The semiconductor industry is undergoing a profound transformation, driven by the strategic integration of artificial intelligence. At SEMICON Taiwan 2025, experts from Microsoft, Advantech, and Nvidia shared insights on the sector’s rapid shift toward “Lights-Out” factories—fully automated, 24/7 operations powered by AI. This evolution is designed to tackle monumental challenges while unlocking new levels of efficiency and resilience.AI Models Evolve for Deeper InsightsAI development has accelerated dramatically in recent years. From early domain-specific models for tasks such as entity recognition or computer vision, AI has advanced into sophisticated multi-domain models that integrate diverse functionalities within a single framework. Notable advancements include Retrieval-Augmented Generation (RAG), which constrains responses for greater accuracy, and multimodal models that seamlessly combine text, audio, and visual inputs.A pivotal breakthrough has been the emergence of reasoning models, capable of delivering deeper insights for complex problems across scientific, mathematical, and manufacturing domains. These models provide richer and more comprehensive analyses compared with their predecessors. Beyond individual models, multi-agent systems are now automating entire workflows, enabling specialized AI agents to collaborate on intricate challenges. One example is Toyota’s “obeya agent,” which integrates multiple sub-agents to assist powertrain engineers in analyzing vast datasets.According to Saj Kumar K, Senior Director of Manufacturing (APAC) at Microsoft, these advanced AI models are already being applied to manufacturing. The Production Copilot functions as a factory agent, analyzing data to identify root causes of downtime and optimize maintenance schedules. The Quality Advisor leverages multimodal AI for visual inspection and defect analysis, potentially rendering traditional image labeling obsolete. AI is also being used to optimize fab scheduling and Automated Material Handling Systems (AMHS), improving overall cycle times. Furthermore, large language models (LLMs) are training humanoid robots to perform complex tasks through single-shot teaching.Advantech’s Edge AI Strategy Targets Speed and PrecisionSemiconductor manufacturing faces unique pressures, particularly from the massive scale of data generated. A single wafer inspection can produce up to seven petabytes of data, requiring enormous computational power and real-time processing. The transition from single-beam to multi-beam inspection demands a tenfold increase in computational speed today, with projections of a hundredfold increase by 2030. Combined with the need for instantaneous data transfer and ultra-low latency to prevent defects, these requirements are driving a shift away from traditional data centers toward powerful edge computing solutions. Meanwhile, global labor shortages and skill gaps are accelerating the need for advanced automation.Advantech is at the forefront of this transformation with a comprehensive Edge AI strategy focused on process optimization, inspection, and fab automation. "Inspection is the heart of semiconductor manufacturers," said Magic Pao, Vice President at Advantech. The company’s strategy is built on four pillars: high computing power, high throughput, low latency, and robust local storage for analytics.Advantech is spearheading this transformation with a comprehensive Edge AI strategy focused on process optimization, inspection, and fab automation. “Inspection is the heart of semiconductor manufacturing,” said Magic Pao, Vice President at Advantech. The company’s approach is built on four pillars: high computing power, high throughput, low latency, and robust local storage for analytics.To achieve this, Advantech embeds AI into every element of its solutions, including products such as SkyRack and the MIC-7000/7500 industrial servers with direct-liquid cooling, both carrying critical SEMI certifications for fab deployment. High-speed connectivity is ensured through Time-Sensitive Networking (TSN) and industrial-grade switches. AI-powered vision systems—incorporating smart cameras with Nvidia GPU acceleration and FPGA-based technology—are central to the precision required for inspection. Advantech also emphasizes ecosystem strength, partnering with Nvidia for advanced computing and Intel for reliable system integration.Nvidia’s Physical AI Drives Industrial AutonomyNvidia is at the forefront of a new industrial revolution, evolving from a chipmaker into an “infrastructure company” advancing Physical AI. This strategy seeks to create “AI factories” that power advanced applications. Andrew Liu, Senior Manager at Nvidia, explained that this vision rests on two pillars: advanced foundation models for robotics and sophisticated simulation environments.Nvidia is developing powerful Vision-Language-Action (VLA) models that enable robots to interpret high-level prompts and images, translating them into precise physical actions and learning from observation or single-shot teaching. To address the scarcity of real-world robotics data, Nvidia relies heavily on its Omniverse platform for digital twins and simulation. Omniverse provides physically accurate, photorealistic virtual environments where synthetic data can be generated and robots can be trained and validated. This approach helps alleviate labor shortages and enables rapid “factory cloning” to mitigate geographic supply chain risks. Nvidia’s integrated platform spans systems for training foundation models, servers for Omniverse digital twins, and embedded platforms for industrial robot deployment.Toward the Lights-Out FutureThe convergence of advanced AI models, strategic infrastructure, and targeted edge solutions is accelerating the path toward fully automated “Lights-Out” semiconductor manufacturing, promising significant gains in productivity, efficiency, and resilience that will redefine the industry’s future. At SEMICON Taiwan 2025, the AI Technology Zone (advised by the International Trade Administration) underscores this transformation by showcasing the full AI ecosystem—from chip manufacturing and IC design to dedicated hardware.SEMICON Taiwan 2025 Smart Manufacturing Forum. SEMI

As the semiconductor market accelerates toward an estimated US$1.3 trillion valuation by 2030, industry leaders are converging on a shared vision. AI is ushering in an era of unprecedented growth and technological transformation.At the CEO Summit 2025, organized by SEMI, thought leaders from Infineon, NXP, Google,DENSO , Tenstorrent, and ASE gathered to discuss the critical role of power in enabling AI, the opportunities and challenges semiconductors bring across applications—from data centers to edge computing—and the importance of global collaboration across the ecosystem.In their inaugural speeches in Taiwan, the CEOs of Infineon and NXP delivered a unified message on the future of AI. Both leaders underscored a critical and often overlooked point: power is the foundational element driving the next wave of AI computing performance. Their presence also underscored Taiwan’s strategic importance within the global technology ecosystem.Infineon Technologies CEO Jochen Hanebeck highlighted that AI’s power demand is becoming a fundamental bottleneck. Without sufficient power, the challenge extends beyond generation to the infrastructure needed to deliver electricity from the grid to the processing core. Ultimately, he warned, society may be unable to enjoy the full convenience and progress AI promises.NXP Semiconductors CEO Kurt Sievers echoed this sentiment, emphasizing that AI’s power demands extend well beyond data centers into "physical AI" applications such as autonomous vehicles and humanoid robots. This shift, he explained, is fueling rapid growth in edge AI. When AI interacts with the physical world, factors such as bandwidth, energy efficiency, latency, and trust become paramount. Consequently, these complex systems require advanced architectures with sophisticated power management for critical functions such as motor control and sensor integration.Sievers further elaborated on this concept, referring to it as "attending A", which he believes will enable truly autonomous machines. He stressed the industry’s immense responsibility to ensure these technologies are developed responsibly for the benefit of humanity, remarking: "We all share quite some responsibility to do this right for the better of humankind." He explained that intelligent systems consist of multiple "agents" that must be carefully orchestrated and enabled to work in harmony.Dr. Hirotsugu Takeuchi, CTO of DENSO , shared insights on how semiconductors are transforming mobility. He outlined the requirements of automotive applications and the need for new SoCs to power the rise of Software-Defined Vehicles (SDVs). He pointed to the breakthrough of SiC technology as a key enabler of a sustainable, carbon-neutral society. Takeuchi emphasized that semiconductors will drive safer, more comfortable mobility and that collaboration with Taiwan and the broader global network will be essential to the future of the automotive industry.Rehan Sheikh, Vice President of Global Silicon Chip Technology & Manufacturing at Google Cloud, noted that demand for AI compute is growing exponentially. He highlighted that AI inference volume doubled in the past year, with Google’s internal data showing a 50-fold increase in monthly processed tokens across its services. Sheikh introduced Google’s latest TPU, Ironwood—a powerful, energy-efficient chip specifically designed for large-scale inference workloads. It represents Google’s first TPU purpose-built to balance high performance with power efficiency in meeting