Ladies and gentlemen, please put your hands together for Jack Huynh. Good morning, IFA. I am so happy to see so many friends and partners in this wonderful city in Berlin. That was just the beginning. Now let me show you what comes next. For decades, we have made computers faster, smaller, more powerful. We put computing on our desk and then in our pockets, and now we are putting intelligence everywhere. Something bigger is happening. For the first time, we are not asking computers to do more. We are giving them an ability to understand what we are doing, what we are trying to accomplish, what matters to us, and ultimately, to work with us. That changes everything. At AMD, we build a compute behind that possibility. From the smallest device to the world's largest supercomputers, we have always taken the hardest problems in computing, and we always believe something simple: Technology matters most when it puts more capability in people's hands. Today, I want to show you what happens when that capability becomes personal. Some ambitions take us beyond Earth. Our technology helped power Artemis II, including critical systems inside the Orion spacecraft that carried astronauts around the Moon for the first time in more than 50 years. Think about that. Humanity going further than it has gone in generations. Others reaching the most personal moments of our lives. Our technology powers Intuitive robotic-assisted surgery systems, giving doctors greater control through less invasive procedures. Imagine trying to move your hand just 1 mm. Intuitive can translate 3-mm movement into 1-mm while filtering out natural hand tremor. 3 mm becomes 1 mm. In the operating room, that precision can change what is possible, and some are as big as the planet itself. In Finland, LUMI puts the power of roughly 1.5 million laptops in the hand of researchers studying our climate and using AI to detect cancer earlier. It runs on renewable energy, and even the heat it generates is captured to warm hundreds of homes nearby. From the edge of space to the operating room to the future of our planet, AMD technology powers all of it. These are extraordinary challenges, but the same technology helping us explore further, heal better, and understand our world is also woven into the moments that shape our everyday lives because computing does not just change what machines can do, it changes what people can do. Every time you open your favorite social media app to catch up with friends, discover something new or simply share a moment, AMD EPYC powers the platforms behind the experience. Every day our technology helps connect billions of people around the world. We also power the consoles and handhelds where billions of people around the world come to play, to explore and discover new worlds. You saw a glimpse at that in the opening film of one of my favorite games, Marvel's Spider-Man 2, on the PlayStation 5 Pro. That's what powerful compute can do. It's not just the silicon. Our software helps developers push their imagination further, creating worlds that were once impossible to build because technology should never limit imagination. It should expand it. When you press play on a story you love, AMD EPYC and Threadripper power that technology that brings that experience to your screen. We're behind more of the moments that shape your day than you might even realize. I spent almost three decades in this industry. I've watched computing transform our lives in ways I never could have imagined. We made processors faster. We made devices smaller. We put computing into people's hands. I feel incredibly lucky to have been part of that journey. I've never seen anything like what's happening right now, because this feels completely different. For the first time, we're not just making computers more powerful, we're giving them intelligence; intelligence that understands us, what we're doing, what we're trying to accomplish, what matters to us. The computer stops waiting for you to tell it what to do. It starts working with you. We're only beginning to understand what this could mean. What happens when the most powerful tool we ever created starts working with us? AI is the new electricity. A century ago, electricity changed the world by making power available everywhere. It transformed how we worked, how we lived, how we moved. Eventually, we stopped thinking about electricity. We simply expected it to be there. AI is heading in the same direction. If electricity gave the world power, AI gives it intelligence. Just as electricity became woven to everyday life, intelligence will become woven into the tools we use every single day. When intelligent becomes that fundamental, it changes what all of us can do. We're already seeing it. AI is moving into every part of our lives. It's helping scientists discover new medicines, helping doctors see what they couldn't see before, helping creators bring their ideas to life, helping all this do things that once seemed impossible. This isn't the distant future. It's happening right now. We're only beginning to see what's