The global semiconductor landscape is undergoing a fundamental shift, moving from a focus on raw training power to the practical complexities of large-scale deployment. In an in-depth interview, Wallace Kou, President and CEO of Silicon Motion, detailed how the generative AI has evolved beyond its initial stages. While the market's early gaze was fixed almost exclusively on NVIDIA's GPUs, the High Bandwidth Memory (HBM), and the CoWoS advanced packaging technology, Kou argues that the industry is now entering the "Inference" era that is turning previous under-estimation about storage's importance on their head.The Shift from Training to InferenceThe turning point for this realization occurred during the NVIDIA GTC conference in March 2026. CEO Jensen Huang unveiled the Vera Rubin architecture, a move that signaled a massive spike in demand for NAND flash memory. During the initial AI boom, the industry was preoccupied with training massive models, a process that relies heavily on the lightning-fast throughput of HBM. However, as these models move into the inference phase - where they are actually used by end-users to generate content or solve problems - the access to context, historical data, and massive datasets storage become the primary bottleneck.Kou notes a dramatic shift in market sentiment. Only two years ago, storage was often an afterthought in the AI conversation; today, it is a critical scarcity. "There is currently not a single global cloud service provider or major smartphone manufacturer whose demand for DRAM and NAND is being fully satisfied," Kou observed. This supply-demand gap has triggered a financial windfall for storage module manufacturers and memory giants, with some stock prices skyrocketing up to tenfold as the market reacts to persistent shortages and rising prices.Technical Paradigm Shift: CMX and the Infrastructure of ThoughtAt the heart of this transition is a new architecture introduced by NVIDIA: the CMX Context Memory Storage platform. This architecture is designed specifically to handle the "KV Cache" (Key-Value Cache), which allows AI models to remember the context of a conversation or a complex task during the inference process.The hardware requirements for the CMX architecture are staggering in their scale and technical demands. Each individual Rubin GPU requires 16TB of dedicated storage to function effectively within this framework. At a system-level scale, a single NV72 Vera-Rubin setup can demand more than 1 Petabyte, or 1,000 Terabytes, of total storage capacity. Beyond mere capacity, the CMX architecture facilitates direct GPU access to storage, a feature that bypasses traditional latency bottlenecks and ensures that AI inference remains fluid and responsive.While this creates a massive commercial opportunity for the storage industry, it also places an unprecedented strain on NAND production. Kou emphasizes that this is not just a cloud-based phenomenon. The explosion of Edge AI - AI processed locally on devices - is further complicating the supply chain. For instance, driven by major players like Meta, the market for smart glasses is expected to reach 60 million units this year. These wearable devices require high-performance embedded storage, creating a secondary front in the war for NAND capacity.Silicon Motion's Role: Solving the QoS BottleneckAs the world's leading NAND controller maker, Silicon Motion sits at the intersection of these competing demands. The primary technical challenge in modern AI environments is maintaining Quality of Service (QoS). In a multi-tenant cloud environment, where multiple GPUs are accessing shared storage simultaneously for different inference tasks, data transfer speeds can often fluctuate or drop.To solve this, Silicon Motion has deployed its proprietary PerformaShape technology. This technology ensures that even under heavy, concurrent workloads, the transmission speed remains stable. By stabilizing these data flows, Silicon Motion has positioned itself as an "indispensable stabilizer" in the AI ecosystem.Beyond data path optimization, Silicon Motion is also extending its role into system-level infrastructure by providing enterprise-grade boot drives for leading AI GPU, TPU, and DPU platforms, ensuring system reliability and fast initialization at scale.The Crisis of Imbalance: Kou's "Capacity Persuasion" EffortsDespite the record-breaking revenues, Kou is deeply concerned about the "shadows" lurking behind this prosperity. The current memory market is suffering from a dangerous imbalance. To maximize profits and satisfy the insatiable hunger of AI cloud giants, major manufacturers like Samsung, SK Hynix, and Micron are funneling the majority of their capital expenditure (CAPEX) into HBM and DDR5 production.This strategic pivot has effectively "squeezed" the production capacity available for standard NAND flash. Kou warns that this "AI squeezing effect" could lead to a collapse in traditional sectors. Over the past eight months, Kou has embarked on a global mission, meeting with leaders at Samsung, SK Hynix, Kioxia, SanDisk, YMTC, and Micron. His message is one of "capacity persuasion": he is urging these giants