Ridger Achieves xFusion's FusionXpark™ (NVIDIA GB10-Based) Compatibility Certification, Delivering Industry-Leading Mobility for Practical AI Deployment

The fusion of AI-native storage and computing overcomes the final barrier to enterprise AI adoption, delivering unprecedented mobility, performance, and simplicity for data-sensitive industries.

SINGAPORE, Feb. 10, 2026 /PRNewswire/ -- Ridger, the world's pioneering AI-Native storage provider, today announced the successful completion of full compatibility certification between its flagship MIMO storage system and xFusion's FusionXpark™ Portable Agent Development Platform, powered by the NVIDIA GB10 (Grace Blackwell) superchip. This milestone delivers the industry's first truly integrated, portable AI infrastructure solution, effectively turning the promise of private, desktop-scale supercomputing into a practical reality for businesses worldwide.

From Data Center Chains to Portable Power: Solving the "Last-Mile" AI Dilemma
Global industries bound by strict data governance—including healthcare, legal, financial services, and creative sectors—face a critical challenge: harnessing the power of AI while ensuring data never leaves their secure premises. Traditional approaches force a compromise, tethering powerful AI to immobile, complex, and costly data center infrastructure. This has created a widespread "last-mile" gap, where AI is technically deployable but practically unusable due to logistical and operational burdens.

The "Fast, Light, Edge" Alliance: A Complete Stack in a Mobile Form Factor
The Ridger-xFusion partnership bridges this gap by uniting two purpose-built, best-of-breed platforms. Ridger MIMO, the cornerstone of AI infrastructure, delivers data center-class performance—400 GB/s bandwidth, 500M IOPS, and near 2PB capacity in a compact chassis—with an architecture fundamentally redesigned for AI workloads. Its native support for GPU storage protocols allows seamless, direct attachment of up to 16 FusionXpark™ units without additional switches.

xFusion FusionXpark™ leverages the revolutionary efficiency of the NVIDIA GB10 superchip (TSMC 3nm) to place 1 PFLOPS of FP4 AI computing power on the desktop, coupled with an integrated ecosystem and marketplace for instant deployment of AI agents and applications.

Together, they converge into a synergistic, mobile powerhouse that consolidates the entire AI workflow—from data preparation and model training to fine-tuning and real-time inference—into a single, elegantly simple, and highly portable deployment unit.

The New Architecture: Redefining Efficiency for Private AI
Third-party validated benchmarks and architectural analysis confirm the solution delivers not just incremental gains, but a fundamental shift in private AI infrastructure.

The Foundation: Storage That Fully Saturates Compute
MIMO's extreme, AI-native bandwidth is architected to fully saturate the dual QSFP interfaces of every FusionXpark™ unit. This ensures zero computational cycles are wasted waiting for data, transforming FusionXpark™'s 1 PFLOPS of FP4 power from a theoretical peak into guaranteed, sustained workload performance. It eliminates the I/O bottleneck at its root, ensuring every investment in compute silicon delivers its maximum return.

The System Impact: Unleashing Cluster-Wide Efficiency
This foundational data velocity creates a cascade of system-level advantages. With MIMO as a central, ultra-high-bandwidth data nexus, multiple FusionXpark™ nodes operate as a single, cohesive AI supercomputer. The shared storage pool eliminates data silos and replication overhead, enabling dynamic workload orchestration across the entire cluster for both training and inference. The result is a transformative increase in total usable compute output per dollar—maximizing the return on the entire AI investment.

The User Experience: Data-Center Performance, Desktop Deployment
This immense power is delivered in a form factor engineered for the edge. MIMO is not only compact and devoid of external switch dependencies but also incorporates a near-silent acoustic profile. It brings unprecedented data-center-grade performance directly to where data resides and decisions are made—in labs, offices, or field sites—without the noise, complexity, or infrastructure overhead of traditional equipment, eliminating the traditional trade-offs and fully delivering on the "Light and Edge" promise.

Executive Commentary

Bruce Xue, Spokesperson of Ridger: "Today marks the end of the 'unusable AI' era for regulated industries. By fusing true AI-native storage with accessible supercomputing, we've eliminated the last excuse for not deploying AI privately. The MIMO-FusionXpark™ solution isn't just another product; it's the first 'AI Appliance' for the edge—delivering cloud-scale performance with uncompromising data sovereignty. We're not just solving a storage problem; we're unlocking a trillion dollars of trapped data value."

Vandia Yang, Spokesperson of xFusion: "The future of AI isn't in the cloud; it's wherever your data lives. Our partnership with Ridger crystallizes this vision. Together, we've built what was previously impossible: a sovereign AI supercomputer that fits under a desk. This certification proves that performance and privacy are no longer a trade-off. We are giving every research lab, every hospital, every financial firm the keys to their own intelligence revolution."

