Next is about AI platform realization. I am sure you understand this, but what is important is computing. I explained this last year, but from this fiscal year, we have been engaged in computing like K computer and Fugaku for over many years, and server is also being produced by ourselves. Going forward, what we attract attention is on Fujitsu MONAKA, the computing is centered around this. The day before yesterday, we made a press release. This is going to be launched into the market, and Fujitsu MONAKA will be the world's first 2 nm server chip.
What characterize this is, a s a recap from last year, on a standalone basis, there is a high performance and power consumption efficiency is really high, 65% power efficiency can be realized. Of course, confidential computing on a chip level. There are a lot of different details in the slide 144. This has been already tested and has been provided to other companies for testing, and we will start receiving orders from June. We have been engaged in CPU since five years before. From up until one and a half years, GPU related LLM training was the AI world. Now we have a generative AI, and it is not just CPU, but CPU will be used necessary. Intel and AMD CPU is hard to come by because of that. As for CPU, of course, inference is something that we are looking at.
It is not just servers, but edge and near edge influence is becoming important, so it is a very good timing for us to do this. MONAKA business, there are twofold, chip business and server business. For the first time for Fujitsu, we will take up the CPU business. In terms of methodology, Fujitsu main deal as a server and chip is going to be supplied to other companies as well. The Fujitsu MONAKA CPU incorporated server will be launched in the U.S. and other companies will also sell this product. On the right, the MONAKA model is going to be launched. This is conservative number, I believe, but these are the numbers that we have come up with. JPY 35 billion, which is going to be launched in March, and JPY 500 billion in 2035 on a standalone basis.
We have a roadmap and MONAKA and MONAKA-X and MONAKA-XX, and it is going to be a steady evolution. MONAKA lineup last year, we worked with RIKEN and NVIDIA to make FugakuNEXT announcement. What RIKEN said is that this is the best chip in AI era. Of course, we are aiming for that. There is a big potential that we see in this business. This is about performance metrics. There are numbers do not lie. On the left you see CPU specifications and linear DDR memory bandwidth, so transfer rate as shown. As for server performance, the chips are the basis for the servers and other companies have not disclosed their numbers. We are benchmarking in lab and what other companies are coming up with is shown on the left.
Probably we can aim for top level metrics. On the right the batch size two. All inference is not considered in the MONAKA, but as for edge and near edge inference, we will be able to come up with a very good specification or metrics. Of course, we are aiming for HPC as well. This is the roadmap. It is not just MONAKA, but MONAKA-X. MONAKA-X, there will be two different versions. The first one is CPU only model, the other is CPU plus NPU model. NPU engine will be incorporated as well. This is FugakuNEXT CPU as well. For general purpose, this is being launched. The software stack is completely open software stack, so there is no closed stack that is close to Fujitsu alone. Customers feel assured to use this CPU.
Japanese players are selling CPU. We are the only one among them. The chip from Arm peer software that has been acquired and also Graviton is there, but they have not launched the CPU, but Arm is the central player. We see a great potential there because of this competitive landscape. Now the quantum computers, it is not just computer, but quantum as a whole. It is very difficult technologies. This is a quantum chip. At the end of this year, 1,000 bit quantum chip, which will be the world's first, will be functioning in the Kawasaki Center. Only a single chip can deal with 1,000 bits, which is quite advanced technology. The other one, I am not going to go into details for this, but it is not just hardware for quantum, but there is also quantum application, quantum middleware.
We are working with other companies in quantum application. Obviously Quantum business is something that we are quite aggressive and proactive at. There are news being reported in the media, but we are coming up with still very conservative numbers. We believe that we can achieve these numbers, but there are some business inquiries and negotiations from universities and governments, inclusive of those in other countries. Quantum computing and quantum together can support the AI era. This is a quantum computer roadmap. This, I believe, is a leading roadmap, 256 by 2030, which will be the top roadmap. Fault-tolerant quantum computer will be produced by 2035. MONAKA HPC together with quantum computing can help us provide various applications to customers. AI software stack computing, and the third one, what is necessary for AI era, is network.
Of course, we have been engaged in network business for a very long time, so why now? Of course, data centers is the name of the game. We have been doing the business for carriers. 1Finity NIC is on the left. In the data center, what is going to be important is the scaling up and scaling out, and then scaling across. Between the data centers, if the distance is more than 10 km, then the data is going to be degraded. That is a key challenge. For 1Finity NIC switch and the Broadcom Tomahawk Switch, between Tomahawk and Tomahawk, the NIC will come in to connect them. RDMA or RoCE will directly connect the chips that we have the IP. People in the data centers are quite intrigued and appreciate this quite highly. As you can see from the numbers, the delay is reduced to 1/20.
There is O-Line and V-Line, a track record for us, but what is going to be used is A-Line for the AI. AI software stack and computing and network, combining all these three will help us move forward. Once again, I would like to emphasize is the technologies for national security. We will engage in those technologies, Fujitsu, Kozuchi, MONAKA, and quantum computing and network technologies. Everything is where we will work with the government as well. It is not just everything, but infrared sensors that we produce in Japan and multi-AI agent for defense purpose with a high level of security and quantum encryption or post-quantum cryptography or disinformation countermeasures. As you can see, Lockheed Martin or General Atomics are working with us, who are representative companies in defense. Last but not least, once again, I would like to just remind you of the software, computing, and network.