to reserve a portion of their production lines for the automotive, PC, and smartphone industries."If these foundational industries break because they cannot find parts, Edge AI will have no 'soil' to grow in," Kou warned. He believes that a total focus on the high-margin AI server market could eventually backfire, destroying the broader technology ecosystem that supports AI development.A Stabilizing Strategy: From Cloud to EdgeSilicon Motion is positioning itself as the "transition enabler" for an industry in flux amid an expected 2–3 year supply shortage. As NAND manufacturers concentrate their internal resources on AI-driven initiatives, they are increasingly outsourcing non-core and mainstream projects, such as PCIe Gen5 controllers and embedded solutions. In this shift, Silicon Motion has emerged as a preferred partner to fill the resulting gap.At the same time, as rising prices weigh on demand in the PC and smartphone markets, the company is helping customers pivot toward automotive and AIoT applications, including rapidly growing segments such as smart glasses, which are seeing a surge in shipments this year.One of the most critical areas is the automotive sector, where Silicon Motion has spent a decade building a presence. While memory giants might see automotive requirements as "niche" or low volume compared to AI servers, Kou views them as essential to global stability. When major OEMs consider abandoning these specialized demands due to capacity constraints, Silicon Motion steps in to ensure the global automotive supply chain does not grind to a halt."We are not just looking for a surge in revenue; we want to fulfill our responsibility to the industry," Kou said. By providing stable controllers and storage solutions for AIoT and automotive applications, Silicon Motion is effectively repairing the cracks in a fractured global supply chain.Future Outlook: 2027 and BeyondThe current supply-demand imbalance is not a temporary glitch but a structural reality that Kou expects to persist until at least late 2027 or 2028. Several factors make it nearly impossible to add capacity quickly, for example, land acquisition is increasingly difficult. The lead time for building specialized cleanrooms and procuring critical equipment now exceeds one year.Kou predicts that while the DRAM shortage might begin to ease by the end of 2027, the relief for NAND will likely come even later. In this high-pressure environment, Silicon Motion's role as a key stabilizing force becomes increasingly important.Particularly in emerging sectors such as smart IoT and automotive applications, Silicon Motion delivers reliable controller and storage solutions, filling the vacuum left by production shifts at major manufacturers or by projects lacking sufficient engineering support.By helping global clients navigate the complexities of geopolitics and capacity wars, Silicon Motion aims to ensure that the AI revolution leads to a steady, sustainable future rather than a chaotic collapse of the broader tech industry.AI inference boom fuels supply-demand imbalance until 2027-2028, says Wallace Kou. Credit: Silicon Motion

At COMPUTEX 2026, Fibocom outlined its perspective on the deployment path of edge AI: as AI moves from the cloud to end devices, the key to large-scale adoption lies not only in model performance, but also in the synergy between stable connectivity and local computing power.Based on this insight, Fibocom continues to strengthen its "Connectivity + AI" capability. On one hand, with wireless modules at its core, Fibocom has built a connectivity product portfolio covering FWA, AI CPE, and next-generation Wi-Fi 8, providing a high-speed, low-latency, and highly reliable connectivity foundation. On the other hand, by leveraging its full-stack AI solutions, Fibocom enhances on-device AI capabilities, enabling terminals with local perception and processing power to improve the stability and responsiveness of AI applications in complex environments.For intelligent edge scenarios, Fibocom launched the Fibocom ClawBox, a high-efficiency AI computing terminal natively compatible with intelligent agents such as OpenClaw and Hermes Agent. Built on a high-performance heterogeneous computing architecture, the product integrates CPU, GPU, and NPU, delivering up to 18 TOPS @ INT8 hybrid precision computing power. With typical power consumption at around 5W, it enables high-density, multi-task parallel inference. It supports multi-stream video analytics, complex visual structuring, voice, and multimodal model inference, while offering low latency, high reliability, and localized deployment capabilities. Combined with Fibocom's Skill Hub, it is widely applicable in security, transportation, robotics, and other intelligent edge scenarios, supporting the large-scale deployment of smart industry applications.At COMPUTEX 2026, Fibocom will showcase a range of innovations at Booth K1022, including AI companion solutions, robotic lawn mower solutions, mobile robotics solutions, and more, demonstrating how "Connectivity + AI" is creating value in high value scenarios. Industry partners are warmly welcome to visit and connect with us.