possible. When you give people more powerful tools, you don't just change what they can do, you change what they believe is possible. This is what AI is beginning to do, is rewriting the limits of what's possible. If you want to create, it could turn imagination into something real. If you want to learn, it can meet you where you are and help you go further. If you want to solve, it can help you see what you couldn't see before. Every generation leaves the next with a powerful set of tools. The PC put computing in the hands of millions. The internet put the world's knowledge at our fingertips. The smartphone put computing in our pockets. And now AI is giving computing intelligence. It's no longer what can computers do? It's what can we do with them? AI doesn't just expand what computers can do, it expands what we can do. Now imagine one entrepreneur with one idea. In the past, turning that idea into a company meant assembling a very large team across design, engineering, research, marketing, and sales. Sometimes hundreds of people, even thousands, just to turn an idea into reality. AI changes that equation. One person can now orchestrate a digital team at a scale that was once reserved for the largest companies. This isn't about replacing people, it's about expanding who gets to build. Because somewhere right now, there's someone with an idea that never made it past the first conversation. They didn't have the team. They didn't have the money. Didn't have the resources. Now, they have a new kind of leverage. Their imagination can go further. Their ambition can go faster. AI doesn't diminish human ambition, it gives ambition leverage. This is why I believe we're entering a new golden age of opportunity. AI can reduce the distance between an idea and a result. It can expand who can learn, build, create. The next great company may begin with one person, one computer, and one idea ambitious enough to change the world. Look at how quickly we got here. AI is advancing at a pace we have rarely seen in human history. Just a few years ago, AI became conversational. Millions of people could simply talk to a model and get an answer. Just one year later, AI learned to see, hear, understand the world in ways that felt much closer to how we experience the world. Then AI learned to reason. It could break down complex problems, connect the dots, and work its way towards an answer. For the first time, it began to feel less like a tool retrieving information, and more like a partner helping to solve the problem. Then, in 2025, AI became agentic. It stopped being assistant simply answers, become assist that can act. An agent can plan, use tools, check its work, and keep going. That changed the scale of what one person can direct. That complete changes the multiplier. One agent can help one person do more. A team of agents can work in parallel. One can research, one can write, one can code, another can test. The multiply keeps growing. Give them more skills, more tools, more capabilities, and the work just compounds. AI is no longer just making one task faster. It's multiplying what one person can accomplish. It gives ambition a whole new level of leverage. That new lever comes with a cost. Every productivity multiplier comes with a compute multiplier. More agent means more reasoning, more context, more models, more actions. All of it requires more compute, and hence, more tokens. The more useful AI becomes, the more compute it demands behind the scenes. That is why Agentic AI is accelerating token consumption at an astonishing pace. We are moving from short interactions to AI systems that can think, act, and keep working. Think about a standard chat with ChatGPT that we use every day. You ask a question, it gives you an answer. That exchange might use a few hundred to a few thousand tokens, then it stops. That is the model most of us know today. An autonomous agent is very different. Now imagine an autonomous agent. You say, "Plan a family vacation to Italy." It checks your Google Calendar, finds flights that work for everyone, finds the right hotel, builds a plan around what your family loves. Then you say, "Let's spend two more days in Rome." It changes the flight, adjusts the hotel, and keeps going till the trip is ready. That's a completely different kind of workload. Instead of a few thousand tokens, hundreds of thousands, millions. Now imagine all that happening all day. A single user can consume millions of tokens a day. Now imagine that across millions of users. That's an entirely different scale of computing. It won't be the same for everyone. For a consumer, AI might manage your travel, your shopping, your schedule, your messages, your everyday life. Millions of tokens a day. For a creator, it might research, create, edit, and produce tens of millions. For an entrepreneur developer, you might code, test, analyze, research, and keep working around the clock 50 million tokens or more. This is the new scale of AI, and we can already see that curve. In just 1 year, monthly AI token processing jumped from about 0.7 quadrillion- 1.7 quadrillion. But that was before agents. Once AI starts working continuously, the curve changes dramatically. By 2030, estimates reach 120 quadrillion AI tokens processed every single month. To compare that future demand with compute we have today, the gap is enormous. We will need much more compute, and we also need to rethink where that compute lives and how efficiently we use it. Because the future of AI isn't just about more compute, it's about better compute. The economics are really under pressure. 