All of these are being engaged by us, but there are very few players like us. It is very rare that you have a single company addressing all these seriously. Including these technologies, we see a great potential, and we have to make sure that this will be a reality. The inference world has a great opportunity, not just for Japan and just for the rest of the world as well. We have spent five years to come up with this, so we would like to pursue business opportunities. As Isobe said, a technology is centered next decade is what we are shooting for. We would like to do business driven by technologies. Thank you.
Thank you. Next presentation is technology-driven business and value-based pricing. We will welcome Mr. Onishi.
Good morning, everyone. I am Onishi, the CRO at Fujitsu. CFO Isobe and CTO Vivek, and following these two members, I want to cover two points, the market demand, and also update on our growth according to the demand. The second point is the value created by AI, and we want to leverage on the value to the growth and in order to acquire such growth, how do we adjust the pricing? Where we are at on this pricing strategy and what is expected to come. That is what I want to cover today. First, Fujitsu growth outperforms the market. I shared a similar slide last year at this IR day.
Showing the IT market growth in Japan from 2024- 2029, it has been set to be at 6.7%. On the left, this is the growth by different industry at the different timing though from 2022 - 2025 or four years growth, the defense in finance and public sector manufacturing are all outperforming these numbers. For the market share, it is either number one, two or three, but in each market we are enjoying a dominant position. I also briefly touched on this in my part. Consulting lead. Using Wayfinders to identify what the challenges are in the business of customers. Once we identify them, how do we solve them? How do we create values? Consultants will be pioneering to developing those solutions. We call this as the orders by consulting lead as these are consulting-led deals.
We mentioned that this is the main purpose of starting this consulting business last year in 2024 and this 2025. Also for 2026, we are looking at to grow by 50% and 100%. We are on a growth track right now. In 2030, we are trying to accomplish JPY 650 billion out of this business. These consulting-led deals and the quality of the deals. For the new deals, they tend to be quite sizable, and they are meeting our plans. Highly profitable businesses deals are obtained. Looking at different industries, let me update the situation. In each industry, regardless of the domestic project or the overseas international project, there are changes in environment and there are the themes which is suitable for these different changes, like IT themes and technology themes.
Beyond different industries, we offer certain solutions or there are some industry specific solutions. Let me use two examples for Uvance for industry. This is for industry specific series are introduced for the last several years. The first one was Uvance for Retail, Uvance for Finance. Those are the first two solutions. For the last few years we have been doubling this business for Uvance for Retail. Sustainable contact center is one example. Flexible Commerce is another example. Those are where we see strong inquiries and growing our results. Uvance for Finance, the sales offices and co-working are focused, those are focused themes to grow a lot. Regardless of industry, we are seeing growth with the sibling platform. Fujitsu [ARO] is what we offer as a service. in 2024, we started to offer this solution, and this actually grew by 5x. It is almost like 500% growth every year.
That is how fast the growth is, and we are receiving a very good, strong customer, the large customer. Of course, I will not be able to share the name, but that is where we see the growth. From a pipelines perspective, these are the pipelines by different teams, how much growth we had in pipelines. It is quite interesting to see that as Vivek mentioned, in physical AI area, this is what we call digital twin. This is intelligent society category. These deals are actually emerging, not in the size-wise, but compared to last year, this is showing a year-on-year growth. With digital twin for the disaster reduction of disasters and actions for the aging society using digital twin to simulate and trying to optimize the population. We are getting a lot of inquiries from public sector for that purpose.
With physical AI, the roadmap for the next generation factory is one example. Utilizing physical AI to accomplish such a roadmap. We are getting a lot of inquiries on that theme. Let me elaborate a little bit more. The first, this is the business deal with MONAKA, the chemical manufacturers. They need to test the durability of the product using AI. As a basis for that, they are using MONAKA and they introduced HPC using MONAKA. Similarly, physical AI, this is for next generation manufacturing plant or PDM or BOP beyond that, or robot integration. Covering all these technologies now we are talking to automotive manufacturer to utilize this physical AI. For the intelligent society, for example, the depopulated municipalities are facing healthcare issues and also talking about trying to collaborate with the MaaS. We run the simulation, then we also optimize the operation.
That is something we started to see in the new mid-term plan, we are trying to invest in those categories. Now we already started to see some specific deals or projects coming up in those areas. I try not to talk about this too aggressively, but from my perspective, the other player system, the competitor systems, how much we have replaced their system is something that I want to talk about. Also we have our customers. From our customers, we also try to protect our deals. This is shown as the wins and losses. This is my important KPI, and I just wanted to share some information on that. For these four different industries, we are competing on the retail, but the other areas, we are showing a lot of wins, both in the yen terms, also the size, the number of deals.
I am proud of these results. Public sector, finance, manufacturing, we are winning compared to the competitors. We are taking over those projects, or we regained the project back from the other players. For the public sector, we gained business, which was held by the other players for over 30 years. In finance, we do cover core mark, but the front line in business where we didn't have much exposure, but we started to win. It's common for those projects is the common thing is how we can utilize technology is actually well-accepted by the customers. The technology is created by Vivek's team is actually utilized by the frontline members, the system engineers and sales and consultants, and they are able to build a storyline to the customers and to convince the customer, and the value was much superior to the competitor's proposal.
That is the reason why we were able to get such great results. Competition is still quite tough. Within this competition, in order to have much stronger position is accomplished by the technology-driven business using AI at its core. Generally speaking, AI trend is coming up right now in 2026. AI has taken a role to be more autonomous, using the AI agent, and that's what we are seeing in 2026. In 2030, AI-driven business management, it will become a common condition, and it will penetrate more. For the coming five years, AI service market. The overall market is growing at 6.7%, but this AI-driven management will grow by 26%, and the market size will go beyond JPY 60 trillion. That's what we are forecasting right now. How do we address such a growing market?