MSI, a global leader in high-performance computing and industrial solutions, returns to COMPUTEX 2026 (Booth #J0605a) to unveil its strategic AI roadmap. This year's showcase centers on a seamless continuum from data center scale to autonomous edge execution, featuring liquid-cooled AI platforms and supercomputers built on NVIDIA MGX, NVIDIA DGX Station, and NVIDIA DGX Spark architectures.Cloud Foundation: Liquid-Cooled Infrastructure for Hyperscale AITo meet the demands of modern AI data centers, MSI is introducing high-density platforms that prioritize thermal efficiency and performance: CG681-S6093 6U Liquid-Cooled AI Server (based on NVIDIA MGX): Built on NVIDIA MGX architecture, this server supports dual AMD EPYC processors and up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition Liquid Cooled GPUs. It delivers the compute density required for large-scale AI inference - with support for a wide range of agentic, physical AI, scientific computing, simulation, graphics, and video workloads.High-Speed Connectivity: The platform is equipped with NVIDIA ConnectX-8 SuperNICs, providing up to 8×400Gbps Ethernet connectivity for distributed AI environments.    Rack-Scale Scalability: MSI's liquid-cooled rack-scale architecture supports up to four CG681-S6093 GPU systems within a 48RU configuration. Networking is anchored by NVIDIA Spectrum-4 SN5600 Ethernet switches and SN2201 out-of-band switches for high-performance AI cluster connectivity.Deskside Development: The Desktop AI SupercomputerBridging the gap between the data center and the developer's desk, MSI presents high-performance AI computing for local workflows: XpertStation WS300 (NVIDIA DGX Station): Powered by NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip with up to 748GB coherent memory and 7.1TB/s HBM3e bandwidth, supporting Windows-based AI development while enabling high-speed CPU-GPU data sharing for large AI models and data-intensive workflows.Efficient AI Refinement: The WS300 is designed for AI model development, fine-tuning, and inference, utilizing a compact, liquid-cooled design with dual 400GbE networking powered by NVIDIA ConnectX-8 SuperNICs to sustain peak performance. (Recently showcased at GTC Taipei).Edge Execution: EdgeXpert and Autonomous IntelligenceMSI is bringing data center-level performance directly to real-world environments through the debut of its edge supercomputing platform: EdgeXpert AI Supercomputer: Built on the NVIDIA DGX Spark platform, EdgeXpert enables enterprises to deploy smarter, faster, and more scalable AI agents and applications at the edge.OpenClaw/Hermes Agent on MSI EdgeXpert: MSI showcases prominent AI Agent frameworks, providing open-source Agentic AI structures that support sustainable local operations and self-optimization capabilities.EU CRA Compliant Agentic AI with Galene Elettra: Powered by a Multi-Agent System (MAS), this solution enables intelligent decision-making across complex workflows while maintaining compliance with the European Cyber Resilience Act (CRA). Legal AI Suite: A specialized platform for enterprise AI, streamlining legal research, document analysis, and IP governance. Smart Campus Patrol: The Smart Campus Patrol Vehicle demonstrates real-time computer vision for precision inspection and smarter industrial operations. Tarot AI Experience with Reachy Mini: An interactive demonstration of Agentic AI, blending robotics with generative AI to deliver personalized engagement. End-to-End AI in Action Across IndustriesMSI delivers purpose-built Edge AI solutions developed with leading industry partners across diverse vertical markets:Smart Manufacturing & Semiconductor:The Edge AI Box MS-C910E with Memorence AI enables real-time machine vision. Partnering with Qiming Tech, MSI leverages the Edge AI Box MS-C939 to deliver real-time, high-precision automated optical inspection (AOI) to optimize semiconductor production yield.Voice AI & Driver Safety: Powered by Ubestream, the Slim Box MS-C926 provides real-time translation, while the Embedded Box MS-C927 enables instant voice-to-order experiences for retail. For driver safety, the In-vehicle Box MS-C932 runs real-time AI Driver Fatigue Detection to enhance on-road security.Smart Transportation & Precision AgricultureIntegrating edge inference with frontline mobility and field operations, MSI optimizes commercial transit and smart farming workflows:Smart Transportation × AI Vision Solutions:MSI is expanding its smart transportation portfolio with mobility solutions powered by Edge AI and in-vehicle vision. The lineup features fleet management tablets, telematics boxes, smart rearview mirrors, and AI-enabled ADAS and DMS systems for fleet monitoring and video