93% of companies are exceeding their AI budgets, and agents will only increase demand. We can't just spend more. We have to compute smarter. Here is one simple illustration. At 15 million output tokens per day, the cloud cost shown here is about EUR 300 a day. That's nearly EUR 100,000 a year for one active user. Fundamentally, another curve is moving in the opposite direction. Models are getting smaller and more capable. Personal devices are getting much more powerful. Capabilities that once required a data center are now within reach of a personal device. That changes the architecture of AI, and it changes the economics. That is the idea at the heart of what comes next. The next era of computing will be defined by intelligence that's closer to you, more personal, more capable, more powerful, built around you. We are entering the era of personal AI. Personal AI understands your context. It stays close to your data and expands what you can imagine achieve. It isn't only a chatbot that you visit, it becomes part of computing experience that travels with you. Personal AI comes down to three things: local compute, more intelligence running directly on your device; privacy and control, your most personal data stays yours, and you decide what gets shared and where; personal context, AI understands your goals, preferences, and the way you work. Bring these together, and the PC becomes something fundamentally more personal. Let's start with local compute, and start with our current generation of Gorgon processors. To understand the pace of change, look at the models themselves. Last August, OpenAI released gpt-oss-120b, a 120 billion parameter open-weight model. It scored 80 on GPQA, a demanding graduate-level reasoning benchmark. That was impressive. Just months later, something remarkable happened. Then came Qwen2.5. A model with just 9 billion parameters scored even higher on the same benchmark. Better performance, 13x fewer parameters. That is the trend that matters. AI is getting smaller, faster, and much more capable, and that efficiency is coming directly to the PC. Ryzen AI can run models locally up to 24 billion parameters. Work that once required a cloud subscription can now happen right on your PC. Model size isn't the only constraint. Memory is becoming the bottleneck. The more capable the model, the more memory it needs to run effectively. That is what led us to build Strix Halo. With 120 GB of unified memory, Strix Halo can run models as large as 200 billion parameters locally right on your device. AI models that once required the cloud can now run on your desk and be carried in your backpack. Local doesn't have to mean second best. Claude Sonnet 5 scores 70.3 on a software engineering benchmark, but Llama 3.1, an open-source model running entirely on Strix Halo, actually scores even higher. A model running on your own PC can outperform a leading cloud model on this benchmark. No cloud, no per-token costs. Frontier class AI is moving onto the PC. Look at the economics. 10 million output tokens on Sonnet 5 costs roughly EUR 90. On your PC there's no per-token cloud charge. You own a compute, use it once. Use it a thousand times. The more AI you use, the more that difference matters. This is where it starts to feel different. On Strix Halo, you could run 10 AI agents locally working in parallel. Your laptop is no longer just a tool you operate, it's a team that works with you. That changes what a PC can become. Smaller models are becoming incredibly capable, but so are larger models. That's what led us to push the boundary further with Ryzen AI Halo. We increased unified memory from 128 GB- 192 GB, increasing the ability to run models from 200- 300 billion parameters. Once again, local does not mean second best. Claude Fable 5 is a leading cloud model. GLM-5.3-Flash running locally on Ryzen AI Halo delivers better performance on this benchmark. A 320-billion parameter running on your PC, no cloud, no per-token charge. That is the power of local AI and the economics move with it. 10 million output tokens on a cloud model shown here is about EUR 500. At high volume that's not small optimization. It changes what is practical to build. Behind me is our Ryzen AI Halo box assist system we built specifically for developers to explore this new class of local AI. I'm super excited to make it available in Europe soon. Our OEM partners are also super excited about Ryzen AI Halo as well. Yesterday Lenovo introduced the ThinkCentre X here in Berlin powered by Ryzen AI Halo. Small enough for your desk, powerful enough to run 300 billion parameter models locally. This is just the beginning. We're going to see an entire new generation of AI systems built around this capability, and we're just getting started. We didn't stop at the desktop. I want to share something very special with you for the first time here in Berlin. Powered by Ryzen AI Halo, we have created an entirely new class of laptop built for the era of personal AI. Codenamed Sundance, years in the making, this is the next generation HP ZBook. 