As Isobe mentioned earlier, also Vivek also mentioned here, sovereign orchestration, to be able to cover all the way from a professional service all the way to platform end-to-end services offered. In my position, especially this professional service layer, let me elaborate a little bit more. FDE has become a buzzword now. Takahashi-san's presentation will cover this, but let me also share my perspective as CRO. FDE model. There is a way finder as a partner. What Palantir proposed in the original model was like this, echo, delta, and development. Those are the three elements. On the online conditions, FDE is considered to be just a replacement of the on-site engineer, but the bigger difference was it is seen in this echo.
To be able to understand what the challenges are in customer's business and then collaborate with FDE and using Kozuchi and Takane, which is in the backend. Making sure that it's actually fed back into those backend systems, that is one of the biggest strengths that we can offer. By doing so, how are we changing the profit generation? That is the latter half of the presentation. Value pricing is seen a lot, and that's been actually heard as a buzzword in the market. For the value pricing, what is this value? I think that's coming from three elements. First is speed. We can accomplish this faster. Also, making sure not to get behind by the war. Japan's being considered to be behind in many different areas, but making sure that we are not behind is quite important for Japan.
Second point, this is the value itself, the business contribution, the growing top line, and improving profitability, cost reduction. Directly meet these expectations and to grow by the joint work. Lastly, sustainability. Using a lot of powers and tokens are moving everywhere to increase the cost a lot. That could happen too. But in a long period, in a sustainable manner, to be able to utilize this technology is the third value.
Looking back, Shimazu and myself have been in this business always, but the cost-plus pricing is a methodology. We have set the price, as you can see in this diagram, we start from the cost, the personal cost and infrastructure and platform development cost, materials, manufacturing resources cost, and then direct expenses. Then there is a certain level of profit that is added on top of that. It is not about customer value at all. On the top, you see the man-hour business. Unless you depart from that, IT companies will not see their stock prices go up. Now, I talked about four-layer model, and through this lens, how should we change the pricing? From a bird's-eye view perspective, in the upper layers, like professional services and business applications, this should be outcome-based.
There is pricing based on outcomes generated by our services, and then SaaS or sovereign platform and infrastructure. In these layers, usage-based should be used. Pricing based on service usage value. This is how we look at it. Let me go into each of those in terms of how we promote a pricing model and what are the actual examples that we see in the real world. As for outcome basis, monetizing outcome is what it is about, and profit share is one of them. For example, there is a JV with the customers, and the profits out of that can be distributed between them. Express service, it is about speed. If something took three years and now it takes only two years, and this one year of benefit should be also distributed to us as a value.
Hybrid pricing is after system started to function on a consultative basis. While we are supporting this hardware, if there is a service to provide the operation of the customers, then we would also benefit from the performance-based incentive by providing the benefit to the customers. Use case and PaaS, FDE or agile-type business for this. There is a fixed framework of output, and then there is a flat rate that is applied to bill the customer. In terms of examples, with regard to profit share, on the left you see major logistics company, the co-logistics platform or joint transportation platform is established, and we make money, and the trucks are operated on the platform. Cost reduction would be the benefit, and CO2 emission reduction and stable supply of truck drivers is another side of benefits. On the right, you see success-based fees.
The Japanese companies based in Europe. There are two different types of success-based fees. The two phases, you come up with AI agents with PaaS, and if this is adopted as MVP, then we can get success fees. Then once this MVP works and the full-scale business started, then we can also benefit from the fixed commissions on the revenue that they make. These are the businesses that are already being started up. Another one is a usage-based. In terms of pricing model, there is a per-code-based pricing and data volume-based pricing. Also as for recognizing revenue, there is a tier usage-based pricing. For example, mobile carrier's pricing scheme is something similar. Savings-based pricing, it is related to sustainability. If we contribute to less consumption of resources, then we can benefit from there.
Revenue share could work in this area as well. Let me share with you some examples. As for source, the per-code-based pricing, this is SMBC Nikko Securities, and we have already jointly made a press release. This is about modernization. Legacy source is reading and automatic generation is done. With that, 1/30 is the reduction that we see in the document generation time. On the right, you see the Uvance for Retail solution example. The retail business operators has a store operation and we are giving support with AI agent to the store managers. The sales forecast and inventory procurement and stock shortage and disposal and recovery and weather-related information, these will be all supported with AI agent. Different stores have different sizes. The larger the store is and the more complex the merchandise is, the more complex the forecast would be.
This is the data volume-based pricing. This is a really typical example. In the retail business, we have expertise in this industry. There is a tacit knowledge that we have, but this has been actually incorporated into AI agent for them to be able to support people. The space where AI agent works is what we call data volume, and we monetize that data volume that is used. This is also success-based fees, baseball and entertainment, Hokkaido Nippon-Ham Fighters and NTT Docomo are working with this ticketing solution. For the business, it is different from JV business. Hokkaido Nippon-Ham Fighters and Docomo are the business operators, and they use this Ticket Revolution solution. Then the more the tickets they sell, then the more kickback that we receive. This is a completely B2C model.