analytics. Utilizing real-time AI processing, MSI helps logistics and fleet operators improve management efficiency and strengthen road safety.Smart Agriculture & Drone Integration: MSI introduces an intelligent agricultural solution that seamlessly integrates autonomous drone technology, Edge AI computing, and ground control systems.Built for harsh environments, this comprehensive platform combines drone ground control stations, rugged tablets, T-Box connectivity modules, and centralized multi-drone management platforms. It is engineered for critical use cases including automated field inspection, precision spraying, crop monitoring, and pest/disease identification. By leveraging real-time video analytics, edge inference, and cloud data synchronization, the solution empowers agricultural operators to boost efficiency, streamline operational workflows, and accelerate the transition to smart farming.Extreme Field Mobility (MS-NE21): The NE21 Rugged Industrial Tablet features an Intel 13th Gen Core i Series (Raptor Lake-U) processor, supporting up to 32GB LPDDR5/LPDDR5X memory and 2TB PCIe SSD. Built to MIL-STD-810G standards with an IP65 rating, it survives 4-foot drops and functions from -10°C to 50°C. It offers a 650-nit sunlight-readable 11.6" display with glove/wet-touch modes and a continuous hot-swappable battery system (64Wh to 98.1Wh).Industrial Panel PCs & Robust Hardware FoundationTo support heavy enterprise workloads, MSI highlights its high-reliability industrial hardware portfolio:Industrial Panel PCs: Features include the MS-1A81 (21.5") for Smart Healthcare clinical workflows; the MS-1A22 (12.1") and MS-1A32 (15") for factory floor monitoring; and the MS-1A91 (10.1") and MS-OP01 (15.6") for secure Smart Locker Systems.Hardware Backbone: For multi-industry infrastructure durability, MSI showcases its robust hardware portfolio featuring the Intel Wildcat Lake, NXP, and NVIDIA Jetson Thor series platforms, alongside industrial 4U rackmount systems.Sustainability & EV Charging InfrastructureBringing intelligent infrastructure to the energy transition, MSI presents its smart EV charging solutions:Eco Series Home EV Charger: Awarded the Taiwan Excellence Award, this residential smart charger delivers up to 22kW three-phase output. Powered by NXP industrial MCUs and AI smart control, it integrates with solar storage. Featuring a UL94-V0 fire-rated enclosure (extinguishing over-heat sources in 10 seconds) and RDC-DD leakage detection for underground parking safety, it holds global safety certifications, RPC certification in Taiwan, and $5M USD product liability insurance.MSI Hyper 80 Dual Fast Charger: Designed for urban commercial hubs, this DC fast charger delivers 80kW power distribution within an industry-leading 30cm ultra-slim chassis to maximize space efficiency.Visit MSI at COMPUTEX 2026, Booth #J0605a, Hall 1 (1F), to experience the future of the cloud-to-edge AI ecosystem.Credit: MSICredit: MSI

Solid State Storage Technology Corporation (SSSTC), a subsidiary of Kioxia Corporation and a leading global SSD provider, is showcasing its enterprise SSDs designed for immersion cooling and engineered for AI-driven data centers, along with a comprehensive portfolio of industrial and enterprise SSD solutions.As generative AI and high-density computing evolve, thermal management is critical. To address this, SSSTC has optimized its SSDs for immersion cooling environments by enhancing corrosion resistance through specialized materials, component protection, and structural design. The lineup includes the SATA ER3, ER4, and ER5 series, as well as the PCIe U.2 PJ1 and EJ5 series.These SSDs are optimized for immersion cooling environments, where systems are submerged in non-conductive dielectric fluids. By leveraging the high heat capacity and convective properties of liquids, heat can be efficiently dissipated through fluid circulation and heat exchange. This approach reduces reliance on traditional air cooling while improving Power Usage Effectiveness (PUE) and overall system reliability.SSSTC is also highlighting a range of industrial and enterprise SSDs optimized for AI and edge applications. Industrial SSDs support Edge AI and harsh environment deployments, with operating temperatures ranging from -40-degrees Celsius to 85degrees Celsius, along with anti-vibration and shock-resistant designs for outdoor and industrial environments. The pSLC architecture enhances endurance for sustained write-intensive workloads, while a multi-tier PLP (Power Loss Protection) framework -  including hardware PLP, firmware PLP, and PLN -  provides flexible data protection.Enterprise eTLC SSDs are designed to deliver stable performance for AI workloads, offering endurance options