EDOM Technology (TWSE: 3048), Asia's best solutions provider, today announced an expanded collaboration with NVIDIA to help enterprises accelerate Edge AI adoption by integrating the NVIDIA Edge AI platform, open-source models, and system optimization technologies. The collaboration enables organizations to lower the barriers to deploying large AI models while reducing overall deployment costs.As generative AI rapidly evolves from proof-of-concept to real-world applications, enterprise priorities are shifting. Rather than simply pursuing larger AI models, organizations are increasingly focused on running AI reliably on existing hardware platforms, accelerating deployment, and maximizing return on investment. Across industries from smart manufacturing and autonomous robotics to healthcare, Edge AI has become a key driver of digital transformation where efficient deployment is critical to successfully bringing AI projects into production.The way enterprises adopt AI is also changing. Instead of building models from scratch, organizations are increasingly leveraging mature foundation models and open-source AI ecosystems to accelerate development, followed by application-specific optimization and deployment. However, as AI models continue to grow in capability, several deployment challenges such as memory capacity, inference performance, and system integration have become major hurdles.EDOM addresses these challenges by combining the NVIDIA Jetson platform with NVIDIA JetPack, NVIDIA CUDA, NVIDIA TensorRT, NVIDIA Jetson AI Lab, and leading open-source AI models and inference frameworks. The company delivers a comprehensive portfolio of services spanning platform selection, model integration, model quantization, memory optimization, and system validation. This end-to-end approach enables enterprises to maximize hardware resource utilization, reduce deployment costs, and shorten the journey from proof of concept (PoC) to production.A notable example is NVIDIA Reachy Mini Jetson Assistant, which demonstrates how Headless Mode, NVIDIA Cosmos-Reason2 open VLMs, model quantization, and optimized inference frameworks enable multimodal AI workloads—including vison-language reasoning, Speech-to-Text (STT), and Text-to-Speech (TTS)—to run simultaneously on the NVIDIA Jetson Orin Nano 8GB platform. This showcases how software and system optimization can efficiently execute multiple AI models and reduce memory footprint on resource-constrained edge devices. Similar architectures are increasingly being adopted across smart manufacturing, autonomous robotics, and healthcare applications, supporting use cases such as production line inspection, intelligent vision systems, voice-enabled interaction, and autonomous decision-making. By performing AI inference directly on edge devices, organizations can achieve low-latency performance while enhancing data security and operational efficiency."Generative AI has entered the stage of real-world deployment," said Jeffrey Yu, CEO of EDOM Technology. "Today, the biggest challenge for enterprises is no longer finding AI models; it's successfully deploying AI into products and operational environments. The success of an AI project depends not only on hardware performance, but also on deployment efficiency, system integration capabilities, and overall return on investment. EDOM's role extends beyond supplying platforms and components. We help customers integrate NVIDIA technologies, open-source AI models, and system optimization strategies to accelerate time-to-market, reduce deployment risks, and enable AI to deliver measurable business value."Backed by years of expertise in Edge AI and embedded system integration, EDOM provides end-to-end solutions encompassing selections powered by NVIDIA technologies, open-source AI models integration and application optimization, hardware design, and system deployment. The company has successfully enabled Edge AI applications across smart manufacturing, autonomous robotics, healthcare, smart retail, and smart city initiatives. Looking ahead, EDOM will continue collaborating with NVIDIA and the open-source AI community to help enterprises overcome the challenges of deploying large AI models, accelerate the transition from proof of concept to large-scale deployment, and transform AI innovation into tangible value across the industry value chain.