On the right, you see the major telecommunication operators and overseas AI service business vendors. It is AI computing broker, and we contribute to resource consumption reduction, then benefit from there as a fee, in the form of fee. Uvance is a central role, and we have seen examples as well. In the modernization and conventional SI, you can automate AI development and source code, but you can have higher quality products at faster base and how you can monetize that part is something that we should consider. There is an AI pricing COE that has been established. But this is not enough, and there is also SE leaders and consultants are involved in the practical team at what level of which customer you are talking to and what you are talking about and what we should be committed to.
And as a result, by increasing that commitment from us, what would be the potential risks that would be increased and how you have to support that through contracts? From quality assurance perspective, how you should approach this? These are all practically reviewed, and the team members will go over to the customer site and start communication and receive feedback. How best and convincing you can proceed is something that they are talking about. Express service is really a fast service, and you benefit from that by providing that value. It used to take three years, and then it takes only two years now. What are the values there? Generally speaking, this here can be used to calculate how much value you can get from there.
But ultimately, when you get to the customer site, then if it is JPY 5 billion that has been saved, then how to allocate and distribute this JPY 5 billion between customer and ourselves is a key challenge. Maybe JPY 3 billion can be received, but JPY 2 billion should be used for ticketing service improvement. It is not the case that nothing works, but there are increasing number of customers that allow us to put this on the negotiating table. Then there are various talks that have started. You have to be logical or scientific sometimes, or is it really about how you proceed with negotiations? You have to accumulate your experiences in order to win actual business and reap the benefit and harvest the benefits. Ultimately, the value pricing can accelerate growth. Let me talk two or three minutes more.
As a target at this moment, the value pricing application rate in the conventional SI monetization Uvance, it is about 20%. Most of them are Uvance, and in 2030, 60%, 2035, 80%, those are the targets for us. The gross profit contribution in 2030, at least 3%, and 2035, 5% and minimum increment. I just would like to emphasize this is the minimum number. As I said earlier, the value pricing can be classified to two different types. The pod type and express types, this is the first order function. You are always pressured to reduce prices, and delivery team has to increase productivity and reduce the cost, and they are pressured to reduce price. The express service is something that we can address that request. Profitability improvement is just a defensive side.
It is not increment to the profits, but you can stop the profits from declining. Of course, if the size is smaller, then there is more projects that you can deal with in a fixed period of time. Also scaling up, this is success fee and also the usage-based billing. This is an offensive one, like ticketing and success-based fees, you can increase this to enhance expected return. What is important is that in a tabloid magazine, you may be skewed toward the defensive approaches, and in a capital market, people are looking at the right-hand side, but you have to promote both in my view. This is my last slide.
So as I have been saying, there are market opportunities right there in front of us, and technology-driven value creation that Vivek talked about is the source for our competitiveness and how you can monetize that and reap the benefits. It is the cost-plus type methodology, which has been from the perspective IT vendor, that is not the way that we should go from now on. We have to look at AI pricing COE and we can open up a very good path forward with this new method. That is all. Thank you.
Thank you. Now we want to move on to the next presentation, Mr. Takahashi. This is to leverage AI transformation to drive the next phase of growth for Uvance.
Thank you very much. I am Takahashi, I am in charge of solutions. So, I want to talk about using AI transformation to drive the next phase of growth for Uvance, and I want to talk about some specific initiatives as well, and a path forward. I will skip my self-introduction here. Let me first look back at progress to date until 2025, as you can see. We accomplished JPY 703 billion in revenue, so we were able to grow the Uvance business to have 30% of the total revenue. Looking at the past few years that we have implemented various initiatives to accomplish the further growth, especially with the partnership Palantir.
We are using data scientists as a forward deployed engineers, and also in December last year, we acquired BrainPad, and they existed for 17 years in Japan, and they were the business to offer data engineering. We acquired this business, and they are now under our group, so we are reinforcing the overall workload. Also in this June, we had a strategic partnership with many different Frontier. Using our own IP and using such a cutting edge AI model, so we can actually combining them to create a new added value. As explained by Onishi, AI transformation market has been changing at such a rapid speed. It has been accelerating, and the recent evolution of AI is quite fast lately. From six months to 12 months, the AI model used to take that long, but now a new model has been out every three weeks.
Because of that, the conventional DX for improving efficiency in business, not just such a DX, but by building the cutting edge AI, so we can actually improve the process, day-to-day business process, and we can transform the business portfolio to transform the organization itself. Using AI to rebuild the business, and that is what has been required in the market. AI transformation, what we call as AX. It is not just the business efficiency improvement, but the business transformation itself. That is what it means. AI transformation. Now, once we execute this transformation, I believe there are several gaps existing. First is the value gap. This is about the AI cost, the token cost. If the value is created enough to meet such a cost, there is a value gap.
At the same time, as we see many Frontier AIs are utilized, when you conduct the testing of your data, especially the know-how and tacit knowledge, when we utilize, if there is any testing, actual data is protected or not, and the customers have some concern about ensuring their GP as well. The third one is a technology gap. AI technology has been evolving so fast, if we have to select the right model to suit that business. Also there is a big workload in order to accommodate the evolution, and it is difficult to accomplish this by a single company. Someone like Fujitsu can support providing customization to offer a various kind of AI model to be suitable for the customer. That is where we can offer the value.
Those are the three gaps that we are trying to overcome, and Fujitsu is able to offer full-stack services. Vivek also mentioned, Onishi mentioned as well, in case of Uvance, the Wayfinders is the service that we can offer, which AI to be utilized. Also through consulting, we have this FDE to actually execute to the application layer. Selecting the most optimal AI model, not just for the cost reason, but also combining multiple AI models to improve the inference accuracy as well. By optimizing the cost but also improving inference accuracy, those are the area where FDE can offer value to the application layer in the middle for AI agent. This is the agent framework, the Guardrail. In order to optimize the token, we do the routing service, and also by combining the optimizing model combination, we improve the accuracy.