of 1 and 3 DWPD over a five-year period for varying workload intensities. Under sustained workloads, they maintain more than 90% random IOPS consistency, minimizing performance fluctuations. Firmware optimized for high-density computing enables low latency operation, while capacitor-based PLP and immersion cooling support ensure reliable performance in demanding deployment environments.With more than 18 years of in-house firmware development expertise, SSSTC understands diverse storage requirements across industries and provides flexible customization options, including configurable over-provisioning, lifespan and capacity optimization, performance and power tuning, and application-specific firmware development. SSSTC remains committed to helping customers build stable, efficient, and sustainable AI storage infrastructures.Founded in 2008, SSSTC became a subsidiary of Kioxia Corporation in 2020 and delivers high-quality SSDs through in-house firmware and NAND expertise. For more information, visit the SSSTC website.Remark: PCIe is a registered trademark of PCI-SIG.Credit: SSSTC

Elitegroup Computer Systems (ECS), a leading global provider of motherboards, mini PCs, and computing solutions, will participate in COMPUTEX 2026 from June 2 to 5, 2026, at Taipei Nangang Exhibition Center, Hall 1, Booth J1317a. Under the theme Power AI Computing, ECS will present its latest motherboards and LIVA Mini PCs, highlighting how compact and scalable PC platforms can support AI Agent workloads, Edge AI processing, smart healthcare applications, and embedded deployments.ECS will demonstrate how LIVA Mini PCs can be flexibly deployed in edge computing environments to support AI-assisted information retrieval, private knowledge base applications, healthcare data monitoring, and embedded commercial deployments. Through these demonstrations, ECS will highlight the role of LIVA Mini PCs in data processing, application execution, real-time monitoring, and vertical use cases. ECS will also showcase motherboard platforms with high-performance expansion capabilities, providing customers with a broader choice of computing foundations for AI and edge applications.Showcasing AI Agent Applications in PC EnvironmentsECS will showcase OpenClaw AI Agent applications running on an AMD desktop PC at its booth, demonstrating how AI Agent capabilities can be applied in PC-based environments. The demonstration will cover common scenarios such as system status queries, information search, and content summarization, showing how AI Agents can help users streamline daily operations and improve information processing efficiency. Through this demonstration, ECS will further present the flexibility and practical value of integrating AI Agent applications into commercial PC environments.LIVA Z11 PLUS. Credit: ECSLIVA One H810. Credit: ECSExtending Edge AI into Healthcare and Private Knowledge ApplicationsECS will showcase the LIVA Z11 PLUS mini PC in two Edge AI and data-driven scenarios: healthcare monitoring and private knowledge base applications. In the healthcare demonstration, the Z11 PLUS will support hemodialysis simulation and FHIR BOX applications, showing how a compact mini PC can serve as an edge computing node for medical data collection, real-time monitoring, and data format conversion.The knowledge base scenario will run a local database with a natural language interface, enabling users to query product and business information more intuitively. This highlights the role of mini PCs in enterprise information access, private data environments, and on-site applications where sensitive information needs to be managed locally. Powered by Intel Core Ultra processors, the LIVA Z11 PLUS provides high-speed storage, dual networking, and USB4 connectivity to support data-intensive Edge AI applications.LIVA Z15 PLUS. Credit: ECSExpanding the LIVA Lineup from AI-ready Performance to Embedded FlexibilityBeyond AI application demonstrations, ECS will present its full LIVA Mini PC lineup and next-generation platforms for commercial, edge, and embedded deployments. The new LIVA Z15 PLUS, built on the Intel Wildcat Lake platform with integrated NPU-based AI acceleration, will be a key highlight of ECS’s LIVA showcase, addressing high-performance commercial use, AI-assisted workloads, and edge computing applications.ECS will also feature the LIVA One H810, extending the LIVA One series' upgradeable socket-type design with the Intel Core Ultra LGA1851 platform. For low-power and embedded applications, the LIVA Z4F offers fanless reliability, while the LIVA Q4 combines an ultra-compact form factor with 45W USB Type-C power input for mobile, space-constrained, and flexible installation environments.

