LITEON Technology (TWSE: 2301) is showcasing its expanding artificial intelligence infrastructure portfolio at electronica 2026 in Munich, Germany, from November 10–13. Under the theme “LITEON Powers AI,” the company is presenting technologies spanning high-density AI data center power delivery, advanced connectivity, 5G networking, and Edge AI. The portfolio includes power shelves supporting NVIDIA’s next-generation AI infrastructure architectures, energy backup systems, near-chip power conversion, and intelligent edge computing solutions designed for industrial and enterprise applications.
LITEON Technology (TWSE: 2301) is showcasing its expanding artificial intelligence infrastructure portfolio at electronica 2026 in Munich, Germany, from November 10–13. Under the theme “LITEON Powers AI,” the company is presenting technologies spanning high-density AI data center power delivery, advanced connectivity, 5G networking, and Edge AI. The portfolio includes power shelves supporting NVIDIA’s next-generation AI infrastructure architectures, energy backup systems, near-chip power conversion, and intelligent edge computing solutions designed for industrial and enterprise applications.
The showcase reflects a growing challenge for the technology industry: as artificial intelligence models become more computationally demanding, the infrastructure supporting them must deliver greater electrical capacity, improved efficiency, and reliable connectivity.
AI Data Centers Drive Demand for High-Density Power Infrastructure
One of LITEON’s major exhibition areas focuses on power systems designed for NVIDIA MGX and Open Compute Project (OCP) Open Rack V3 architectures.
As AI accelerators and large language models increase computing requirements, conventional server power infrastructure is facing new limitations. Modern AI systems require more electricity per rack, creating additional demands on power distribution, voltage conversion, backup systems, and thermal management.
LITEON is addressing these requirements through a portfolio covering power delivery from rack-level distribution to individual computing components.
The company is demonstrating a 110kW Power Shelf designed to support the NVIDIA Vera Rubin NVL72 architecture. The system incorporates active current sharing and input current smoothing to maintain power stability under changing AI workloads.
It is also presenting a 72kW Power Shelf in a compact 1U form factor, targeting high-output power delivery while conserving valuable rack space.
Another major highlight is LITEON’s demonstration of NVIDIA’s 800 VDC power architecture. Through an interactive virtual rack, the company is illustrating configurations of 330kW, 880kW, and 1.2MW.
The demonstrations highlight how data center power distribution could evolve toward megawatt-scale infrastructure as next-generation AI systems increase rack power density.
Moving Power Conversion Closer to AI Processors
Beyond rack-level power distribution, LITEON is showcasing technologies designed to improve power delivery closer to computing components.
Its portfolio includes Capacitor Backup Units (CBUs), Intermediate Bus Converters (IBCs), Voltage Regulator Modules (VRMs), and high-efficiency Power Bricks.
These technologies address different stages of the power delivery chain. Backup units provide additional energy support during power disturbances, while converters and voltage regulators help transform and deliver electricity at the levels required by servers and processors.
Reducing unnecessary conversion losses is increasingly important in high-density AI environments, where even relatively small efficiency improvements can have significant implications for power consumption and cooling requirements.
This component-to-system approach positions power electronics as a critical part of AI infrastructure rather than a supporting subsystem.
Extending AI Infrastructure From Cloud to Edge
LITEON’s exhibition extends beyond hyperscale data centers to enterprise networks, industrial computing, and edge applications.
Its power portfolio includes 800W and 2,000W Common Redundant Power Supplies (CRPS), high-power Power over Ethernet (PoE) injectors, DIN Rail power supplies ranging from 12W to 960W, industrial AC/DC and DC/DC power solutions, and medical-grade power systems.
These products support environments where reliability, continuous operation, and compact power delivery are important.
On the connectivity side, LITEON is presenting its Connected AI Edge ecosystem, including Open Radio Access Network (O-RAN) radio units supporting FR1 and FR2 frequency bands.
The company is also demonstrating an AI-RAN solution built on NVIDIA AI Aerial, a platform designed to support AI-enabled telecommunications infrastructure.
The convergence of artificial intelligence and radio access networking could enable telecommunications operators to use computing resources more efficiently while introducing AI capabilities into network operations.
LITEON is also introducing its Fixed Wireless Access (FWA) solution following its integration of U-MEDIA. The broader connectivity portfolio includes all-in-one small cells, Ethernet switches, and Wi-Fi access points.