We need to, of course, avoid a hallucination and prompt injection. Fujitsu Guardrail can create the new value by combining with the Frontier. These two layers are supported by Sovereign AI Platform. Especially for customers who handle highly sensitive information, we offer Sovereign Environment, and there are many different pairings available. From the Fujitsu's proprietary AI technologies or to the latest Frontier AI models. We can deliver both trusted environment and flexibility to incorporate advances in technology. To deliver such AI Transformation to customers in the most effective way, we believe Uvance must also evolve. Therefore, we are promoting this Uvance for industry. We are utilizing more knowledge in specific industry by leveraging on the context, such as tacit knowledge is utilized. That is the actual sources for the customer's competitive advantage.
We need to make sure to have them learned by AI so that we can result the outcome and the judgment, know-how, those are not in data. They are existing in each individual's knowledge, and such a knowledge to be made into the context so that AI can understand them as well. By making them into context, AI agent based on such a context to be able to make a judgment, also execute the actions on a day-to-day operation. Such different policies dependent on individuals can be actually utilized by the whole organization, and so that way we can scale those outcomes across the organization. Why can we accomplish this? Because for over the years, Fujitsu has accumulated the expertise in developing operating customers' business systems, but also having a deep knowledge of industries and business processes behind those systems.
So we systemize that knowledge as industry context that AI can use rather than limiting it to individual customer engagement. Then we believe we can deliver AI transformation services I described earlier to customers faster and with greater quality. So Uvance traditionally focus on providing broadly applicable offerings centered on social challenges fit to standard. That was the main focus. But now going forward, we will transform Uvance into a model that delivers AI transformation starting from industry-specific context. More than combine the outcomes and insights gained from different industries and expand them towards solving broader cross-industry social challenges and creating new market. So we will make them into the context and trying to create the specific outcome. This is the FDE that we are trying to offer.
As Onishi explained back then, since 2020, Fujitsu's been working with Palantir, who is a pioneer in FDE concept, so we partner with them. Through this collaboration, we learned mindset and ways of working as defined FDEs, and we developed for us. The key point is FDE consulting, working together as mentioned by Onishi, so we run this feedback loop. Then, we will actually feedback this as IP into our technology. Because it's important with the speed and the quality, we can actually offer the most optimal solution to customers. So we are talking about FDE having thousands of FDE in the market, but we actually develop such a quality resources so we can maximize the value offered to customers. That is our main focus. The important point, we are talking about circulating back to the customer.
This FTE, this is not just for the customers in Japan, but this is well appreciated by the global customers. Let me use the next page to explain these examples. This is an example from real estate business, and this is a U.S.-based customer. They are operating 100 different countries around the world. The evaluation is actually the block of all these, and tacit knowledge, like years of buildings and size of that land is not the only information, but they also need to understand the demographics surrounding the property, who are living nearby, what are the communities nearby? Such information is also creating value that contributing a lot to the value of the property. So the spread of the actual market can actually change by 20% to 30% from the standard reference value. Such a tacit knowledge to be understood by the AI first.
The company can utilize the intelligence, so making sure to build such a context, and using them, the customer can accomplish the speed and accuracy of evaluation as the appraisal of real estate. There is an improvement 45% in four weeks. It's not just ending after four weeks. There is additional workload. This is going to be a recurring work. FTE work is actually contracted as a pod size, and then by investing into application, we can make this into recurring business. FDE is not a permanently stationed personnel. They just go into the business of customers, and output has to be made on a permanent basis. This will contribute to revenue increase and cost reduction, and that is something that we have to do as FTE. The second point is about the retail industry area activities that was also mentioned by Mr. Onishi.
Before sharing a specific example, let me talk about the future image of agentic commerce. Ahead of 2030, the leading role in consumer activity will shift from brick-and-mortar stores and e-commerce sites to personal AI agents. Personal AI agents is not just a purchase history, but the individual's lifestyle, health status, family composition, and also future plans included. Taking these into account, you select the optimal products and services for customers. Omni Africa, which is an African vendor, and also we are working with various industry players to create new services. AI agent is about development of services customized for that customer, and this is integrated optimization from production, sales, purchase, and delivery. Manufacturers, logistics, and finance institutions, various AI agents are working with us in real-time basis so that we can provide the best purchase experience to the consumers.
In this agentic commerce, we take advantage of our industry knowledge to make retail AX evolve, and we are hoping to create new demand and value in the world of e-commerce. As a first step for this is the AEON Food Style example. First of all, as you can see here, AI agent and store manager actually talking to each other about store strategy. The order placement and shipment is something that they have to do on day-to-day basis, and they are too busy with that, and they may not be able to spend time for strategic store operation discussion. It is not just efficiency in operation, but how to make it into actual actions is a key. Using multiple AI agents, it is not rationalization, but sophistication has been made a reality. First, order placement.
For example, store has tacit knowledge like local customer base generations and local preference. These are a part of the tacit knowledge of the store manager. By incorporating this into AI, you can make it possible to do automatic order placement. Traffic flow inside the store can be enhanced with the image analysis. What sort of layout should be made to make it easier for customers to purchase? By also detecting stock shortage, you have to make sure that products that the customers would come to buy would be there, and that will increase the actual sales, but customer satisfaction will be also enhanced. In order to enhance this added value, there is AI agent that is there, but behind this there are multiple AI agents that are running. This is one of the examples that I was able to share.
