[Speech Summary] The DNA of Challenge: Shaping the Past and Future of Semiconductors Human being’s imagination has continuously created new technologies and industries. Semiconductors are a core technology that turns human being’s imagination into reality. The advancement of AI is exponentially increasing the volume of data and computation, further heightening the importance of semiconductors.

In particular, the era of Agentic AI and Physical AI requires new semiconductor innovations, including high-performance computing, HBM, and advanced packaging.

Samsung Electronics has taken on the challenge of the semiconductor business even under unfavorable conditions, creating inflection points in its business through memory, system semiconductors, and foundry. Innovations such as DRAM, V-NAND, FinFET, EUV, and GAA are the result of continuous efforts to overcome the limitations of existing technologies.

Future competitiveness comes not only from technology, but also from comprehensive innovation capabilities that connect memory, logic, foundry, and packaging. Such innovation cannot be achieved alone; it is made possible through trust-based collaboration with customers, universities, research institutes, materials companies, and equipment companies.

Leaders who create inflection points need Desperate Ownership, a positive attitude, passion, continuous self-development, and fairness. Ultimately, the leaders of the future must share dreams and visions, identify and take on future inflection points, lead challenges and collaboration, and become the Brain of the organization.

[Speech Summary] The State and Future of AI in Korea: Challenges and Opportunities Put in energy and capital, and out comes intelligence—tokens. Few countries possess every layer of the AI full stack, from energy and chips to infrastructure, models, and applications; Korea is one of them. Semiconductor in particular, which account for more than half the cost of an AI data center, are the layer where Korea holds real bargaining power on the global stage. Add a cultural base of perfectionism and rapid technology adoption, the geopolitical leverage, and an industrial base that combines manufacturing and electronics with software and strong application sectors such as defense, shipbuilding, healthcare, and content—all working at once—and now is the moment for Korea to move beyond being a component supplier and become a “full-stack AI nation.”

This talk surveys the AI economy, where value is migrating up the stack toward applications, and the global full-stack competitive landscape, then addresses the two critical challenges Korea faces at the crossroads between a component nation and a full-stack nation: trust and talent. The first principle of the AI era is that speed determines direction. Innovation comes from how fast one can run the loop in which humans define, AI executes, and humans verify—and trust between institutions, within organizations, and between humans and AI is its fuel. Finally, the talk notes that the bottleneck is no longer AI capability but the human capacity to use it, and underscores the value of talent that understands what to build, why, and with whom.

[Speech Summary] New standards create new growth ! Today, shifts in policies among global powers have weaponized national self-interest and natural resources, dividing the global market. With only 0.7% of the world's landmass, 0.63% of the global population, and virtually no natural resources, South Korea faces an increasingly challenging competitive environment on the world stage. For South Korean enterprises—whose growth is inevitably driven by trade—we have entered an era of global competition where creating value and achieving sustainable growth are only possible through "world-first" and "Only-One" innovative technologies.

Currently, the global semiconductor industry is advancing at an unprecedented pace as AI integrates into everyday life. In particular, semiconductor manufacturing technology has come to play the most crucial role in enhancing human intellectual capacity and quality of life. However, because technologies and architectures have not fundamentally deviated from the transistor framework established in 1948, this era demands a new paradigm of innovation. More than 99% of even the greatest innovations are results that others have already conceived and executed; focusing on the uncharted 1% that others have overlooked to establish new standards is the essence of new growth tailored for the AI era.

New growth is possible only upon new standards. No matter how superior a technology may be, the existence of competitors allows buyers to dictate prices; in contrast, even if an innovation has shortcomings, the absence of competitors empowers the creator to set the price. This is the difference between mere technology and true innovation, and precisely why innovation is essential. Driven by such world-first and Only-One innovative technologies, Jusung Engineering is committed to presenting a new paradigm for future advanced industries and relentlessly leaping forward.

