AGI Technology will showcase a diverse lineup of next-generation storage products at COMPUTEX 2026, centered around this year's core theme: "Think Big, Build Grand." Highlighted exhibits include DDR5 R-DIMM, PCIe Gen5 SSDs, and new additions to the AGI Smart Life industrial solution series. Additionally, AGI will feature its patented thermochromic heat-dissipation memory, alongside a strategic collaboration with StorArt on memory cards and a major external storage lineup featuring a heavyweight Japanese anime crossover.AGI is aggressively expanding into the extreme gaming and edge computing markets, empowering users to build a more robust foundation for high-performance computing.AI Solutions: Debuting DDR5 R-DIMM Server Memory to Launch a New Era of AI ComputingAs AI server performance bottlenecks shift toward memory bandwidth, AGI provides robust data throughput for next-generation processor platforms to meet the demands of AI computing and big data analysis. The initial rollout features mainstream high-speed specifications of 5600 MT/s and 6400 MT/s, with capacity options of 32GB and 64GB. Utilizing strictly screened high-quality original ICs, these modules ensure extreme stability and signal integrity under continuous 24-hour high-load environments—a storage solution tailor-made for next-generation AI data centers.Crossover Innovation & Major Collaboration | Expanding Industry Horizons and Unleashing the Hardcore Spirit of DAN DA DANAGI is expanding industry horizons through a strategic partnership with StorArt, pushing performance to new heights. A major highlight is the heavyweight crossover with the hit anime DAN DA DAN, featuring high-end internal DRAM, SSDs, and a brand-new Rugged External SSD (ED268). This powerhouse combines USB 20Gbps speeds with IP-rated durability and 2-meter drop protection. By merging elite tech with pop culture, AGI invites the world to embrace a "Bold" attitude at COMPUTEX 2026！Smart Life Solutions | Entering the Industrial Market with Grade-A Applications to Build a Reliable Foundation for Edge ComputingTo meet the rising demand for high-frequency data processing and edge computing, AGI's "Smart Life" series delivers industrial-grade solutions engineered for harsh environments. This specialized storage lineup ensures absolute data and video integrity with "Zero-Frame-Loss" performance.During the exhibition, the AGI booth will feature five major themed experience zones. Visitors who complete the stamp collection challenge can redeem an exclusive specialty coffee and enter a lucky draw for a chance to win premium DRAM products valued at over NTD10,000.AGI sincerely invites all content creators, gaming enthusiasts, and storage technology professionals to visit our booth and explore the new possibilities of next-generation storage.Join AGI at COMPUTEX 2026 from June 2–5 at Taipei Nangang Exhibition Center, Hall 1, 4F. Visit us at Booth L1117a to explore our latest innovations.AGI introduces new memory solutions engineered for peak performance.Credit: AGI

With demand for AI edge computing continuing to grow, Retronix announced that its Sparrow Hawk AI edge computing platform has officially entered mass production and is now available to the global market. The platform is powered by the automotive-grade SoC R-Car V4H from Renesas Electronics, delivering up to 30 TOPS (Dense) AI inference performance. Sparrow Hawk is designed to support a wide range of edge AI applications, including robotics, smart manufacturing, and other intelligent edge systems.According to Retronix, Sparrow Hawk integrates high-performance AI computing capabilities with a comprehensive I/O expansion architecture, enabling diverse edge computing scenarios such as computer vision, industrial automation control, and smart device development. The platform provides a wide range of interfaces including PCIe, USB 3.0, DisplayPort, CAN-FD, AVB Ethernet, and MIPI CSI/DSI. It also features an expansion connector compatible with industry-standard accessories, allowing developers to flexibly integrate peripheral modules and accelerate AI application development and deployment."We are pleased to see the global launch of Retronix's Sparrow Hawk Edge AI Computing Platform, powered by the automotive-grade R-Car V4H SoC," said Aish Dubey, Vice President and Head of the SoC Division at Renesas Electronics. "We value our collaboration with Retronix and look forward to continuing our work together to bring more innovation to the edge AI ecosystem."Regarding product compliance, Sparrow Hawk has obtained several international certifications, including FCC, CE, UKCA, and VCCI. These certifications indicate that the product meets international standards for electromagnetic compatibility (EMC), safety regulations, and other regulatory requirements, enabling Sparrow Hawk to be legally marketed and deployed in major regions including the United States, the European Union, the United Kingdom, and Japan.Sparrow Hawk was developed with an Upstream First software strategy that helps reduce entry barriers for developers. To meet diverse application needs, it fully supports two major software distributions. The Debian BSP is primarily aimed at industrial applications, providing a familiar software installation experience, while the Yocto BSP continues the high performance and reliability DNA of the R-Car series in the automotive field. Developers can access source code, Startup Guides, and detailed technical tutorials directly through the R-Car Open Source Community GitHub page. The community has established FAQ and Q&A forums, allowing global developers to communicate with experts in real-time and accelerate the transition from prototype to production.For hardware configuration, Sparrow Hawk is available in 8GB and 16GB LPDDR5 memory versions. The Basic Kit includes the Sparrow Hawk SBC, while the Complete Kit adds a 32GB microSD card, a fan HAT and a debug serial cable for a ready-to-use development setup.The Sparrow Hawk single-board computer and development kits are now officially available through Shimafuji Electric, with broader availability planned through major electronic component distributors such as Mouser Electronics and Digi-Key Electronics. Retronix Technology will work closely with Renesas Electronics to provide comprehensive technical support for developers and customers worldwide.For more technical details, please visit: Retronix Official Website.

