MSI unveils Nvidia-based AI push across cloud to edge
Tue, 2nd Jun 2026 (Today)
MSI unveiled an enterprise artificial intelligence strategy at Computex 2026, built around a cloud-to-edge product lineup spanning data centre systems, deskside machines and edge deployments.
The announcement focused on infrastructure based on Nvidia platforms, including MGX, DGX Station and DGX Spark designs, alongside industrial hardware and electric vehicle charging products for business and public sector users.
Data centre push
At the data centre end, MSI introduced the CG681-S6093, a 6U liquid-cooled AI server built on Nvidia MGX. The system supports dual AMD EPYC processors and up to eight Nvidia RTX PRO 6000 Blackwell Server Edition liquid-cooled GPUs.
Designed for large-scale AI inference, the server can be deployed in rack-scale configurations with up to four GPU systems in a 48RU setup. Networking in that configuration uses Nvidia Spectrum-4 SN5600 Ethernet switches and SN2201 out-of-band switches, while the server itself includes Nvidia ConnectX-8 SuperNICs with up to 8x400Gbps Ethernet connectivity.
The focus on liquid cooling reflects broader pressure on suppliers to manage heat and power use as AI workloads expand in corporate and hyperscale environments. Across the server market, vendors have been moving to denser designs to support graphics-heavy systems without sharply increasing rack footprints.
Desk-side systems
MSI also presented the XpertStation WS300, a desktop AI supercomputer based on Nvidia DGX Station. The machine uses Nvidia's GB300 Grace Blackwell Ultra Desktop Superchip and is intended for local AI development, model fine-tuning and inference.
According to MSI's specifications, the WS300 offers up to 748GB of coherent memory and 7.1TB/s of HBM3e bandwidth. It also includes dual 400GbE networking through Nvidia ConnectX-8 SuperNICs and uses a compact liquid-cooled design.
The desk-side category has drawn growing attention as companies look for ways to let developers and research teams work on larger models locally rather than relying entirely on remote compute resources. That can reduce latency during experimentation and gives enterprises another option between centralised data centre systems and lightweight edge devices.
Edge deployment
For edge computing, MSI introduced the EdgeXpert AI Supercomputer based on the Nvidia DGX Spark platform. The system is aimed at enterprises deploying AI agents and applications in operational settings outside traditional data centres.
MSI tied several software and solution demonstrations to EdgeXpert. These included the OpenClaw and Hermes agent frameworks; a system described as compliant with the European Cyber Resilience Act through a multi-agent architecture; a legal AI suite; a smart campus patrol vehicle using computer vision; and a robotics-led generative AI demonstration with Reachy Mini.
The edge strategy also extended into sector-specific systems. In manufacturing and semiconductors, MSI highlighted the Edge AI Box MS-C910E for machine vision and the MS-C939 for automated optical inspection, developed with Qiming Tech. In voice applications and transport, the Slim Box MS-C926 was shown for real-time translation, while the Embedded Box MS-C927 was presented for voice-based retail ordering and the in-vehicle MS-C932 for driver fatigue detection.
MSI also outlined a broader transport portfolio that includes fleet management tablets, telematics boxes, smart rear-view mirrors, and AI-enabled advanced driver assistance and driver monitoring systems. In agriculture, it presented a package combining autonomous drones, edge computing, rugged tablets, connectivity modules and multi-drone management for inspection, spraying, crop monitoring, and pest or disease identification.
Industrial hardware
Alongside its AI systems, MSI showcased a wider industrial computing range. That included the MS-NE21 rugged industrial tablet, built around Intel's 13th Gen Core i Series processors, with support for up to 32GB of memory and 2TB of PCIe SSD storage.
According to MSI, the tablet meets MIL-STD-810G standards, carries an IP65 rating, survives four-foot drops and operates in temperatures from -10C to 50C. It has an 11.6-inch sunlight-readable display and a hot-swappable battery design.
The industrial lineup also included several panel PCs for healthcare, factory monitoring and smart locker systems, as well as hardware based on Intel Wildcat Lake, NXP and Nvidia Jetson Thor platforms, and industrial 4U rackmount systems.
Charging systems
MSI also presented smart EV charging products. These included the Eco Series Home EV Charger, which delivers up to 22kW three-phase output and integrates with solar storage, and the Hyper 80 Dual Fast Charger, an 80kW DC fast charger housed in a 30cm chassis designed for tighter urban sites.
The breadth of the launch underlines MSI's effort to extend beyond its traditional association with PCs and components into AI infrastructure, industrial systems and adjacent business equipment. The latest lineup stretches across a full chain of computing environments, from liquid-cooled racks in data centres to specialised devices deployed in vehicles, factories, campuses and fields.
One of the clearest signals from the portfolio is MSI's attempt to tie Nvidia's latest architectures to practical business uses in sectors such as manufacturing, transport, agriculture, healthcare and legal services, while also building out the rugged and industrial hardware needed to support those deployments.
MSI said its edge and infrastructure products are intended to bring data centre-level AI performance into operational environments where real-time analysis and local processing are required.