[Speech Summary] AI driven Business Value Creation Strategy(TBD) As global AI competition intensifies, the importance of Sovereign AI has surfaced, positioning AI as a central pillar reshaping industrial structures. Leveraging these transformative shifts, LG AI Research is creating tangible value through AI technologies that collaborate in real time with domain experts. In manufacturing, AI precisely analyzes process data and control variables to closely support engineers' decision-making. In finance, it autonomously performs research and risk management to assist experts in high-level strategic planning. The true capability required for enterprises preparing for the AI era lies in building a sophisticated collaboration framework by enhancing AI literacy among leaders and practitioners while clearly defining field problems. Based on these practical experiences, this session delivers actionable insights into bridging technology adoption and real business performance.

[Speech Summary] Rethinking HBM for the Next Leap in AI The proliferation of Agentic AI is driving demand for unprecedented levels of performance, capacity, power and thermal efficiency in AI systems. Samsung Electronics is leading the continued evolution of HBM through HBM4E and HBM5, while also introducing zHBM, a new 3D memory architecture that goes beyond the limitations of conventional 2.5D structures. zHBM stacks memory directly on top of the processor and applies ultra-high-density TSVs and hybrid bonding to dramatically improve performance and energy efficiency. Furthermore, through close co-design of memory and processors, Samsung Electronics aims to overcome the Memory Wall and open up new possibilities for next-generation AI systems.

[Speech Summary] It’s the memory, stupid – AI era The rapid growth of AI models and services is making memory a critical factor in AI system performance. Memory bandwidth and capacity are not scaling as quickly as compute performance, while reasoning models and Agentic AI are dramatically increasing context lengths and token counts. As a result, the importance of memory continues to grow.

Meanwhile, AI algorithms and serving systems are evolving through memory-efficient attention, prefix caching, and hierarchical memory management. However, greater efficiency enables longer contexts, higher concurrency, more agent steps, and broader AI usage. Similar to the Jevons Paradox, even as the cost per token decreases, total token processing and memory traffic may actually increase.

Agentic AI is expected to accelerate this trend further, as autonomous agents continuously interact with LLMs, external tools, and other agents, driving greater demand for computing infrastructure. Sustaining the next wave of AI will therefore require not only software innovation, but also more efficient accelerators, memory, and storage. This talk explores why memory is becoming a key technology in the AI era.

[Speech Summary] AI Semiconductor Strategy in the Era of the Intelligence Production Economy As AI evolves from training to inference and onward to agentic AI, the rules of the semiconductor game are being rewritten. The center of gravity in AI infrastructure is no longer the compute core but memory: KV caches expand nonlinearly in inference workloads, tiered memory hierarchies anchored by HBM are becoming the new standard, and the linear equation mapping GPU shipments onto memory demand no longer holds. Yet this supercycle does not license unconditional optimism. Hyperscaler capital expenditure has outgrown internal cash generation and shifted toward debt financing, and the decisive variable behind the widening gap between announced investments and actual construction is power—a GPU takes eighteen months to procure, a grid a decade. The fate of Korea's three mega-projects, anchored by semiconductor clusters and gigawatt-scale AI data centers, will likewise be determined not by silicon but by electrons. The grammar of geopolitics is also being rewritten. The twenty-first-century order is reorganizing around the bottlenecks of intelligence—an era of techno-geopolitics in which the United States builds fortresses through export controls, China carves roads of efficiency, self-sufficiency, and open source, and national competitiveness shifts from GDP to GDI, Gross Domestic Intelligence. This lecture integrates migrating technological bottlenecks, infrastructure constraints, and the techno-geopolitical order into a single strategic frame. Bottlenecks migrate with technology, and power moves with them. The emerging bottlenecks of the AI era; AI memory, advanced packaging, power technology, overlap substantially with Korea's strengths. Beyond defending its HBM dominance, Korea must pursue memory–system co-design, robust public fabs, AI transformation of manufacturing (M.AX), and integrated energy–semiconductor security becoming an intelligence state indispensable to both fortresses while captive to neither.

[Speech Summary] Manufacturing Innovation Beyond Vision, Into Action Manufacturing technology has advanced alongside the history of human civilization. From the First to the Fourth Industrial Revolution, manufacturing innovation has opened new possibilities in every era, and Hyundai Motor Group has continued to advance its manufacturing technologies as part of this broader transformation.