VIA Labs, Inc. (VLI), a leading supplier of USB4, DisplayPort, SuperSpeed USB, and USB Power Delivery Controllers, today announced the launch of its first MST Hub VL610 at Computex 2026. Following the market success of the VL605 USB-C to HDMI 2.1 signal converter, VL610 is a new-generation chipset designed to address the growing demand for multi-display expansion. Supporting up to three high-resolution displays simultaneously, it sets a new benchmark for USB-C docking solutions. Attendees can learn more about VLI's products during Computex 2026 at Booth N0614 in Hall 1, Taipei Nangang Exhibition Center.The VIA Labs VL610 is a highly integrated DP 2.1 HBR3 Multi-Stream Transport (MST) hub designed for USB-C docking stations and high-end video adapter applications. Compared to its predecessor VL605, which supports a single HDMI 2.1 output, VL610 significantly enhances display performance. VL610 series includes two SKUs: VL610 supports three video outputs, while VL610D supports two video outputs, addressing different docking design needs. Its flexible architecture includes one fixed HDMI 2.1 FRL transmitter, one configurable port supporting DP++ or HDMI 2.1 FRL, and a third configurable port supporting DP++ or HDMI TMDS (VL610 only). VL610 supports a single display up to 8K60Hz or 4K240Hz, and up to three displays at 4K60Hz or QHD144Hz. It also enables up to six independent audio and video streams, with a single DP output supporting up to four MST streams. Full support for color formats and audio ensures a high-quality multimedia experience.In terms of advanced display technologies, VL610 integrates a DSC 1.2a decoder, supporting decompression to HDMI output or direct pass-through to compatible displays. It also supports cross-platform Variable Refresh Rate (VRR), delivering smooth visuals for high-end gaming and professional imaging applications. VL610 also integrates ECDSA-256 asymmetric authentication, enabling secure firmware updates and protection against malicious firmware attacks.VL610 also features a unique Logo Bitmap display capability with event-triggered graphics. It can display brand logos, warning messages, or guidance screens in scenarios such as host disconnection, USB-C port anomalies, or link errors, helping users quickly identify and resolve issues. This feature enables brands to proactively communicate with users during idle or fault conditions, enhancing both brand visibility and user experience."VL610 represents a key milestone in VLI's multi-display signal conversion technology. Evolving from the single HDMI 2.1 output of VL605 to the triple-stream architecture of VL610, we have translated customer needs into product innovation. VL610 is positioned to become the core chipset for next-generation high-end USB-C docking stations, further strengthening VLI's leadership in USB-C display interface solutions.", said Wayne Chang, Director of PM at VIA Labs.

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."