Next is the third example. This is initiative. Uvance has been always aiming for cross-industry initiative, and this is one example, that data and customer attach points in one industry would be also applied in the totally different company to create completely new ecosystem. With cross-industry data and the connection being made safe, you can actually create value for society as a whole and create new services. We are working with SMBC Group and also Uvance . This is SoftBank and SMBC Group. We started from medical data and SMBC Group and SoftBank Group has tens of thousands of consumer data. Data foundation infrastructure that can create new services for consumers, which is what we are aiming for. There are 4,000 medical institutions and 60,000 users. By integrating data, you can optimize medical expenses, which is one of the first benefits.
JPY 50 trillion worth of healthcare expenses can be reduced by about JPY 5 trillion , which will be about efficiency improvement in healthcare expenses. This initiative is really important. It is not just health data, but this can be used for many other industries as well. For example, drug discovery is one of them. Patient's data can be used to avoid drug loss and drug lag, and also financial insurance design and food that is suitable for different customers can be developed. By having infrastructure data platform that is common, we can have a cross-industry value for society as a whole. I have talked about cross-industry examples and industry-specific examples, but now I like to talk about sovereignty. AI transformation is supported by a sovereign platform and cybersecurity.
I talked about this, the AX challenge, and there is a trust gap where people cannot use AI in a safe manner or with peace of mind. We can ensure the sovereignty of the customers as one of the solutions. In this world where AI data is the core of AX, who is managing the infrastructure under which legal system, using which technologies? This is something that customers can get to choose, and that is something that we are coming up with. Fujitsu looked at five different types of sovereignty, data, operation, legal, security and technologies. We are providing architecture that is best for that particular customer, using these types of perspectives. This is domestically produced CPU, MONAKA, which was press released and the server that incorporates them. We can trace back to the source of the parts.
There is a traceability that is being assured and actual data processing will not be visible to external parties. We can protect the confidential information. MONAKA server is the core, and then we can take advantage of Kozuchi and Takane, the proprietary technologies. With the vertical integration, we can provide services. What is important is full stack structure, which was explained by Vivek. There are many customers expressing their interest. Major telecommunication players and All-Japan Federation of Labor Unions and automotive manufacturers are expressing interest, and sovereignty is really important for customers that would like to ensure that. The cybersecurity business that we made announcement last month with the leapfrog evolution of AI, cyber attacks are becoming increasingly and rapidly sophisticated. We was providing individual products and also operation on behalf of the customers and other cybersecurity services by product.
Fujitsu will provide the management cyber defense governance, which integrates management, risk management and defense implementation, operation and improvement. By providing the high degree defense service, we can provide business continuity and corporate value protection. The first one is inherent defense architecture. You are not adding the defense on hindsight, but this has been embedded in the system and its business. On the zero trust concept, if there is an invasion, then dummy file can be provided to the hackers so that there is going to be inherent defense. The other is adaptive defense intelligence. We are using cutting-edge Frontier AI together with our LLM guardrails technology. We would increase the accuracy of defense, and we will keep evolving this, and by doing so, we can make sure the security of the customers.
The data that we gained from security services can be fed back to the top management of the customers so that the defense structure can keep being updated and redesigned. We can provide circular type model so that the recurring service revenue can be produced. The cyber attacks are being advanced on a day-to-day basis, and we are providing the protection from the top management perspective. We have Customer Zero. We have been receiving a large amount of cyber attacks, so using the experience that we have, we can provide services to the customers. I've been talking very long, but this is the future outlook. What we're aiming for is not just a simple increase of sales, we would like to change the business structure and to pursue the quality of the growth.
By FY 2030, more than JPY 1.7 trillion in revenue for Uvance is what we are aiming for, and the recurring rate of 70% will be something that we're targeting at. Forward deployed engineers will be permanently stationed in the customer sites so that AI platform will be implemented and highly profitable services like sovereignty and security will be pursued and tacit knowledge using industry context or AX using that, those high value-added services can be used to increase recurring business and gross margin. AI is reshaping the world and as we witness that, we are committed to the realization and transformation, and beyond that, we like to create values for the society as a whole. Thank you.
Next, delivery transformation from Ms. Shimazu.
Good morning everyone. I am looking at service delivery at Fujitsu. My name is Shimazu. Let me talk about delivery transformation. For the last few years, every year, we have been improving our gross margin by 2 percentage points, roughly JPY 100 billion, and we've been improving, increasing every year this much. We want to continue making this 2% improvement every year. That is what we are trying to accomplish even beyond now. I want to talk about how to accomplish that. The latter half of my presentation will be the modernization. I have been updating modernization every year at this IR Day, so I want to share the latest situation. AI-driven delivery. Service solution, not just modernization and traditional IT service, there's also Uvance as service solution. That's what we cover under service solution.
The important thing is it's not just a cost improvement when we talk about AI-driven delivery. As Mr. Onishi mentioned, we also want to offer the speed to the customer, and that is what Fujitsu is trying to offer. For the last few years, we have been growing the margin by 2% every year. How did we accomplish this? First, the domestic engineers, there are about 16 system engineering companies. They were now consolidated under Fujitsu Japan and offshore for the last three years even comparing before that to before the three years, we have doubled. We have increased the use of offshore to be worth 100,000 man- month. We implement the automation and standardization and in addition, the value pricing explained by Mr. Onishi and AI-driven delivery will be further promoted. With this, we intend to improve our gross margin by 2% every year.