Today, however, the manufacturing environment faces fundamentally different challenges: growing customer demand for high-mix, low-volume production; increasing pressure for faster delivery; the rapid growth of China's manufacturing sector and the rise of emerging EV companies; and structural challenges such as a shrinking working-age population and a shortage of skilled labor. Together, these factors are placing significant pressure on manufacturing competitiveness.

Hyundai Motor Group is responding to these challenges with its vision for an Autonomous Factory. By creating a smart manufacturing environment where people and robots collaborate and hardware and software converge, the Group plans to fundamentally enhance manufacturing competitiveness through two key pillars: automation and autonomy. The presentation will introduce key initiatives in high-difficulty automation and strategies to accelerate the transition to the Software Defined Factory (SDF), as well as a roadmap for expanding proven technologies from HMGICS to next-generation factories.

Manufacturing is a driving force behind Korea’s economic growth. Hyundai Motor Group will shape the future of manufacturing through AI and robotics and set a new benchmark for Korea’s manufacturing competitiveness.

[Speech Summary] From Digital AI to Physical AI: AI that Understands and Acts on the Real World This session will focus on the development trajectory of Robot Foundation Models. Recently, various approaches—including large-scale multimodal models, action-based models, world models, and Vision-Language-Action (VLA) models—have been proposed to implement robot intelligence. Active research is underway in both industry and academia to identify which architectures can pave the way toward general-purpose robot intelligence.

Furthermore, we will objectively assess the current state of the technology and examine key challenges that must be addressed for robots to understand the physical world and act autonomously. Specifically, we will highlight major technological directions for next-generation robot intelligence, such as data acquisition and training methods, physical world comprehension, reasoning and planning capabilities, and the integration of simulation with real-world learning.

Finally, from the perspective of transitioning from digital AI to physical AI, we will discuss how future robot foundation models will transform industry and daily life, as well as the technological evolution required to realize this shift.

[Speech Summary] Mission-Oriented AX: Why, What, and How We Create with AI Human society and global industries are rapidly entering the era of AX (AI Transformation), where artificial intelligence leads practical decision-making and physical control. As AI technology evolves into ‘Agentic AI’—which plans and executes autonomously—and ‘Physical AI’—which integrates with physical actions—AI has shifted from being a strategic option to a core element that determines a company’s fundamental competitiveness.

Amidst this paradigm shift, POSCO Group is focusing on the fundamental question: “Why and for what purpose is AI?” To answer this, we have established and are pursuing a ‘Mission-oriented AX’ strategy designed to accomplish ‘what needs to be done.’ This strategy aims to fully integrate AI into the Group’s core businesses—including steel, secondary battery materials, and energy/infrastructure—to create substantial financial performance and business value.

In this session, I will introduce POSCO Group’s three core transformation strategies and their achievements: Process Transformation: Maximizing efficiency and operational standards by utilizing the knowledge and intelligence required for manufacturing and business processes. Work Transformation: Redesigning work methods and supplying the necessary intelligence for daily office operations, research, and development. Value Transformation: Commercializing internally proven intelligence and knowledge into solutions and creating new value through accumulated intellectual capabilities. In particular, I will present POSCO Group’s unique methodology for measuring, managing, and improving the financial value of AI. I will also demonstrate how we are strengthening data-driven AX execution, moving beyond vague expectations. Furthermore, by sharing our governance framework for internalizing AI across the entire organization, I aim to provide practical insights into how a traditional manufacturing giant can be reborn as an AI-native enterprise.

[Speech Summary] BATTERY: THE HEART OF PHYSICAL AI The rapid growth of Physical AI is ushering in an era where AI controls physical-world devices such as robots and autonomous vehicles, making high-performance, high-safety batteries more critical than ever.

Samsung SDI is leading the humanoid market with specialized high-power cylindrical cells, while also developing all-solid-state batteries — a next-generation solution with enhanced energy density and safety.

This presentation will cover the advantages of cylindrical cells for humanoid applications, key technologies in materials, electrodes, and cell structure, as well as Samsung SDI's all-solid-state battery technology currently under development. It will also illustrate the present and future of batteries, and explain why they serve as the beating heart of Physical AI.

