Artificial intelligence dominated COMPUTEX 2026.
From AI-powered PCs and next-generation GPUs to humanoid robots and intelligent automation platforms, virtually every major exhibitor had an AI story to tell. Yet among the many AI showcases spread across the exhibition halls, Innodisk stood out for a different reason.
Rather than focusing solely on AI models or computing hardware, Innodisk showcased an entire ecosystem designed to help organizations deploy AI in the real world.
Under the banner of its Five-Layer Edge AI Ecosystem, Innodisk demonstrated how memory, storage, compute, sensing, connectivity, and software can work together to transform AI concepts into production-ready deployments. The result was one of the most comprehensive industrial AI showcases at COMPUTEX 2026.
Who is Innodisk?
Founded in Taiwan in 2005, Innodisk has established itself as one of the world’s leading providers of industrial-grade memory and storage solutions.
For years, the company has been known for supplying mission-critical storage, DRAM, and embedded computing components used in industrial automation, transportation systems, enterprise infrastructure, and edge computing environments. Today, Innodisk holds the largest market share in industrial storage solutions and ranks among the world’s leading memory module providers.
In recent years, however, the company has expanded far beyond memory and storage.
At COMPUTEX 2026, Innodisk positioned itself as an AI solution provider capable of delivering complete end-to-end edge AI ecosystems spanning hardware, software, sensing technologies, and deployment platforms.
The Five-Layer Edge AI Ecosystem
The centrepiece of Innodisk’s COMPUTEX 2026 showcase is its Five-Layer Edge AI Ecosystem.
Rather than treating AI as a standalone software capability, Innodisk approaches deployment through five interconnected layers:
- Compute
- Memory
- Storage
- Sensing & Communication
- Software
The company argues that successful AI deployment requires all five layers to work together seamlessly. Without reliable sensing, sufficient compute power, optimized storage, and robust software management, many AI projects struggle to move beyond proof-of-concept stages.
Through a combination of live demonstrations and production-ready platforms, Innodisk showed how these layers integrate into complete AI solutions designed for enterprise and industrial environments.
Building Private AI with Full Data Sovereignty
One of the strongest themes throughout the Innodisk booth was enterprise AI deployment.
While many organizations are exploring generative AI, concerns surrounding data privacy, compliance, and intellectual property continue to slow adoption.
To address these challenges, Innodisk showcased AccelBrain, an on-premises AI platform powered by its APEX-X200 system. The solution enables organizations to deploy open-source large language models entirely within their own infrastructure, ensuring complete data sovereignty without relying on public cloud services.
Complementing this is AccelTune, a no-code platform designed to simplify LLM fine-tuning and customization. Running on the APEX-S100 platform featuring Intel Xeon 6700-series processors and support for dual NVIDIA RTX PRO GPUs, the system allows enterprises to create application-specific AI models without requiring extensive machine learning expertise.
Together, these platforms demonstrate Innodisk’s commitment to making enterprise AI both practical and secure.
AI-Powered Industrial Safety
Beyond enterprise applications, Innodisk also showcased several industrial AI solutions designed for deployment in harsh operating environments.
One particularly impressive demonstration focused on heavy machinery safety.
The system utilizes eight ruggedized GMSL2 camera modules rated to IP67 and IP69K standards, providing AI-powered surround-view stitching, blind spot detection, and driver monitoring capabilities. By combining multiple camera feeds with AI analytics, operators gain significantly improved situational awareness, helping reduce risks in demanding industrial environments.
The demonstration highlights a growing trend within industrial automation: using AI not only to improve efficiency but also to enhance workplace safety.
AI on Qualcomm Dragonwing
Another major focus at COMPUTEX 2026 was Innodisk’s rapidly expanding collaboration with Qualcomm Technologies.
Building upon the AI on Dragonwing initiative launched earlier this year, Innodisk showcased one of the most comprehensive Dragonwing portfolios currently available. The lineup spans multiple processor generations and performance tiers, delivering AI capabilities ranging from 20 TOPS to an impressive 700 TOPS.
The portfolio includes:
Dragonwing IQ8 Series
Designed for vision inference and sensor fusion applications, the IQ8-based EXMP-Q801 COM-HPC Mini module delivers up to 40 TOPS while supporting wide-temperature industrial environments.
Dragonwing IQ9 Series
Supporting up to 100 TOPS through dual NPUs, the IQ9 powers modules, starter kits, and fanless edge AI systems aimed at industrial inspection, driver monitoring systems, and multi-stream analytics.
Dragonwing IQ10 Series
The flagship IQ10 platform pushes performance up to 350 TOPS and is engineered for demanding workloads such as autonomous mobile robots, humanoid robotics, vision-language models, and advanced edge AI deployments.
Dragonwing IQ-X Series
Perhaps the most interesting addition is the new IQ-X platform, which introduces native Windows compatibility alongside Ubuntu Linux support. Designed for PLCs, HMIs, and industrial edge controllers, the platform helps simplify deployment within existing enterprise environments.
Live AI Demonstrations
Innodisk reinforced its hardware announcements with several live demonstrations that showcased real-world deployment scenarios.
One demo executed multiple AI workloads simultaneously, including:
- CNN-based defect inspection
- Driver Monitoring Systems (DMS)
- Vision-Language Model PPE detection
Remarkably, the entire system operated within a 30W power envelope while achieving an average inference latency of just 4.4 milliseconds. According to Innodisk, this delivers a 2.6x performance advantage compared to traditional architectures while significantly improving performance-per-watt efficiency.
A second demonstration focused on autonomous robotic navigation using vision-based sensing, depth cameras, and real-time obstacle avoidance. Designed for autonomous mobile robots and automated guided vehicles, the system provided 360-degree environmental awareness while operating at only 32W of power consumption.
Memory and Storage Continue to Evolve
Despite its growing focus on AI, Innodisk has not forgotten its roots.
The company also unveiled an extensive range of next-generation memory and storage products, including:
- DDR5 8000 RDIMM
- CUDIMM
- CSODIMM
- 12800 MRDIMM
- CXL Add-In Cards
- Data Centre SSDs in EDSFF and U.2 form factors
- 218-layer 3D TLC SSDs
These products provide the foundational infrastructure required to support increasingly demanding AI workloads and enterprise applications.
Connectivity and Sensing Innovation
The sensing and communication layer of Innodisk’s ecosystem also received significant attention.
Among the highlights was the debut of the ELPL-82F1 SFP28 25GbE LAN card, alongside the award-winning EGPL-T2F1, recognized as the world’s first SFP+ LAN card in M.2 form factor. The company also showcased a variety of industrial camera solutions spanning GMSL2, MIPI, and USB interfaces, all designed to support modern machine vision and AI deployments.
These technologies play a crucial role in ensuring that AI systems can efficiently gather and process real-world data.
Final Thoughts
While many companies at COMPUTEX 2026 showcased AI hardware, Innodisk focused on something arguably more important: deployment.
The company’s Five-Layer Edge AI Ecosystem demonstrates a deep understanding of the practical challenges organizations face when moving AI from laboratory experiments into real-world applications.
From private enterprise LLM platforms and industrial safety systems to Qualcomm Dragonwing edge AI deployments and autonomous robotics, Innodisk presented a vision of AI that is secure, scalable, and production-ready.
As AI continues to move closer to where data is generated, the importance of edge computing will only increase.
And if COMPUTEX 2026 is any indication, Innodisk intends to be one of the companies helping build that future.