ZEISS, a global leader in optics and optoelectronics, will bring the quality discussion to the official COMPUTEX 2026 Forum stage for the first time this year, highlighting the growing role of quality in scaling AI hardware.As demand for AI infrastructure accelerates, quality is shifting from a manufacturing support function to a direct driver of performance, yield and delivery readiness. While public attention often centers on AI models, ZEISS says reliable hardware execution is becoming a decisive factor in AI deployment.Behind every AI interaction are massive data centers powered by thousands of GPUs. As systems scale from chip to rack, defects in semiconductor packaging, printed circuit boards (PCB/A), cooling systems and high-speed interconnects can affect uptime, deployment speed and total cost."With compute demand surging, manufacturers face record orders, but the challenge is delivering at scale with consistent quality," said Clive Yen, Global Head of Electronics Customer Segment, ZEISS Industrial Quality Solutions. "As systems grow more complex, quality becomes critical to reliable deployment. This is why we work across Taiwan's ODM ecosystem and the full AI server value chain to enable consistent, scalable quality.""At scale, even small defects can become major bottlenecks," said Tonmoy Kundu, Global Head of Sales, ZEISS Research Microscopy Solutions. "Manufacturers need faster insight, tighter process control and trusted failure analysis to accelerate next-generation AI hardware."ZEISS says it offers one of the industry's most comprehensive quality portfolios across the AI hardware value chain, supporting customers from semiconductor packaging and PCB inspection to liquid cooling, optical connectivity and final rack integration.At the forum, ZEISS will showcase solutions for advanced high-bandwidth memory (HBM), where rising stack heights and shrinking interconnect dimensions require high-resolution, non-destructive inspection and deep defect analysis.The company will also present metrology solutions for co-packaged optics (CPO), where ultra-tight tolerances for FAU and MPO connectors are essential to maintain alignment, coupling efficiency and long-term transmission reliability in 51.2T+ networks.At the exhibition hall (Booth J1109 | TaiNEX Hall 1, Taipei), ZEISS will showcase technologies spanning wafer process control, advanced packaging, X-ray inspection, electron microscopy, light and digital microscopy, and coordinate measuring machines. Applications will focus on chip manufacturing, PCB reliability, thermal management systems, connector quality and L10-L11 rack mechanical parts assembly.COMPUTEX 2026 runs June 2-5 in Taipei, where ZEISS will position quality as a foundational enabler of the next wave of AI growth. ZEISS will speak at the official COMPUTEX 2026 Forum on June 4, 4:30 p.m. to 4:55 p.m. at TaiNEX 2, Room 701, presenting "Quality Innovation Across the AI Chip-to-Rack Stack." The session will feature Tonmoy Kundu and Clive Yen. 

SK hynix Inc. (or "the company", www.skhynix.com) announced today the launch of the iHBM solution that embeds integrated cooling elements(ICEs) within the high-bandwidth memory(HBM) package for next-generation HBM products.Heat management has become a critical challenge as HBM technology advances with higher stacking and faster speeds to cater to the surging demand for AI data processing.The efficient management of power density in the Die-to-Die Physical Layer (D2D PHY) - the interface connecting HBM and GPU - has emerged as a key factor defining the competitiveness of next-generation HBM.With the iHBM solution, the company has taken a structural approach to addressing the heat management issue. Existing HBM products rely on an indirect cooling method that draws heat away through the core die. In contrast, the iHBM solution places ICEs directly in the D2D PHY area where heat concentration is the highest, creating an additional 'heat dissipation path'. This latest heat management solution helps reduce thermal resistance by 30% and enables chips to operate stably even in high-temperature and high-pressure conditions.The company's mass-production capabilities also serves as a key advantage. SK hynix's Wafer Level Packaging(WLP) process, based on its market-proven Mass Reflow Molded Underfill(MR-MUF) technology, enables stable high-volume production of iHBM-equipped chips. Furthermore, the solution offers high design compatibility with existing System-in-Package(SiP) architectures, allowing customers to adopt the new thermal technology with minimal design adjustments.Through the iHBM solution, slated for deployment in next-generation HBM products, including HBM5, SK hynix aims to increase the stability and operational efficiency of HPCs, AI data centers by meeting heat management standards required in high-density and high-bandwidth environments."iHBM is an optimal solution for thermal management, combining our memory design capabilities with advanced packaging technology," said Kangwook Lee, Senior Vice President and Head of PKG Development at SK hynix, adding "The company will cement its AI memory leadership by taking preemptive steps to offer values needed in the AI environment for its customers."