190 GB of unified memory. What if I told you I can create an entirely new world locally on this machine? Let's have some fun. Press play. Seeing is believing. Look at how beautiful this laptop is. An entire 3D world created from a single prompt, all running locally on this beautiful machine. Take an idea like building a world real time and build it wherever you are. This is personal AI in your hands. One of my favorite games, Microsoft Flight Simulator, playing flawlessly on this laptop. Create worlds during the day, game at night. Real world in one hand, a world I imagine the other. This is the future of Personal AI and we're just getting started. The models are ready, the computer's ready and the most powerful open weight AI can now run where you need it. From the cloud to the data center to the device in front of you. The frontier is no longer somewhere else. It's coming to you. But keeping alone is not enough. Personal AI also has to earn our trust. That brings us to privacy and control. Personal AI will touch the most valued parts of your life: your files, your conversations, your preferences. The closer intelligence gets to us, the more important it becomes to protect what makes the intelligence personal because the more AI knows about you, the more trust matters. Your context should belong to you, your data, your choice. Personal AI should never come at the cost of privacy. Personal AI should be private when it needs to be. Sensitive work staying on your device close to you under your control. It should be portable when you want it to be. Your context should move with you across the device and experience that you choose and should be shared only when you choose. You should decide what leaves the device, where it goes and why. Today most AI starts at a cloud. But imagine if intelligence started with you. Your PC handles what it can locally and when you need more scale it reaches into the cloud. You don't have to decide whether work runs; your PC does. That's the shift. Intelligence starts with you and extends to the cloud when you need it. That brings us to the final part of personal AI, personal context. This is the part that makes intelligence truly personal. The most powerful AI won't simply know more about the world, it will understand more about you. Because AI becomes truly useful when understanding not just the question but the person asking it. Computing already has a familiar stack: applications, models, compute. But personal AI needs one more layer, context, the layer that tells AI who you are, what matters to you, and what you're trying to accomplish. Without context, AI just gives an answer. With context, it gives the right answer for you. Let me give you a simple example. Two people ask AI the same question: "Help me prepare for tomorrow." For one person, tomorrow means a customer meeting and three decisions. For another, it means an exam and a study group. Same question, completely different need. That's the power of personal context. AI doesn't just understand the question, it understands what the question means to you. We talked a lot today about the incredible pace of innovation, and I believe the PC will be the heart of what comes next. AMD and Microsoft have been building together for decades, and I've never been more excited about what we are creating together. We share a joint vision of making AI more personal, more accessible, and more useful for everyone. I'm super excited to welcome my great friend and partner, Pavan Davuluri, Executive Vice President of Windows and Devices at Microsoft. Jack. Pavan. How are you? Great to have you. Good to see you. Good to see you. Thanks for being here in Berlin. How you doing? Of course. I'm well, thank you. It's great to be here. Thank you, Jack. Pavan, you and I travel the world so much, and AI is moving at a pace that we couldn't even have imagined. Absolutely. What excites you the most? What are you seeing that fundamentally changes the PC? For us, when we think about the space we're in, the shift that is happening now is technology is more capable than ever before, helping us act on new ideas, take on new challenges. AI is evolving the Windows PC from a tool you use to a partner that can help us achieve more. We're seeing rapid innovation, as you pointed out across the entire stack, silicon models, the OS, and the application layer with each advancing across the board. As you noted, agents, I think, are the next part of this evolution. They turn intent into action, whether you're working with a team or as an individual. With your permission, agents