With the first phase of the European Union's Cyber Resilience Act (CRA) set to take effect imminently, German-based automated cybersecurity compliance provider ONEKEY, together with its Taiwanese local partner Trinity Technologies, notes that the CRA elevates product cybersecurity from a "one-time audit" to a "long-term governance" framework. Its review documentation and reporting processes are subject to an extremely rigorous compliance framework. To this end, ONEKEY and Trinity Technologies jointly recommend that Taiwanese ICT companies prepare early, leveraging localized professional support and automation tools to complete verification, and transform compliance pressures into market competitiveness.Cybersecurity Management Obligations Extend from Shipment to the Entire Lifecycle, Making Reporting Mechanisms a New Challenge for EnterprisesThe CRA will first implement vulnerability reporting and response mechanisms on September 11, 2026, and will take full effect on December 11, 2027. At that time, companies that have not achieved compliance will not only face substantial fines but may also see their products banned from sale in European countries. The CRA requires that all "Products with Digital Elements" (PwDE) fully incorporate cybersecurity management during planning, design, development, and maintenance, and address vulnerabilities throughout the entire product lifecycle. Consequently, most information and communications technology (ICT) supply chains—including smart home devices, connected devices, industrial control products, automotive electronics, and software solutions—are directly facing the impact of CRA regulations.Although the CRA is set to take effect in early September, a survey by ONEKEY reveals that as many as 68% of companies remain unfamiliar with the specific details of the CRA. ONEKEY notes that most manufacturers face four common misconceptions regarding the CRA regulations. First is a misjudgment of the implementation timeline, overlooking the fact that the "24-hour vulnerability reporting obligation" will take effect on September 11, 2026;second is a misunderstanding of responsibility, failing to elevate product safety to the level of "product governance"; third is a lack of compliance evidence, as a single penetration test is no longer sufficient to meet EU audits; and finally, manufacturers must be responsible for the overall safety of their products, which means the entire product supply chain must also comply with the CRA regulations.Five-Step Guide for Manufacturers to Build a Reproducible Operational Model, Starting with a Product Line InventoryJan Wendenburg, CEO of ONEKEY, stated: "We are at a turning point where global products face stringent cybersecurity requirements. With the EU's implementation of the CRA, cybersecurity is no longer a 'nice-to-have' for products, but rather a 'passport' for entering the European market. Many manufacturers possess world-class hardware engineering capabilities, but when faced with complex software supply chains, they often lack real-time, auditable transparency. ONEKEY's mission is to help clients bridge compliance gaps, guiding them from a 'good enough' approach toward full lifecycle cybersecurity governance."To transform this situation, ONEKEY leverages integrated platform technology to weave SBOM automation, vulnerability prioritization, impact assessment, and compliance guidance throughout the entire product lifecycle. This helps manufacturers establish a monitoring system ready for audits at any time, fundamentally resolving compliance challenges. Currently, ONEKEY has assisted global giants in Taiwan—including and ZyXEL—in aligning with CRA compliance requirements. Leveraging its extensive experience in CRA compliance, ONEKEY recommends that manufacturers follow these five steps to progressively achieve CRA compliance, thereby creating a repeatable compliance operating model. This transforms product "compliance" into a powerful competitive advantage, enabling them to seize the initiative in the European market. 1. Assess the product portfolio: Prioritize identifying which networking equipment, industrial computers, or IoT devices will be sold in the European market after 2026 and 2027. 2. Comprehensively establish a Software Bill of Materials (SBOM): For the aforementioned products, create an SBOM by scanning source code and binary code, ensuring it includes legacy firmware and vendor-provided firmware.3. Establish a Product Security Incident Response Team (PSIRT) process: Ensure that a dedicated team can assess the scope of impact within 24 hours of a vulnerability disclosure, fulfill reporting obligations, and proactively implement remediation.4. Implement "Shift-Left" Development: Integrate binary analysis and SBOM checks into every software development and firmware release cycle as a final pre-shipment verification.5. Turn Evidence into a Marketing Asset: Encourage the public disclosure of these automated detection results and compliance evidence to demonstrate product security as a key differentiator to European buyers.Incorporating AI Compliance Pathways Aiming for 100% Certainty into Automated TestingRegarding the application of AI in automated testing technology, ONEKEY emphasizes that while AI adoption certainly improves efficiency, it must also adhere to non-negotiable standards of determinism for cybersecurity compliance. Unlike the generative AI commonly found in the market today—which is prone to "hallucinations" or misjudgments—the AI models adopted by ONEKEY focus on the auditability and accuracy of results. Wendenburg stated, "In the realm of legal compliance, 'mostly correct' is 'wrong.' When EU auditors request proof of compliance, vendors cannot provide a result that is merely a guess generated by AI."To implement "Responsible AI," ONEKEY's technical architecture combines in-depth binary static analysis with rigorously validated deterministic algorithms. This means that every decision made by the system—whether scanning firmware, generating an SBOM, or matching against CRA regulatory requirements—is backed by clear logical reasoning and technical evidence. This approach ensures the system does not produce false regulatory determinations, preventing enterprises from falling into compliance traps due to AI false positives and even facing unnecessary legal litigation risks.Wendenburg emphasized that cybersecurity compliance requires not an AI