the demands of the AI boom.Jim Keller, veteran chip architect and CEO of Tenstorrent, emphasized his mission to champion open-source architectures and foster a spirit of global collaboration at a time when AI computing is becoming increasingly complex and costly. Keller explained that Tenstorrent aims to democratize high-end AI computing, making it more accessible, faster, and more open. The company’s strategy rests on three pillars: RISC-V, OpenAI, and a unified software stack. At the heart of this approach is a commitment to open standards and collaborative development, challenging the industry's traditionally closed ecosystems.During the fireside chat, speakers reflected on Taiwan's evolving role in the global semiconductor landscape. Long established as a manufacturing hub for logic chips, Taiwan is now emerging as the ideal platform for integrating diverse technologies and solving system-level challenges, thanks to its mature ecosystem, advanced packaging capabilities, and open-minded approach. Executives from Infineon, NXP, and Tenstorrent all stressed that no single company or region can address these challenges alone. The future of the industry, they agreed, depends on deep cooperation across sectors and geographies, combining expertise in areas ranging from high-performance computing to power and sensing technologies.In closing, ASE CEO Dr. Tien Wu referenced A Chip Odyssey, Taiwan’s first documentary film spotlighting its globally leading semiconductor industry. He described the film as a powerful reminder to both senior leaders and younger generations to maintain confidence in the steady progress of semiconductor development.The film chronicles Taiwan’s journey from humble beginnings to its emergence as a critical player in the global industry. Dr. Wu shared that he was deeply moved by the story of two generations of talent—each bringing vision, diligence, and leadership—who together created a lasting legacy. He extended this reflection beyond Taiwan, noting that the United States, Japan, Korea, and Europe have each experienced their own “chip odysseys.” This, he emphasized, illustrates that Taiwan's story is part of a larger global network of interdependence, where collective collaboration produces outcomes greater than the sum of individual contributions.SEMICON Taiwan 2025 CEO Summit highlights collaboration to drive semiconductor innovation. DIGITIMESTenstorrent CEO Jim Keller champions open-source architectures and global collaboration to democratize AI computing. DIGITIMES

SK hynix announced today that it has begun supplying its high-performance mobile NAND solution ZUFS 4.1 to customers, marking the world’s first mass production of this solution.The solution's adoption in the latest smartphones reinforces SK hynix's technological excellence in the global market. ZUFS 4.1 will enhance smartphones’ powerful on-device AI capabilities, offering users an innovative experience.SK hynix successfully completed the qualification process for the solution in June this year through close collaboration with customers. In July, the company began mass production and started supplying the product.ZUFS, or Zoned UFS, is an extended specification of UFS that applies Zoned Storage technology, which stores data in different zones based on its usage and characteristics.When installed in a smartphone, ZUFS 4.1 enhances the operation speed of the operating system (OS) and improves data management efficiency. As a result, it mitigates read performance degradation over extended use by more than four times compared to conventional UFS, enabling a 45% reduction in app launch times. Unlike conventional UFS, which writes new data by overwriting existing data, ZUFS 4.1 writes data sequentially. This data storage method has resulted in a 47% reduction in AI app launch times.These performance characteristics position ZUFS 4.1 as an optimal solution for today's mobile environment, where on-device AI and large-scale data processing are essential.In addition, SK hynix has significantly enhanced the error-handling capabilities of ZUFS 4.1 compared to version 4.0 developed in May 2024. By detecting errors with greater precision and clearly communicating the necessary corrective actions to the central processing unit, the latest solution is expected to significantly improve system reliability and recovery performance.“ZUFS 4.1, which we have successfully begun to supply, is the first solution developed and mass-produced through collaboration aimed at optimizing Android OS and storage devices. Looking ahead, its applications are expected to expand,” said Justin Kim, President & Head of AI Infra at SK hynix.“We will continue to supply NAND solutions that meet customer needs in a timely manner, while strengthening partnerships with global companies to strengthen our competitive edge in the AI memory sector.”Credit: SK Hynix