can work across your apps, your services, your files to help you get things done. Now for us, for the PC, instead of just executing commands, the PC will understand workflows and just help people get done what matters most to them. For us, I think that really only works if it's built on the foundation of trust, security, and user control, and I think that's the role that Windows plays for us. Yeah, Pavan, we completely agree. I mean, Pavan, you and I worked there for almost 20 years. That's right. I'm blanking. You know, Strix Halo is my baby. Yes. Right? And we brought so much compute, so much unified memory that you asked for. Yeah. What are you most excited about that we're building together for the next generation of agentic experiences? First, let me start by saying, Jack, we are deeply grateful for the partnership with AMD. Thank you, Pavan. From advancing Windows PC to gaming consoles to the cloud, we've accomplished some ambitious goals together with our deep co-engineering. You know, Jack, for me, it feels like just yesterday we were kicking off Strix and Strix Halo. Yes. Thinking about memory bandwidth on the memory controller- Yes -and getting the best image pipeline for camera. Yep. Yeah. Getting a great NPU for performance. Yes. Yes. It was actually even code named MS1 at the time. That's right. Yes. Here we are- Yes -celebrating the work on Kraken Halo. Yes. It brings some impressive capabilities like the 16 Zen 5 CPU cores, a highly powerful RDNA GPU, a very efficient NPU, and the 192 gigs of memory that you described. With the work on the Windows team for bringing unified memory access and workload scheduling, you can now run larger models than ever before and run really sophisticated AI workloads locally. Here's what that looks like. This is a Qwen 3.8, 125 billion parameter model running locally on Kraken Halo, fully leveraging the GPU. The performance you see here just reflects the incredible work our teams have done together, Jack. This is a model they just released last week, I think. So it's super impressive to see the performance of the system. Oh, do you guys know Pavan pushed me so hard on Strix and Kraken Halo. More CPU, more GPU, more AI- I made for it. -better performance-per-watt, better security. Correct We always deliver. We love working with you, Pavan. As these systems become just much more capable, how does Microsoft make it just so much easier for developers to create net new experience on Windows? It's a great point. Yes. I think for us, breakthroughs in computing really become real through experiences developers create. Our goal in Windows is have Windows be the best platform for developers. At Build earlier, we shared how Windows is becoming an open platform for secure agentic experiences. For us, agents run in a secure isolated boundary with Microsoft Execution Containers so that users, their data, and their devices are protected. Now we're helping developers start building even faster. We're introducing Project Zenith, a ready-to-code Windows experience on developer class hardware. Developers get Visual Studio Code, WSL, GitHub Copilot CLI, PowerShell, all preinstalled and with a curated set of Windows settings optimized for coding on high-end devices with a minimum of 64 GB of UMA. This ensures that developers have enough memory, enough performance to run large models locally. I'm really excited to share Project Zenith will become available on AMD Ryzen AI Halo out of the box. Pavan, I'm so excited for Project Zenith. So excited we'll be first to have the experience on our Ryzen AI Halo boxes. This is a complete game changer. Pavan, we talk about how we can bring hardware, software, Yeah OS at the very beginning of development. That's how we change an industry. Things that once required a cloud, can now run local on a PC as you just saw with- Right -the latest Qwen model. Pavan, what else do you have for us? Show us what's next. Yeah. Let me show you- Yes -what agents using local Ryzen AI performance can do and how they take on more increasingly complex workloads. It starts with a simple prompt like asking agent to write a function like you saw earlier in the benchmark. From there, agents could start performing more demanding tasks all locally, like a forensic analysis for a sensitive set of files and data, for instance. Going as far, for example, to create an entire designer portfolio on the fly for creators, also possible locally. Both examples for me highlight the spectrum between creating developer code that runs on agentic basis and being able to build websites. To me that's just a simple example Jack, of starting with a prompt. Yes. 125 billion parameter model running on Ryzen AI Halo locally, moving from a simple prompt to a complex analysis and creating new content. The quality's very good too. I mean, that website. Blazing fast. Very good. Super fast. Amazing example. To me it's not just about making the PC faster, it's creating this net new experience that you and I have always talked about. Yes. We're fundamentally changing what a PC can do. As AI