Advanced Chip and Circuit Materials (ACCM) today introduced Celeritas SMC (Silicon-Matched Core), a production-ready core for advanced IC substrates engineered to match the in-plane coefficient of thermal expansion of silicon while running on established organic-substrate manufacturing lines. Available now from ACCM's Wisconsin facility, the material is designed for large-body chiplet packages, embedded-bridge architectures, and other advanced packages in which conventional organic cores face growing warpage and thermomechanical reliability limits.Celeritas SMC provides the low CTE, stiffness, dimensional stability, and electrical performance that have driven industry interest in glass core, without requiring through-glass vias, glass-specific metallization, brittle-panel handling, or wholesale line retooling.Addressing the Warpage Challenge and Silicon Die Bridges in Next-Gen AI PackagesAI and high-performance computing packages are growing faster than the materials beneath them. As package bodies exceed 100 mm per side and silicon bridge dies are embedded directly into the substrate, the thermal expansion mismatch between a conventional organic core at 12 to 16 ppm/degree C and silicon at roughly 3 ppm/degree C drives warpage, embedding stress, and interconnect fatigue that packaging engineers can no longer design around. The industry consensus is clear: the core must move toward the CTE of silicon.Glass core emerged as one route to that goal, and its target properties are real. Its commercialization, however, depends on glass-specific via formation, metallization, inspection, surface-preparation, and handling capabilities that are not broadly deployed in high-volume substrate manufacturing, together with substantial new capital investment across the supply chain. Celeritas SMC delivers the silicon-matched CTE, dimensional stability, and electrical performance sought from glass core through the manufacturing infrastructure the industry already operates.Silicon-Matched Properties Seamlessly Integrated into Existing LinesCeleritas SMC matches silicon from room temperature to 100 degree C and is tunable by construction, eliminating the bulk in-plane CTE mismatch between embedded silicon and the core. Vias are formed by standard mechanical and laser drilling. Build-up films bond directly using established lamination processes, without glass-specific surface preparation or adhesion-promotion steps. Panels handle, laminate, and singulate on the installed equipment base using the process fabricators already use.The property set extends well beyond CTE: a glass transition temperature above 300 degree C for compatibility with the full assembly and rework thermal budget, a high stiffness, a dissipation factor of 0.004 at 10 GHz with lower loss grades available for the most demanding signal integrity applications, moisture absorption of 0.1 percent at saturation, and core thicknesses from 100 to 1,200 µm and above, covering everything from thin coreless-adjacent constructions to thick high-rigidity platforms for the largest package bodies.Eliminating the Massive Capital Investment of Glass Cores"The industry does not have a low CTE problem, it has a low CTE at acceptable cost problem," said Tarun Amla, PhD, Founder, President, and CEO of ACCM. "Glass core asks the entire substrate supply chain to retool around through-glass vias before volume yields and economics have been established. Celeritas SMC gives packaging engineers silicon-matched CTE, high stiffness, and low loss on the equipment they already own, with the drilling, metallization, and lamination processes they already know. That is the difference between a roadmap concept and a material customers can put on their lines today."Optimizing Thermomechanical Stress Across the Packaging StackA low CTE package does not remove thermomechanical stress from the system; it relocates it. A package pulled down toward silicon CTE transfers the mismatch to the board-level interconnect, which becomes the new reliability bottleneck as package bodies grow and package-to-board integration tightens. ACCM is positioned on both sides of that interface: the same material platform behind Celeritas SMC enables printed circuit boards with tunable CTE from 2 to 10 ppm/degree C, allowing a graded CTE architecture from the die through the substrate to the system board. For embedded-bridge designs, CoWoS-L-class constructions, and other large-body chiplet packages, Celeritas SMC enables CTE management within the package substrate. For substrate-free architectures such as CoWoP, the broader ACCM material platform can provide tunable-CTE system-board materials, addressing the package-to-board interface where thermomechanical stress is otherwise concentrated.Securing Supply Chain Resilience with US-Based Manufacturing"Every panel of Celeritas SMC is manufactured in Wisconsin, on a site that has made advanced laminates for American electronics for decades," said Keshav Amla, Founder and COO of ACCM. "Customers bring us a stackup, we supply material and applications-engineering support, and their fabricator processes it on equipment already installed in the line. Domestic production, direct applications engineering support, and an initial process-verification path measured in weeks give our customers both speed and supply-chain resilience as they plan their next-generation packages."Celeritas SMC is available for customer evaluation immediately, in core thicknesses from 100 to 1,200 µm and above. ACCM also offers ultra-low-loss build-up films engineered to pair with the core for customers seeking a complete low-loss, CTE-managed dielectric stack. A technical datasheet and application notes for embedded die and large-body substrate constructions are available under NDA.For more details, you can read a companion FAQ on ACCM's website: ACCM Celeritas SMC FAQ or reach out for more information, a meeting, or to start an evaluation or program through ACCM's contact page: ACCM Materials Inquiry