Path to Deployment and Strategic Vision

The certified MIMO-FusionXpark™ integrated solution is available for order now through Ridger's official website and authorized partners, with shipments commencing at the end of February 2026. 

This partnership, anchored by the NVIDIA GB10 platform, delivers a definitive blueprint for the future of private AI: a practical, closed-loop ecosystem of compute, storage, and intelligence that is as portable as it is powerful. Ridger remains committed to advancing this "Fast, Light, Edge" paradigm through continued R&D and ecosystem collaboration, driving the next wave of secure and sovereign AI adoption that is liberated from fixed infrastructure.

About Ridger
Ridger is defining the future of data infrastructure as the world's first AI-Native storage provider. Its flagship product, MIMO, is built on a ground-up architecture designed to unlock the full potential of AI and accelerated computing, prioritizing unparalleled speed, density, and edge readiness. Learn more at https://ridger.tech.

About xFusion
xFusion is a leading provider of computing infrastructure solutions. Its FusionXpark™ is a portable Agent Development Platform built on the NVIDIA GB10 superchip, bringing datacenter-scale AI performance to the edge and private environments. With its integrated AI agent marketplace and development tools, FusionXpark™ simplifies and accelerates the deployment of enterprise AI applications. Learn more at https://www.xfusion.com/en/product/fusionxpark?utm_source=homepage. 

 

** The press release content is from PR Newswire. Bastille Post is not involved in its creation. **

Ridger Achieves xFusion's FusionXpark™ (NVIDIA GB10-Based) Compatibility Certification, Delivering Industry-Leading Mobility for Practical AI Deployment

SHANGHAI, April 3, 2026 /PRNewswire/ -- As AI becomes essential in K–12 education, many teachers face a practical question: How to teach AI through hands-on, project-based learning without advanced coding skills?

A recent workshop in Kathmandu University offers a clear answer. Conducted by DFRobot, an innovator in STEAM education, the three-hour workshop at the School of Engineering brought together teachers and department heads from across disciplines. Participants completed two hands-on AI projects while exploring how to translate these experiences into their own classroom teaching.

From Awareness to Classroom Practice

According to the Computer Science Teachers Association, while most educators support AI in the curriculum, many lack confidence in teaching it effectively. This workshop directly addressed that gap—moving from awareness to practical implementation.

Learning AI Through Hands-On Building

The training adopted a face-to-face, project-based approach, combining technical learning with pedagogy. It focused on deploying AI capabilities—such as speech and vision—on edge devices, connecting abstract concepts to interactive classroom applications.

Using UNIHIKER K10 hardware and Mind+ graphical programming software, participants completed two progressive projects:

Voice Interaction

Using the UNIHIKER K10 and the Mind+ graphical programming platform software, the workshop introduced voice-based human–machine interaction as an accessible entry point into AI learning. Teachers began by creating a simple voice-controlled system with commands such as "turn on the light," experiencing a fundamental AI interaction loop: wake → recognize → execute.

Building on this foundation, they extended the system to control on-screen movement through voice directions, transforming a basic function into a more engaging and interactive experience. Rather than focusing on complex programming, the training emphasized how simple voice commands can be translated into practical classroom applications.

Through this process, teachers gained a clear understanding of real-time voice interaction and its classroom potential.

Vision Interaction

The workshop introduced the HUSKYLENS AI vision sensor and its face recognition capabilities, helping teachers understand how machines perceive the world through visual data. Trainers explained the core workflow of face recognition—including detection, alignment, encoding, and matching—providing a clear technical framework for classroom application. Building on this foundation, teachers connected HUSKYLENS with the Mind+ programming environment and implemented real-time recognition tasks. Extending this approach, they developed a "Smart Pet" system by integrating HUSKYLENS with the UNIHIKER K10. The system could recognize different types of cats—such as Orange Tabby, Striped Tabby, and Siamese—and respond with corresponding interactive states.

Through this hands-on process, teachers gained a clear understanding of computer vision concepts and how visual input can drive interactive systems, enabling more engaging and project-based learning in the classroom.

A Model That Can Be Applied in AI Classrooms

The workshop highlights a practical approach to AI education that can be replicated across schools:
Accessible: Entry-level coding is all that's needed — suitable for K–12 learners.
Adaptable: Projects can be transformed into games, smart systems, or classroom tools
Structured: Aligned with project-based learning and real-world problem-solving

As part of the assessment, each teacher was asked to outline how the projects could be adapted for their own classrooms—highlighting a core objective of the training: not just using tools, but enabling curriculum design and effective knowledge transfer.

More importantly, it emphasized a critical shift—from understanding AI to applying it. By experiencing the full creation process, teachers gained the confidence to bring AI into their own classrooms.

** This press release is distributed by PR Newswire through automated distribution system, for which the client assumes full responsibility. **

How to Teach AI in the Classroom: A Hands-On Teacher Training Workshop at Kathmandu University