What is AI-driven delivery? As you see at the very top, this is conventional development. When you develop system, you define requirements, design, development, and testing. Humans were leading these processes. We actually manage at three stages to accomplish AI-driven delivery. For level one, I think people are using a lot of AI, using prompt to obtain various information. For the level one, this will be still prompt-based conversation with AI to develop system. From level two onwards, AI agent will start playing a role. In level two, AI agent specific to each process exist, and they execute their tasks. Between different processes, there are people involved to do the confirmation to move forward to the next process.
For level three, in between those processes, AI agent will actually be engaged to run as an AI orchestrators, and engineer will be involved at the beginning when inputs are needed to the AI, and the final confirmation at the end. Those are the three stages we pursue to utilize AI more. If this is really feasible or not. We have a strength in the fact that for over the years we were engaged in building systems at customers, so we have that knowledge based on them. We also have a managing capability to run the development project. First, we can have all this knowledge to be learned by AI. Secondly, there are experts in AI, also qualified members for AI. We have a lot of them. We will use them in many different projects.
We have our own LLM, large language models, meaning as we develop the system, if we find we need something like this, then we talk to Vivek, then that will come back immediately. They will be able to reflect such a request. It is not just our own Takane, but we also can utilize Frontier AI. We can have a strategic partnership with them. As we start using AI, there has been also the increase in token cost. We started to realize that cost increase. In order to control this token cost, there are also know-how, and we also have technology to allow that to happen. In our international market, there are businesses who offer such a service, but we can actually cover that in-house by ourselves, and that is another strength we have. We have already started to implement AI-driven deliveries.
We have the agent for each specific process, and that is being utilized by a manufacturer A. This is a project worth a few billions of yen. The designing and development phases, they utilized agents. The overall schedule wise was reduced by 20%, but by using these agents, actually, they expect to reduce by 30% instead of 20%. For a company B, their financial business, they also use agents for designing and developments, and they actually reduce their time by 50%. This is roughly about 20% for the overall process, but the certain process they have accomplished reducing the time by 50%. In level three, this is quite high. It is a high hurdle, a very high difficulty. We made an announcement in February. We have this our own package for local municipalities and the medical package.
There are legal changes taking place every year, and then we're going to have to incorporate them, reflect them into packages. It took three months in the past to implement those changes. But with the use of multi-agent AI, now it's reduced down to four hours from three months. We do have such an experience already being implemented. The future of AI-driven delivery. We already have expanded the scope of AI-driven delivery. Other than the project requested by customers, most project. Well, of course, there are difference in levels, but we have already expanded the scope of AI-driven delivery. Now we are moving on to next phase of exploration, going deeper to increase more level twos and level threes.
Ideally, we want to accomplish 100% a level three, but first in 2030, this will be the breakdown we expect to accomplish with the AI-driven deliveries. In express service, there is also a request of cost reduction from customers. We do offer values, but the selling price must come down. When it comes to engineers, maybe we use 100 engineers, but we only need 40 or 60 engineers. In the past, sometimes we had to reject some of the project because of shortage of engineers, and there were about a few tens of billions of yen worth of those lost opportunities. But now we are able to take those businesses to make sure that we are not losing the revenue. That is accomplished by AI-driven delivery. Next, modernization. We are trying to support customers to moving away from legacy assets.
In 2022, we intend to complete the mainframe business and UNIX server business, basically maintenance to be completed in 2035. We have announced that. We started modernization as a business since then. Not just us, but also other players are now in a modernization business in the market. Last year, we were number two. It was a very small difference. But this year we are trying to be number 1 in market share in Japan. The market environment for modernization is seeing a CAGR growth of 10.2%, and for the large and mid-size businesses, roughly 80% of those customers still using legacy systems are mainframes and UNIX servers. Currently, there are 250 customers using mainframes. UNIX users, there are 500 users still using UNIX. They need to modernize their system. Once again, the numbers.
In 2023, we started modernization business, and last year, we made a business worth JPY 392 billion. We are achieving a large growth every year. This year, again, another big growth to accomplish JPY 470 billion. That's what we are trying to accomplish. Of course, the gross margin will be improved. We have a modernization version for AI-driven delivery. We intend to improve by two percentage points in gross margin. The future outlook for modernization. The modernization of our legacy system, in 2028, it's going to be the peak to implement modernization, then it will start peaking out. The knowledge and know-how that we gain from modernization of our legacy systems, and it's also true that there are still 80% of customers using legacy system.
So we will utilize that knowledge to address those customers, and this will grow the business to be JPY 700 billion in 2030, with market share going over 30%. That is what we are trying to accomplish with modernization business. Even beyond that, customers will be using more AI in their business. So there is the modernization to shift AI use more towards the foundation of the customer's business. That is going to be the pillar towards 2035 as a modernization. So we will evolve our modernization business furthermore. This concludes my presentation. Thank you very much for your attention.
All right. The last one is towards sustainable enhancement of corporate value. Mr. Isobe, once again, please.
Thank you very much for staying with us. We are running over the scheduled time significantly. Each Corporate Vice President, once they start talking about their own topics, they keep talking, and there is a lot more that they had talked than scheduled. If this has been conveyed to you, well, then that would have been better, but there will be always alphabets and keywords and abbreviations that they use. To what extent we are able to make ourselves understood is a question and concern, but there will be various opportunities to have communication with you. I am hoping that those corporate vice presidents can be getting involved in the future opportunities as well. We have run over the scheduled time. We could speak my part, but that is not something that we should do.