The wave of factory automation is reshaping global manufacturing at an unprecedented pace.  To respond to the mounting pressure for higher efficiency, tighter quality control, and greater operational resilience, manufacturers across industries are accelerating investments in automation technologies.  Statistics show that the global factory automation market was valued at approximately $36.01 billion in 2024 and is projected to experience strong growth at a CAGR of 11.1% through 2030.However, modern production systems are no longer defined by isolated subsystems, but by tightly integrated architectures where motion control, sensing, compute, and connectivity must operate as a coordinated whole.  As a result, "system complexity" is becoming the defining challenge for engineering teams.In response to this trend, Arrow Electronics, a leading global technology solutions provider, supports this transition by combining broad component access with system-level engineering and integration expertise—helping customers move more efficiently from concept to deployment.Converging trends redefining new-generation smart factoriesToday, several global trends are driving the transition to interconnected, data-driven production environments.  Manufacturers are no longer satisfied with the operation of single components; instead, they require seamless interoperability across motors, drives, sensors, cameras, and industrial networks—often from multiple vendors.  These requirements span core automation subsystems including machine vision, motion control, power management, thermal, and industrial networking architectures.Particularly in the field of AI-enabled automation, machine vision and edge-based analytics are increasingly used for defect detection, dimensional measurement, and robot guidance, which places extremely high demands for high data throughput and low-latency processing.  Market data reflects the growing importance of imaging technologies.  The global machine vision market is expected to grow from $20.4 billion in 2024 to $41.7 billion by 2030, at a CAGR of around 13%, with Asia-Pacific accounting for over 40% of demand.  At the same time, the global servo motor market is projected to expand from $13.5 billion in 2024 to $20.1 billion by 2030, reflecting steady adoption of high-performance motion control technologies across industrial applications.On the other hand, global labor shortages, rising wages, and environmental regulations for energy efficiency and carbon reduction are pushing manufacturers to place greater scrutiny on system-level efficiency and architecture redesign across the factory floor.  Faced with such a complex multi-supplier ecosystem, manufacturers are urgently seeking partners capable of providing cross-domain technologies.  Arrow's role is to bring together these technologies into validated and scalable system architectures.Five major engineering challenges on the factory floorBehind the pursuit of high performance, engineering teams often face enormous design challenges on the factory floor:Motion control: Systems must deliver high-speed servo performance under vibration, load variation, and continuous operation, and the multi-axis synchronization systems adds another layer of complexity.Machine vision: Systems must maintain consistent inspection under variable lighting conditions, and the large volumes of data generated by high-resolution 2D and 3D imaging require real-time processing to keep pace with production throughput.Power and thermal management: Dense electronic systems operating under continuous duty cycles must balance efficiency, protection, and heat dissipation within compact form factors.Networking architectures: Networks must support deterministic communication and handle high-bandwidth data streams from sensors and cameras to ensure low-latency and high-reliability decision-making at the edge or in the cloud.Safety and operational risks: Failure of any subsystem to meet safety and compliance requirements can result in costly downtime and operational risks.To address these challenges, manufacturers have gradually shifted from component-level optimization to system-level design.  Modern factory automation solutions demand tightly integrated architectures where sensing, motion, power, and compute are designed together from the outset.  Early validation of system interactions—such as synchronization between vision and motion, thermal performance under peak loads or real-time communication requirement of industrial network —can significantly reduce downstream integration risk and shorten deployment timeline.System-level design – The key to competitive advantageIn response to this trend, Arrow Electronics leverages its extensive component portfolio, system-level engineering capabilities, and integration expertise to help customers accelerate the implementation of automation solutions, from proof of concept to practical deployment.  Arrow Electronics' unique strengths enable it to help customers achieve this goal, with comprehensive support including:Full BOM coverage and broad component portfolio: Providing complete supply spanning sensing, motion, power, and compute components.Proven reference designs and architectures: Offering reliable architectures for critical industrial applications such as machine vision and motor control.Extensive application engineering support: Our expert team's expertise covers motor control, power electronics (including SiC and IGBT technologies), edge AI platforms, and industrial networking technologies.Pre-validated integration frameworks: Effectively linking a complete ecosystem of multiple suppliers to enable seamless integration.Hoffei Hou, vice president of Arrow's Sales Components business for Taiwan, stated: "As factory automation continues to deepen, the real challenge is no longer the introduction of a single technology, but how to integrate diverse technologies such as machine vision, motion control, sensing, and connectivity to achieve a highly efficient and scalable system architecture.  With system-level design integration capabilities, engineering support, and complete supply chain services, Arrow Electronics helps customers accelerate the process from design verification to large-scale deployment, turning technological complexity into a competitive advantage."Turning complexity into advantage: System integration becomes a key capability for manufacturingToday, as automation technologies become increasingly prevalent, manufacturers’ competitive advantage is no longer whether to adopt advanced technologies, but how to integrate them effectively to maximize overall benefits.  With the rapid evolution of diverse technologies such as machine vision, intelligent sensing, motion control, energy management, and industrial networks, manufacturers are increasingly valuing partners with cross-domain integration capabilities and industry experience.  This helps them shorten adoption cycles, enhance system stability, and reduce operational risks.Looking ahead, factory automation systems will become more interconnected and sophisticated.  By combining expertise in design engineering, embedded software validation, and supply chain management, along with localized engineering support, manufacturers can more effectively implement innovative technologies into practical applications.  This will continuously improve production efficiency, operational resilience, and market responsiveness, accelerating the realization of digital transformation results.