Global Unichip Corp. (GUC), the Advanced ASIC Leader, will showcase Jotunn8, a next-generation data center AI inference processor developed by VSORA, at the TSMC Europe Technology Symposium.This exhibition highlights the successful collaboration between GUC and VSORA, combining advanced AI architecture with leading-edge ASIC implementation and packaging technologies to deliver a high-performance, scalable inference solution for data center workloads.Enabling advanced AI inference through close collaborationJotunn8 is VSORA's flagship AI inference processor, purpose-built for inference workloads. It delivers ultra-low latency and very high throughput by addressing the memory wall bottleneck, enabling cost-efficient deployment of large-scale AI models.GUC provided comprehensive turnkey ASIC services for Jotunn8, managing the full implementation from netlist to manufacturing. The project demonstrates GUC's expertise in complex system integration, including: Advanced chiplet architecture design and integration, High-bandwidth memory integration with HBM3E PHY and controller, 2.5D die-to-die connectivity using GUC's 17.2 Gbps GLink-2.5D interconnect, Advanced packaging implementation leveraging TSMC CoWoS-S technology (3x reticle size), Implementation on TSMC's 5nm process node,Full-system co- optimization of signal, power, and thermal integrity (SI/PI/TI), Power and IR optimization to enhance overall system efficiency.This collaboration underscores GUC's capability to deliver complex AI and HPC ASICs with optimized performance, power, and scalability.Leveraging a strong ecosystem with TSMCJotunn8 also reflects the strength of GUC's long-standing collaboration with TSMC. As a key partner in TSMC Open Innovation Platform (OIP) ecosystem, GUC enables customers to efficiently adopt advanced process nodes and packaging technologies, bridging innovative architectures with leading-edge manufacturing.By leveraging TSMC's advanced CoWoS packaging and 5nm process technologies,GUC ensures high-quality silicon execution and a reliable path to production for demanding AI applications."We are proud to showcase Jotunn8 at the TSMC Europe Technology Symposium. This achievement highlights our strong collaboration with VSORA and demonstrates GUC's ability to bring complex AI processors to silicon using advanced process and packaging technologies," — said Patrick Wang, Senior Vice President and Chief Revenue Officer at GUC."Jotunn8 is designed to unlock a new level of efficiency and scalability for AI inference. Our collaboration with GUC has been key in successfully translating this architecture into a highperformance silicon solution,"— said Khaled Maalej CEO of VSORA.Driving next-generation data center AIThrough this collaboration, GUC and VSORA address the growing demand for high-performance,energy-efficient AI inference in hyperscale data centers. Jotunn8 enables new levels of throughput and latency optimization, supporting large-scale AI deployment across cloud and enterprise environments.