We have to sum up and conclude in some way or other for those presentations that are made. How all these presentations are connected to each other. Ultimately, each strategy will contribute to cash generation, and with the allocation of cash gained, Management Vision 2035 financial KPI achievement is what we see. I will skip this over. This has been uploaded already, and this is how each of the corporate vice presidents made presentations. Here, this is what was explained at each section. It may have been difficult for you to see, but most of our major quantitative targets are sorted out here. They are different color coding. Blue part is more or less the revenue increase, and pink is about profitability improvement. The big box on the top is a leading innovation for AI sovereignty, Mahajan explained.
AI core new business creation area, at JPY 3 trillion or over in turnover revenue is what we are targeting at in new business areas. The service solution business evolution, Onishi, Takahashi, Shimazu explained about these, and quantitative numbers are listed up here. The market demand, which is the basis, remains strong. With that background, Uvance is likely to do continued growth, and modernization will increase a share with accelerated basis, and AI-driven delivery will also speed up our service delivery and productivity improvements. What we can achieve as a result of accumulation of these is on the next slide. As for the revenue increase, what was represented in blue is summarized here. In the large box is service solution business expansion part. Domestic IT market is increasing and Uvance and modernization steady growth will give us 6%-8% of growth in 2030.
The cash that we gain from these businesses can be allocated to investments in new business areas, Kozuchi and Physical AI, MONAKA, and quantum computing. The new business areas can achieve a JPY 3 trillion worth of sales that we are targeting at. For these new business areas, most of the driving force will be realized after 2030, but as Mahajan said, in the MONAKA market, there will be some sales recognized in next fiscal year. About tens of billions of yen worth of sales can be generated from there. Together, the total revenue will be by 2035, 6%-9% CAGR is what we are targeting at. This is about productivity or profitability improvements. Pink color-coded one is summarized here. Speed and quality and business contribution, what customers wanted, that has to be transformed into pricing.
With AI-driven delivery, the sustained gross margin improvement of about 2% or more per year is what we are aiming for. Together with revenue growth and productivity improvements, this is the target. The revenue CAGR of 6%-9% through 2035, and through sustained productivity improvements, we aim to reach an operating margin of 25%-30% by 2035. These are the aims that we have. In terms of revenue scale, we aim to double from JPY 3.5 trillion in the last fiscal year to double the amount, around JPY 7 trillion-JPY 8 trillion by fiscal 2035. In operating profit, we aim to increase from JPY 0.4 trillion in fiscal 2025 to over JPY 2 trillion by fiscal 2035. On a personal note, if we can say JPY 3 trillion and JPY 10 trillion, that would be better, but these are the numbers that we have in mind.
Together with profit increase, we also would like to expand the core free cash flow to 4x- 5 x the level of FY 2025 with the profit growth and capital efficiency improvement. With the cash flow increase, allocation or the pool available for allocation grows larger. The left graph shows the base cash flow prior to executing growth investments, what we call base cash flow. Actual base cash flow for the three-year period of the previous midterm plan totaled, on a cumulative basis, JPY 1.3 trillion. In 2025, on a single year basis, JPY 0.7 trillion, and this includes these asset recycling of Shinko Electric and other one-time cash inflows. Excluding this, it would have been JPY 0.4 trillion. In fiscal 2035, we are expanding our cash generation capacity to 4x-5x on a single year basis to JPY 1.5 trillion- JPY 2 trillion.
The base cash flow, the source of funds for allocation, will total JPY 10 trillion over the next 10 years. We plan to allocate this JPY 10 trillion, as shown on the right, JPY 6 trillion for growth investment and JPY 4 trillion for shareholder returns. Half of the growth investments, in other words, JPY 3 trillion, will be directed toward expanding our service solution business like Uvance and Modernization. In FY 2025, this amounted to JPY 200 billion for the single fiscal year, but we plan to expand this to about 1.5 x more. The remaining JPY 3 trillion will be invested in creation of new technology powered businesses. Obviously, both are the same, but to accelerate our growth, we are open to applying leverage at the right scale and the right timing. Shareholder returns will total JPY 4 trillion. There are no major changes to our policy.
We will maintain our policy and total return ratio of 60% as a general guideline, and we aim to expand returns in line with profit growth. There are no significant changes to our approach to dividends and share buybacks. To sustainably grow corporate value, it is essential to maintain a balanced approach between two key pillars, advancing proactive growth investments and expanding shareholder returns, which must be guided by not just profit distribution, but also by capital efficiency. We will thoroughly execute the investments necessary to achieve business growth, as explained by corporate vice president today. Naturally, we will clearly demonstrate the resulting growth through our annual results. At the same time, we will steadily expand shareholder returns.
It is precisely this balance that will lead to the achievement of the financial KPIs set forth in our Management Vision 2035, and EPS CAGR exceeding 15% and ROE exceeding 20%. We have explained the various strategies covering a 10-year period. We have looked at this 10-year period for the vision. With the rapid changes of recent times, not just 10 years from now, but even one or three years from now, the landscape is likely to differ significantly from our current projections. However, it is precisely these changes that represent growth opportunities for Fujitsu. That is why it is extremely important to set out a vision for 10 years from now to continue steadily demonstrating our progress and results each year. At the same time, to remain flexible and adapt ourselves in response to various changes.
As the culmination of all this, I would like to conclude today's presentation by promising to achieve financial KPIs of an EPS CAGR exceeding 15% and ROE exceeding 20%. We are making this promise of these financial KPIs. Thank you very much for your attention today.