Innodisk Corporation - Experts & Thought Leaders

The showcase combines AI compute platforms, industrial storage, connectivity, sensing, and software technologies to support scalable and secure AI deployment across enterprise and industrial applications. Innodisk, a global AI solution provider, will showcase its complete edge AI ecosystem at COMPUTEX 2026, demonstrating how enterprises and industrial customers can accelerate AI deployment for production-ready applications. Built across five essential layers—compute, memory, storage, sensing and communication, and software—the portfolio brings together AI platforms, industrial modules, and real-world application demos to support secure, scalable AI adoption. Secure enterprise AI deployment At the centre of this year’s showcase are Innodisk’s on-premises AI solutions for secure enterprise AI deployment. AccelBrain, powered by the APEX-X200, enables fully on-premises deployment of open-source LLMs with 100% data sovereignty. Complementing it, AccelTune provides a no-code LLM fine-tuning experience, demonstrated on the APEX-S100 featuring Intel® Xeon® 6700-series processors and supporting dual NVIDIA RTX PRO GPU. Together, these solutions help enterprises build secure, application-specific LLM workflows without relying on public cloud infrastructure. Enhance operator awareness Beyond enterprise AI, Innodisk will demonstrate how edge AI can be deployed in demanding industrial environments. A heavy machinery safety solution showcases eight ruggedised GMSL2 camera modules rated IP67 and IP69K, delivering AI-driven surround-view stitching, Driver Monitoring System (DMS), and Blind Spot Detection (BSD) to enhance operator awareness and safety in harsh field applications. Innodisk, an Intel Gold Partner, will also present the APEX-E400 edge AI system, powered by Intel® Core™ Ultra Series 3 processors. Built on a heterogeneous CPU-GPU-NPU architecture, the system supports up to 16 simultaneous streams with parallel AI model inference. A live Intel OpenVINO™ demo will highlight adaptive workload distribution across the tri-engine architecture for optimised edge AI performance. Automatic emergency braking In collaboration with Qualcomm Technologies, Inc., Innodisk continues to expand its AI on Qualcomm Dragonwing™ series for compact, power-efficient edge inference, showcasing the latest Dragonwing processors, across the IQ9, IQ10 and IQ-X Series, in the Innodisk booth. Powered by the Dragonwing IQ-9075 processor-based COM-HPC Mini platform, there will be live demos showcasing AI screw thread inspection and multi-stream vision processing, while an AMR demonstration further integrates GMSL and depth cameras for spatial awareness, ROI safety zone detection, and PWM-triggered automatic emergency braking. Automatic emergency braking Software enablement is another key layer of the ecosystem. Innodisk integrates Edge Impulse’s end-to-end edge AI machine learning operations (MLOps) platform with iCAP, Innodisk’s intelligent cloud management platform, to demonstrate Agentic AI capabilities for automated remote model updates and deployment. Supporting these AI systems is Innodisk’s industrial-grade memory, storage, connectivity, and sensing portfolio. The company will showcase next-generation DDR5 8000 RDIMM, CUDIMM, and CSODIMM, along with industry-pioneer 12800 MRDIMM and a new CXL Add-in Card (AIC). Storage highlights include Data center SSDs in EDSFF and U.2 form factors, as well as 218-layer 3D TLC SSDs. Ruggedised camera solutions In connectivity, Innodisk debuts the ELPL-82F1 SFP28 25GbE LAN card, alongside the award-winning EGPL-T2F1—honoured at Embedded World 2026 as the world’s first SFP+ LAN card in M.2 form factor. The booth will also feature ruggedised camera solutions across GMSL2, MIPI, and USB interfaces, complemented by an interactive OCR container identification demo running on an integrated edge system. Together, these demonstrations show how Innodisk brings together its complete five-layer AI ecosystem, helping customers deploy AI securely, reliably, and at scale across enterprise and industrial environments. Experience Innodisk’s advanced edge AI solutions at Booth No. K0106, Taipei Nangang Exhibition Center, Hall 1.