that writes poetry, but a digital audit assistant with a high degree of "certainty." Through this rigorous technical approach, ONEKKEY assists Taiwanese companies in navigating the complex provisions of the CRA to produce legally valid compliance evidence that withstands official EU verification, reducing the error rate in regulatory judgments to near zero and fundamentally protecting the company's market reputation and operational rights.Trinity Technologies Strengthens Localized Technical Support to Help Manufacturers Become Trusted Security Partners in the EUChen Zhaoren, founder of ONEKEY's Taiwanese partner Trinity  Technologies, pointed out that as a global hub for the ICT industry , time has become the most pressing cost factor for enterprises in Taiwan as the EU's CRA regulations are set to take effect. He observed that while Taiwanese contract manufacturers have traditionally relied on "fast delivery" and "value for money" as their core competitive advantages, future market dynamics will shift toward "Secure by Design" and "supply chain transparency." Currently, EU brands have begun formally shifting the responsibility for providing Software Bill of Materials (SBOM), vulnerability reporting obligations, and up to five years of security maintenance to their Taiwanese supply chain partners. This means that Taiwanese manufacturers can no longer remain on the sidelines of this compliance wave.To help enterprises navigate these stringent regulatory hurdles, Trinity  Technologies is committed to localizing world-class tools like ONEKEY to assist Taiwanese companies in reshaping their R&D and quality assurance management processes. Through comprehensive protection technologies that integrate source code and binary firmware, Trinity  Technologies helps manufacturers establish automated compliance evidence chains and vulnerability response mechanisms without sacrificing cost-effectiveness. This not only accelerates clients' procurement review processes but also helps Taiwanese manufacturers transition from mere hardware suppliers to "security strategic partners" trusted long-term by the EU market, enabling them to withstand low-cost competitors in a regulation-driven competitive environment.Smart Compliance Wizard and Fast Start Program  Helping Enterprises Simplify Processes and Accurately Identify GapsTo help clients quickly grasp, organize, and implement the CRA regulations, ONEKEY has launched the "CRA Fast Start" acceleration program—specifically designed for equipment, machine, and system manufacturers—to assist companies unsure where to begin. Through automated software, the program addresses three key areas—readiness assessment, systematic vulnerability management, and 24/7 continuous monitoring—to guide companies step-by-step in identifying gaps in their existing processes and precisely pinpointing discrepancies with compliance standards.Additionally, ONEKKEY actively participates in the EU-funded Digital Europe program "CRA Compliance Wizard" and has integrated it into its products to provide automated guided workflows. Customers need only upload firmware binary files; the system automatically analyzes vulnerabilities and compares them against regulatory requirements, guiding enterprises through technical and organizational questions to generate a CRA-compliant "Statement of Compliance" and audit trail with a single click. This streamlines the previously cumbersome process of manually creating documents—which required significant legal and engineering resources—transforming it into an intelligent software tool that simplifies regulatory compliance in daily operations.If addressing EU CRA regulations is likened to a military campaign, "CRA Fast Start" serves as a "training and advisory program" that helps enterprises inventory their resources, formulate strategies, and build a defense system; meanwhile, the "Compliance Wizard" is the "intelligent detection weapon" used by R&D and QA personnel in actual operations to scan for product vulnerabilities, accurately answer regulatory questions, and generate certificates of compliance. The two complement each other, not only simplifying complex processes but also helping manufacturers transform compliance costs into a strategic advantage for entering the European market.Actively Preparing for the Phase 1 Deadline: Turning Compliance Pressure into a Strategic Advantage for Entering EuropeAs the September 11 deadline for the first phase of vulnerability reporting obligations draws near, Taiwanese ICT manufacturers can no longer afford to sit on the sidelines. The entry into force of the CRA not only signifies a comprehensive raising of the EU's product cybersecurity standards but also marks a survival-of-the-fittest test of supply chain resilience. Faced with extremely complex review documents and the stringent 24-hour reporting deadline, traditional manual response models can no longer meet compliance requirements. If manufacturers fail to promptly establish an automated vulnerability management and continuous monitoring system, they not only face the legal risk of product withdrawal from the market but may also lose their competitive edge as EU brands reshape their supply chains.Amid this tidal wave of compliance, the first step in proactive preparation lies in implementing a "product safety incident response team," "shifting security left," and "supply chain transparency." Through ONEKEY's automated compliance tools and Trinity Technologies' localized professional support, enterprises can start by inventorying their product lines to rapidly build highly reliable SBOMs and technical evidence chains. This transformation effort should not be viewed as a burdensome "compliance tax", but rather as a "market passport" for building international trust and countering low-cost competition. By taking action now and leveraging automation to bridge process gaps, companies can confidently turn regulatory challenges into a powerful moat for deepening their presence in the European market when the new regulations take effect in 2026.ONEKEY and Trinity Technologies meet to discuss CRA compliance solutions. Credit: DIGITIMESONEKEY CEO Jan Wendenburg explains CRA regulations and strategies for Taiwanese ICT companies. Credit: DIGITIMES

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!

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.

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.