becomes more personal, what do you see the future of the PC? That's such a lovely question. We think about that at Microsoft a lot. For us, we see AI becoming the scaffolding for human potential, expanding what is possible, what people can achieve and imagine themselves. We believe the PC will become the home for unmetered intelligence grounded in your context, like you mentioned, with your personal understanding under your direction. Local compute will deliver more personal, more responsive, more private experiences working with the cloud seamlessly. As agents become more capable, trust and security will continue to remain that deep foundation. That co-design engineering we talked about is bringing this vision on Ryzen today to life and together with AMD Silicon and Microsoft software, we're shaping the next generation of personal computing. Pavan, I love it. The opportunity in front of us is enormous. Absolutely. It's a once in a generation opportunity. AI's moving so fast and when we move into the PC that responds to you, to a PC that understands you, help, gets things done. That future takes all of us working together. That's right. From silicon to software to OS application. Pavan, I'm grateful for partnership with you within 20 years. I'm looking forward to another 20 years. Thank you so much. Thank you very much. We're really excited. Let's keep building together Pavan. Of course. Take care. Congratulations. Thank you, Pavan. We talked about how AI is changing the PC here with Pavan, but AI has changed something even bigger: What we expect Compute to be. What happens when one system becomes whatever the work requires? What if a PC didn't have to only be a PC? What if a system can give one developer extraordinary compute and become shared AI infrastructure when the team needs it? What if we could bring the kind of compute associated with the modern data center right beside you? We asked ourselves what should our system look like? We actually designed it and then we built it. I have something very special to share with you here today for the first time. A new class of system that I believe will redefine personal AI for years to come. Thank you Peter, so much. Today here at IFA, we are witnessing the birth of something entirely new. This is the most powerful workstation in the world, designed and engineered for completely new era computing, capable of running AI models with more than 1 trillion parameters and we call it the Threadripper Halo Station. Look how beautiful this is. 96 cores on Threadripper PRO, two MI350P Instinct accelerators with the path to four. Up to 2 TB of system memory and up to 576 GB of HBM 3E. All liquid cooled because we want all that power to stay super quiet. This is about as close as you can get to a personal super computer. Peter, you can ship that back to my house. We started with Personal AI. We brought extraordinary compute to the desk with Strix Halo, and we just brought supercomputer class compute to the workstation. Now we need to make that compute open, accessible and ready for the world. That takes software, that takes an ecosystem and one of our key partners in making that happen is SUSE. Please join me in welcoming Dirk-Peter van Leeuwen, SUSE CEO to the stage. Dirk. Hi, Jack. Good to see you. Great to see you. How are you doing? Very well. Thank you. Thanks for joining me here in Berlin. Thanks for having me. How did you like that supercomputer workstation? Amazing. Congratulations. Thank you, Dirk, so much. Dirk, you just saw us make the PC and workstation much more personal. How do we scale it? That is a great question, but that is also exactly the opportunity. AI has made it incredibly easy to build something very, very impressive. The challenge is taking that idea from one machine to something that an entire organization can depend on. Development and production have traditionally been two very different worlds. Developers want freedom. They want to experiment. They want to choose their tools, they want to test new models, and they move very, very quickly. But an application becomes an important business requirement and it needs to be secure. It needs to be reliable, need to be manageable, and of course, if you run it in business critical environments, it needs to be ready to scale. That is why and where the combination of AMD and SUSE become so powerful. A developer can start on Ryzen AI Halo. They can build and test locally, on the right workstation right in front of them. Technologies like SUSE AI Factory and SUSE Rancher provide a path into a secure supported production environment. Then as demand grows, they do not have to throw away what they built, but they can add systems, they can add compute, they can scale the application. So it becomes a journey, and the journey becomes one solution, it becomes production, it becomes many systems, and then the application grows with the business. No, Dirk, it makes perfect sense. The workstation isn't the destination, it's just the starting point. That's an important shift. Absolutely. AMD gives developers extraordinary