Together, these technologies target enterprise networking, wireless broadband, and industrial connectivity requirements.
Edge AI Brings Real-Time Intelligence to Physical Environments
Another part of the showcase focuses on Edge AI Vision and intelligent computing.
LITEON is presenting AI-powered security cameras and AI Boxes capable of performing local AI inference. Rather than sending every piece of data to a remote cloud environment, edge devices can process information closer to where it is generated.
This approach can reduce dependence on continuous cloud connectivity, support faster responses, and help manage data transmission requirements.
Potential applications include smart buildings, healthcare facilities, retail environments, and industrial automation.
For example, AI-enabled cameras can support real-time visual monitoring, while local inference systems can analyze operational conditions and trigger responses without relying entirely on centralized computing resources.
The combination of power electronics, connectivity, and edge computing is particularly relevant to industrial environments, where latency, network availability, and operational reliability can directly affect business processes.
Why Megawatt-Scale AI Infrastructure Matters
The broader technology industry is increasingly focused on the electrical and thermal requirements of AI data centers.
A McKinsey analysis published in 2026 estimates that AI-focused data center racks, which previously required less than 50kW, are moving toward power demands of 200kW to 600kW. Some configurations could approach approximately 1MW per rack later this decade.
These changing requirements help explain the industry’s interest in high-voltage direct-current architectures, advanced power conversion, and more integrated energy management systems.
NVIDIA’s MGX ecosystem and Vera Rubin platform are part of this wider transition toward rack-scale AI infrastructure. Companies such as LITEON are seeking opportunities across the power delivery chain, from rack-level systems to the components that supply individual processors.
The shift also creates opportunities for suppliers working in data center power, networking, optical connectivity, cooling, and energy management.
LITEON’s Strategy Connects Power, Connectivity, and Computing
LITEON’s electronica 2026 showcase demonstrates how the company’s technology portfolio is evolving alongside the requirements of AI infrastructure.
Rather than focusing exclusively on power supplies or individual networking products, the company is bringing together power management, telecommunications, intelligent sensing, and edge computing capabilities.
The strategy reflects a broader industry trend: AI deployment increasingly depends on coordinated infrastructure that can support computing from centralized cloud data centers to distributed edge environments.
For data center operators, the priorities include power density, efficiency, resilience, and scalability. For enterprise and industrial customers, connectivity and local inference are equally important.
LITEON’s challenge will be to translate its broad portfolio into commercially competitive systems as demand for AI infrastructure grows.
Its electronica 2026 demonstration provides a view of how power electronics and connectivity technologies are becoming increasingly central to the next generation of artificial intelligence deployment.
Market Landscape
The AI infrastructure market is expanding beyond GPUs and computing accelerators. Power delivery, energy management, connectivity, and cooling are becoming equally important components of next-generation AI systems.
Five important industry trends
- 800 VDC power architecture: Higher-voltage DC distribution is emerging as an approach to supporting increasingly power-intensive AI racks.
- Megawatt-scale AI infrastructure: Future AI computing environments will require more sophisticated power delivery and energy management.
- Near-chip power conversion: Efficient voltage regulation is becoming increasingly important as AI processor power requirements rise.
- AI-RAN and 5G integration: Telecommunications infrastructure is evolving to combine AI computing with wireless networking.
- Edge AI computing: Local inference enables intelligent applications to analyze information closer to where data is generated.
The OECD’s 2026 Digital Economy and Society Outlook also identifies high power density, advanced cooling, and energy efficiency as major infrastructure challenges for AI-ready data centers.
Top Insights
- LITEON is demonstrating 110kW and 72kW power shelves designed for high-density AI data center infrastructure and NVIDIA computing architectures.
- Its 800 VDC virtual rack showcases 330kW, 880kW, and 1.2MW configurations for emerging megawatt-scale AI infrastructure.
- Near-chip power technologies, including IBCs, VRMs, and Power Bricks, target efficient electricity delivery to high-performance AI processors.
- The Connected AI Edge portfolio combines O-RAN radio units, NVIDIA AI Aerial, FWA, networking, and industrial connectivity.
- Edge AI Boxes and intelligent cameras bring local AI inference to healthcare, retail, smart buildings, and industrial automation.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