Innodisk, a global AI solution provider, announces its participation in NVIDIA GTC 2026. Alongside subsidiary Aetina, Innodisk Group is showcasing the seamless integration of NVIDIA AI architectures into scalable, deployable edge AI applications. By converging its high-performance computing platforms and the proprietary vision sensing camera modules, Innodisk is translating cutting-edge technology into field-ready industrial AI solutions across two verticals: smart healthcare and intelligent mobility. Hardware-software integration As the global AI landscape shifts from cloud-based models to edge execution, Innodisk delivers the hardware-software integration depth that industrial-scale edge AI adoption demands. "NVIDIA technologies are driving the AI revolution, and Innodisk is where that innovation meets industry reality," said Randy Chien, Chairman of Innodisk Group. "Through close collaboration with NVIDIA and two decades of vertical market expertise, we help customers cut through integration complexity and accelerate their journey from advanced algorithms to field-ready solutions." Space-constrained clinical settings Innodisk is demonstrating a Medical Multimodal Vision Language Model (Medical VLM) powered by NVIDIA Blackwell architecture and NVIDIA TensorRT, running on the APEX-X200 Edge AI Computing Platform. Housed in a 16.5-liter chassis, the system integrates an NVIDIA RTX PRO 6000 Blackwell Server Edition GPU (24,064 CUDA cores; 752 Tensor cores), bringing enterprise data-centre-level compute to space constrained clinical settings. By enabling fully local inference to eliminate cloud dependency, the solution ensures ultra-low latency and absolute data privacy. Improving clinical communication Live demos show the system analyzing X-ray and CT images in real time, auto 1/2 generating diagnostic report drafts, and converting medical findings into patient-friendly explanations — streamlining radiologist workflows and improving clinical communication. Innodisk and Aetina have jointly developed an Intelligent Mobility Solution powered by NVIDIA Jetson AGX Orin platform. By leveraging Innodisk’s custom capture card, the team integrated eight GMSL2 camera modules on a single platform, overcoming hardware integration challenges to deliver real-time multi-channel video. The solution features Surround-View Stitching and active Blind Spot Detection (BSD), while an integrated AI Driver Monitoring System (DMS) alerts against fatigue, distraction, and mobile phone use, delivering comprehensive safety coverage for heavy-duty vehicle operations. Comprehensive safety coverage Visit Innodisk at NVIDIA GTC 2026 Join Innodisk at Booth #7039 and Aetina at Booth #139, San Jose McEnery Convention Center, March 16–19. Explore the full portfolio of edge AI infrastructure engineered for demanding workloads. Event Details: Dates: March 16–19, 2026       Venue: San Jose McEnery Convention Center, San Jose, CA       Innodisk Booth: #7039 │ Aetina Booth: #13

The joint solution combines Innodisk’s APEX-P200 edge AI system, RealSense depth cameras, and EverestLabs’ RecycleOS platform to enable accurate material identification and sorting directly on recycling facility production lines. Innodisk, a global AI solution provider, announces a strategic collaboration with RealSense and EverestLabs—a pioneer in AI-powered recycling automation—to deliver a scalable, real-time AI vision solution engineered for demanding conditions of industrial recycling environments. Industrial recycling environments The joint solution integrates Innodisk’s rugged APEX-P200 edge AI system with RealSense’s depth camera and EverestLabs’s RecycleOS platform, enabling accurate material identification and sorting directly on the plant floor—where split-second decisions, environmental resilience, and continuous uptime are critical. Material Recovery Facilities (MRFs), recycling operators, and waste management companies face mounting pressure to improve recovery rates, material quality, and labour efficiency while meeting stringent regulatory requirements. Traditional sorting methods struggle with the volume and complexity of mixed waste streams, resulting in contamination, lower material value, and operational inefficiencies. Multi-camera 3D analysis The joint solution integrates Innodisk’s rugged APEX-P200 edge AI system with RealSense’s depth camera At the heart of the solution is the Innodisk APEX-P200, an industrial-grade edge AI platform engineered to withstand harsh recycling conditions, including constant vibration, dust, variable lighting, and temperature fluctuations. Powered by a 13th-Gen Intel® Core™ i7 processor and an NVIDIA RTX 2000 Ada GPU delivering up to 120 INT8 TOPS, the platform provides the high-performance compute required for real-time, multi-camera 3D analysis. By processing AI inference at the edge, the solution enables immediate decision-making without cloud dependency, helping facilities maintain throughput while improving material recovery and purity. "Advancing sustainable recycling requires an edge AI solution that provides reliability in real-world industrial conditions," said Victor Le, President of Innodisk USA Corporation. "Working with EverestLabs and RealSense, we're delivering platforms that run continuously on production lines, helping facilities improve both recovery rates and environmental impact." Real-time sorting decisions RealSense’s depth camera provides factory-calibrated 3D depth with aligned colour imaging, enabling precise measurement and segmentation of mixed, overlapping materials moving at conveyor speeds. As these high-precision imaging data streams require high-performance and reliable computing, Innodisk’s APEX-P200 serves as the ideal engine for the task. This combination of spatially accurate data and robust processing significantly enhances AI model performance, supporting consistent, real-time sorting decisions. “Getting recycling right comes down to reliable perception in harsh, high-speed conditions,” said Nadav Orbach, CEO of RealSense. “Together with Innodisk and EverestLabs, we’re bringing production-grade 3D sensing to the line so operators can increase purity and recovery without slowing throughput.” AI-powered recycling systems EverestLabs integrates the APEX-P200 and RealSense technology into its AI-powered recycling systems EverestLabs integrates the APEX-P200 and RealSense technology into its AI-powered recycling systems, deploying advanced Vision AI models that help Material Recovery Facilities improve operational efficiency, reduce contamination, and scale AI-driven automation across facilities. “Our systems operate in some of the world’s toughest industrial environments and require embedded-grade reliability, high-accuracy Physical AI, and real-time responsiveness,” said Ravichandra Malapati, Principal Engineer at EverestLabs. “Innodisk’s APEX-P200 meets all three requirements.” This collaboration reflects Innodisk's strategy of partnering across the value chain—from vertical market specialists to advanced sensing innovators. With flexible edge AI platforms and deep integration expertise, Innodisk enables partners to rapidly deploy production-ready solutions tailored to industry-specific requirements at scale