[Speech Summary] Strategic Direction and the Journey Toward Manufacturing AI Achieving South Korea’s ambition of becoming one of the world’s top three AI powers will depend fundamentally on the AI Transformation (AX) of its industries. Manufacturing, a sector characterized by highly skilled expertise, intensive labor, and a strong concentration of human capital, has long been the cornerstone of the nation’s economic growth.

The AX transformation of manufacturing represents the integration of South Korea’s accumulated industrial success formula into the era of AI. The capabilities, knowledge, and know-how cultivated through decades of manufacturing excellence must be converted into digital assets and AI models, creating a new engine of national growth. At the same time, this transformation is not only an industrial opportunity but also a critical mission to address demographic changes and ensure the sustainable participation of future generations in the manufacturing ecosystem.

This lecture presents a strategic vision for Manufacturing AI in South Korea and explores the future direction of the field through HD Hyundai’s AX transformation journey. As a global leader in heavy industry and advanced manufacturing, HD Hyundai provides a compelling example of how AI can enhance industrial competitiveness and drive innovation at scale.

The speaker, Youngok Kim, Chief AI Officer (CAIO) of HD Hyundai, is a Manufacturing AI specialist who majored in Artificial Intelligence and has spent more than 20 years applying AI-driven transformation in the manufacturing sector. Drawing on extensive experience across diverse manufacturing environments, he will share practical AX use cases, key implementation considerations, and insights into the future of Manufacturing AI powered by a Full Stack AI approach.

[Speech Summary] Branding in the AI Era As we consider the theme, “Beyond Followers: Transitioning to a Leading Economy,” much of the discussion has focused on technological leadership. Yet one critical force has often been overlooked: branding, the power that enables the market to recognize and trust that technology.

Branding is not simply about logos or advertising. It is the sum of experiences created by designing every touchpoint between a brand and its customers with a consistent sense of direction. At the heart of those experiences is always an authentic story, in other words, storytelling.

A leading economy cannot be built on technological capability alone, because creating a great product and being chosen by the market are two entirely different challenges. Technology becomes a brand only when the market recognizes its value and customers place their trust in it. A strong brand clearly and consistently communicates why the technology was created and what value it brings.

The way that value is communicated, however, must differ from one industry to another. Semiconductors must establish themselves as technological standards. Robotics must help people accept and integrate new technologies naturally into their lives. Biotechnology must prove its credibility and earn trust over time. Manufacturing must move beyond price competition and gain recognition for higher value.

In this context, AI is not a technology that creates a brand on its own. Rather, it serves as a tool for understanding the market and demonstrating the value of a brand more effectively.

Ultimately, if technology creates possibility, branding is the final step that transforms that possibility into market value.

[Speech Summary] From Autonomous Operations to Space Data Centers: The Rise of Space AI Space exploration and development, once dominated by large-scale government-led initiatives, are undergoing a rapid transition into the private-led "New Space" era. Simultaneously, the global technology landscape is experiencing a paradigm shift driven by Artificial Intelligence (AI). In today's space industry, AI has evolved beyond a conventional tool for data analysis. It has emerged as a fundamental game-changer, reshaping the entire space value chain from complex mission design and autonomous satellite operations to real-time on-orbit data processing and the establishment of space-based data centers. This presentation highlights the imperative role of AI in future space systems and outlines how Korea Aerospace Industries (KAI) is shaping the future of space technology through innovative Space AI solutions and strategic vision.

[Speech Summary] The way forward for K-New Space Space is no longer simply an area for preparing for the future. Recent conflicts, including the war in Ukraine, have demonstrated that satellite imagery, satellite communications, and positioning information are critical to achieving information superiority and sustaining military operations. Space has become essential infrastructure that simultaneously shapes national security, industrial competitiveness, and everyday services.

For Korea, this calls for a new approach beyond simply increasing the number of space assets. Given our limited budget, it is unrealistic to independently develop and own every space capability. Instead, we should build the ability to secure the right space information and services at the right time by integrating government and military satellites with commercial satellites operated by domestic companies and space assets of our allies.