compute at the point where the ideas begin. SUSE helps turn those ideas into something that can operate at scale. That's the bridge. An idea can start with one developer. It can become something that the entire organization is going to depend on. What's important is keeping it open. Yeah, completely agree, Dirk. As it scales, there's another critical principle. Scale shouldn't mean giving up choice. Miles will change, tools will change, infrastructure will change. How do we make sure DevOps and business stay in control, Dirk? That's exactly why openness matters. AI is evolving at an extraordinary speed. The model that's best today is probably no longer the model for six months, let alone tomorrow. The same is true for frameworks, for libraries, for hardware. Developers need the freedom to choose what works and what works best for them, and what they want to be building on. Businesses need to do the same. They need freedom. Maybe your workstation or workload runs locally, maybe it runs in the private data center, maybe it runs on the edge, or maybe it runs in the cloud. An open ecosystem lets you make those decisions based on the workload and the business, not based on what you're locked into. That's important because the choice you make today should not determine what you're allowed to do tomorrow. Open source has been at the heart of SUSE for over 34 years. We actually had our birthday this week. Congratulations. Yeah. We are a German company at heart, so it is great to be here as well. Scaling it is something that we deeply share with AMD. We believe innovation is the strongest when developers have the choice and the customers remain in control. Congratulations on 34 years. That is a big milestone. Thanks. You know, AMD and SUSE share the same DNA of openness. I mean, we have to think about how, as we scale our ambition, we don't want to shrink choice, right? We want to keep everything open. Yeah, that's absolutely right. Make the computer and make the computers work together is really important, Jack. Now Dirk, let's take that one step further. We talk about workstations becoming enterprise applications, but imagine a company with hundreds or thousands of powerful AI systems, workstations, servers, data centers, edge devices. Today we tend to think about those as individual pools of computing. I think the bigger opportunity is how we make them work together. What does that future look like to you, Dirk? I think this is where honestly it gets really exciting. Today these resources, they're powerful, but they're often managed as separate environments. The workstation is one world, the data center another, the edge another. The opportunity is to create a much more unified pool of AI capability. A Ryzen AI Halo Workstation could become part of that larger environment. If a workstation makes sense locally, run it locally. If it needs more capacity, use additional systems, and as demand grows, scale into enterprise infrastructure. The developers shouldn't have to think about each environment as a separate world. AMD brings compute across that entire spectrum. SUSE brings the open infrastructure and software to manage, connect, and scale it. Together we can move forward an environment where a lot of that compute works together as one open platform for AI. Dirk, I couldn't have said it better myself. Start on one system, then scale to many, keep it open, and ultimately make all the compute work seamlessly together. An idea can start with one person, one workstation, and grow to something the entire organization can use. That's the power of personal AI. Dirk, we're grateful for our partnership with you and SUSE. Thank you so much for joining me here today. Thanks for having me, Jack. Thank you, Dirk. Take care. Thank you. Every great era of computing has given people something they did not have before. The PC gave us computing in our hands. The internet gave us access to the world's knowledge. The smartphone gave us that power wherever we went. Now, AI gives us something new: the ability to turn imagination into action. Think about that. Something that begins as a thought can become something you can see, something you can create, and something you can build, and something that can reach far beyond you. One person can start something, and now one person can do so much more. That is the era of personal AI. AI that understands you, AI that works with you, AI that expands what you are capable of, not replacing your imagination, expanding it. We are only at the very beginning, because the most important question is not what AI can do, it is what you will do with it. I believe the greatest breakthroughs of this era has not been imagined yet. The next great idea has not been thought of, the next great invention has not been built, and the next great company has not been created. Maybe it is in this room. Maybe it is you. So imagine something that does not exist, then build it. When you think you have reached that limit, imagine and dream bigger. The best is yet to be imagined. Thank you, and have a wonderful IFA 2026.