According to a report from Liberty Times Net, IEI Integration (威強電, TWSE:3022) management used AMD's (超微) Advancing AI 2026 conference this week to release two new offerings to the market: the NANO-X100 single-board computer, built around AMD's Ryzen AI Embedded X100 series processor, and the company's new "AiEi" edge AI software platform, unveiled at the same event (E1). The report is dated July 24, 2026.
Per the same Liberty Times Net report, the NANO-X100's onboard NPU delivers AI inference performance of up to 50 TOPS, which the report says supports machine vision, image analysis, and other AI inference tasks at the edge (E2). On the hardware side, the platform can be configured with up to 128GB of memory and offers dual 2.5GbE Ethernet ports, USB 3.2, and M.2 expansion slots, which the report notes are intended to accommodate camera connections and various embedded AI applications (E3).
| Spec | Value | Source |
|---|---|---|
| NPU AI inference performance | Up to 50 TOPS | E2 |
| Maximum memory configuration | 128GB | E3 |
| Networking | Dual 2.5GbE | E3 |
| Expansion | USB 3.2, M.2 | E3 |
According to Liberty Times Net, IEI's AiEi edge AI software platform provides real-time image analysis, virtual fencing, and abnormal behavior detection through a centralized management interface. The report states that users can quickly set monitoring conditions and trigger corresponding alerts, with notifications supported via email and SMS (E4).
The same report indicates that IEI plans to expand AiEi going forward by connecting to third-party Vision Language Models (VLM), and by supporting Vision-Language-Action (VLA) models integrated with robotic systems, with the stated aim of broadening the platform's edge AI application scope (E5).
According to a separate report from UDN Money dated July 26, 2026, IEI teamed up with Yu Shan Technology (友上科技) to demonstrate an enhanced AI-recognition collaborative robot solution that integrates an AMD chip, described in the report as suited for human-machine collaborative environments (E7).
The UDN Money report states that AMD, citing its confidence in the outlook for physical AI, announced the Kria AI solutions along with an open robotics partner network at Advancing AI 2026 (E6). Per the report, the Kria AI product portfolio includes a Kria AI System-on-Module (SOM) and a robotics carrier board, both driven by the Ryzen AI Embedded X100 series processor — the same processor family used in IEI's NANO-X100 — as well as a separate Kria AI Robotics Development Platform (E8).
According to AMD's statement as reported by UDN Money, the Kria AI SOM is expected to begin shipping from ODM partners starting in the fourth quarter of this year. The Kria AI Robotics Development Platform has already been sampled to early customers and is likewise expected to ship in the fourth quarter (E9).
The two reports, taken together, show IEI's NANO-X100 and AMD's newly announced Kria AI SOM both built on the same Ryzen AI Embedded X100 series processor (E2, E8), and IEI's Yu Shan robotics demonstration (E7) lines up with AMD's parallel announcement of an open robotics partner network (E6) at the same conference. On the software side, IEI's stated roadmap to add VLM and VLA support to AiEi (E5) points toward the same robotics and multimodal direction that AMD's Kria AI Robotics Development Platform is targeting for fourth-quarter shipment (E9), though the two reports do not state that IEI's roadmap and AMD's shipping schedule are formally linked.
According to a report by technews.tw, the EU AI Act's transparency obligations take effect on August 2, 2026, while high-risk system rules enter a critical implementation phase the same month. The report notes that ambiguous provisions are forcing developers to interpret rules themselves, prompting users to shop around for models with looser restrictions.
According to a report by Liberty Times Net (ltn.com.tw), the Taipei Exchange (TPEx) disclosed that GlobalWafers (環球晶) triggered a settlement default of NT$11.03 million, marking the 11th case this year to reach TPEx's default disclosure threshold, following notifications from First Securities' Chiayi branch and Cathay Securities' Taichung branch.
According to a report from Inside (inside.com.tw), NVIDIA and South Korea's SK Group have built out a chip and AI-infrastructure partnership pegged at over $500 billion, covering memory procurement, AI supercomputers, and data-center construction. NVIDIA CEO Jensen Huang, cited by both Inside and CNYES (news.cnyes.com), forecasts the global semiconductor industry could expand to 10 times its current size within a decade.