As Artificial Intelligence (AI) applications rapidly expand from cloud computing to edge computing, smart manufacturing, networking equipment, and AI PCs, stable and high-efficiency power supply has become the critical foundation for AI system operations. FSP Group will make a major appearance at COMPUTEX 2026 under the theme "Powering AI Together," showcase to  comprehensively demonstrate its AI power solution layout from data centers to end-user devices.Collaborating with Global ICT LeadersThis year, FSP Group has joined forces with multiple AI and Information and Communications Technology (ICT) leaders to demonstrate application results. Partners include Advantech, AIC, ARBOR, Chenbro, CyberTAN, UNEEC, Intel, SPARKLE, Trust5 intelligence and Uniwill.Through live system demonstrations, FSP will showcase how its power solutions are successfully integrated into AI servers, industrial PCs, and smart devices, further driving the rapid implementation of AI business applications.Focusing on Critical Power Infrastructure in the AI EraThis exhibition showcases FSP's AI power solution portfolio spanning from data centers to end devices, focusing on the critical power infrastructure driving the AI era, including: Data Center Power Shelf, Battery Backup Unit (BBU) Shelf, Networking Power Solutions, CRPS Solutions, Industrial PC (IPC) Power Solutions, USB PD Fast Charging Solutions, Grid Resilience Solutions, AI PC Power Solutions.These exhibits fully demonstrate FSP's comprehensive technical strength from core infrastructure to terminal applications.High-Performance Computing and Edge AI ApplicationsData center solutions focus on AI servers and High-Performance Computing (HPC) applications, providing high-efficiency and reliable power support tailored for high-power, high-density, and redundancy requirements. Meanwhile, networking and edge computing applications cover diverse scenarios including PoE switches, network switches, embedded systems, and edge AI servers, aiming to meet the power demands of various AI architectures."Micro Customization" as a Core Competitive EdgeFSP stated that in the face of rapid AI market development and diversified application needs, the company continues to leverage "Micro Customization" as its core competitive advantage. By providing highly flexible and reliable power solutions tailored to different application scenarios, FSP helps customers shorten development cycles and enhance system stability and market competitiveness.Complete Layout for High-End Consumer and Retail MarketsIn addition, FSP showcased a variety of products tailored for AI workstations. This impressive lineup features high-wattage power supplies such as the CANNON 3300W, the TWINS PRO 1400W ATX(PS2) redundant power supply, the high-power TWINS CRPS 3200W redundant power supply, and the 80 PLUS Titanium-certified MEGA TI 1650W White Edition. On the chassis front, FSP introduced multiple 4U Tower/Rack hybrid AI chassis, alongside the S210 chassis, specifically designed for compact performance and certified with NVIDIA SFF READY.  Providing a complete solution from power supply to chassis cooling, this display fully demonstrates FSP's comprehensive product layout in both enterprise-level applications and high-end consumer markets.With the rapid advancement of AI applications, power supplies have evolved from backstage components to the critical core of system performance and stability. Through COMPUTEX 2026, FSP Group will further demonstrate its technical prowess and industrial layout in the AI power sector.For more information on FSP products, please visit.

As generative AI and intelligent applications continue expanding from cloud infrastructure into real-world environments, demand is rapidly increasing for Edge AI platforms capable of delivering real-time computing, low latency, space efficiency, and long-term operational stability. At COMPUTEX 2026, Shuttle Inc. will showcase a series of Edge AI computing platforms powered by Intel Core Ultra processors, highlighting AI-ready edge computing systems designed to accelerate practical AI deployment across commercial and industrial environments.As AI applications continue expanding into retail, healthcare, smart manufacturing, and visual computing environments, Shuttle continues to focus on compact and deployment-ready Edge AI platforms designed for different spaces and operational requirements.At COMPUTEX 2026, Shuttle will present Edge AI platforms ranging from compact 1-liter systems to expandable 5-liter AI workstations, supporting workloads including visual AI, intelligent analytics, digital signage, healthcare applications, agentic AI workflows, and industrial automation.Among the featured systems, the DB860 supports multi-display output and rich I/O connectivity for digital signage, retail analytics, and control center applications. The XB860G2 provides PCIe Gen5 expansion capability for GPU and AI accelerator integration targeting advanced visual analytics workloads. The ultra-compact NT20H delivers up to 99 TOPS AI performance for space-constrained edge deployments, while the fanless BPCAR03 industrial platform is designed for smart factory and AIoT environments requiring wide-temperature support and long-term reliability.In addition to hardware platforms, Shuttle will also demonstrate several Edge AI deployment scenarios across commercial and industrial environments during COMPUTEX 2026.Experience the future of Edge AI with Shuttle at COMPUTEX 2026! We warmly invite our industry partners to join us from June 2–5 at the Taipei Nangang Exhibition Center, Hall 2, 4F (Booth R0114). Connect with our team and explore our latest solutions in action. Click here to book your exclusive meeting or download our newest product highlights today!