As generative AI and large-scale language models rapidly drive enterprise digital transformation, the demand for high-performance computing and storage architectures for AI training and inference continues to rise. INFINITIX, a leader in AI infrastructure management software , announced a technology collaboration with Phison, a global leader in storage technology. This collaboration combines Phison’s aiDAPTIV+ smart storage technology with INFINITIX’s AI-Stack infrastructure management software to create a next-generation, integrated hardware and software enterprise-grade AI training and inference solution.

Breaking Boundaries and Unleashing the Potential of AI Computing Power:

This collaboration leverages high-speed SSDs and smart memory expansion technologies to overcome the hardware limitations of traditional HBM and GDDR. Through deep integration of aiDAPTIV+ and AI-Stack, enterprises can seamlessly incorporate hardware acceleration capabilities into AI workload scheduling within a native Kubernetes environment, achieving end-to-end performance optimization from model training to inference deployment.

AI Authority Trend: Supermicro Partners with NVIDIA to Accelerate Liquid-Cooled AI Infrastructure

Chen Wen-yu, CEO of Digital Unlimited, stated: “AI has entered a stage of large-scale application centered on architecture and platform capabilities. The key is no longer just computing power, but how to effectively manage, expand, and transform the value of computing power. This collaboration with Phison, through the deep integration of AI-Stack and aiDAPTIV+, incorporates storage layer capabilities into the AI ​​infrastructure scheduling system. This helps enterprises integrate heterogeneous computing power, memory, and storage resources in a Kubernetes-native manner, accelerating the training and inference deployment of large models in AI data center environments with a more flexible and cost-effective architecture, and creating a scalable enterprise-grade AI platform.”

Phison Electronics CEO Jiancheng Pan stated , “AI is rapidly evolving from single-GPU computing to system-level architectural innovation across nodes and resources. Phison, with aiDAPTIV+, formally integrates the NAND storage layer into AI memory and computing architecture, redefining the way AI systems are scaled. This collaboration with Digital Unlimited leverages the native scheduling capabilities of Digital Unlimited’s AI-Stack management software to enable NAND storage, memory, and computing power to work collaboratively in enterprise environments. We believe that this hardware-software integration model will become the next standard for enterprise-level edge AI infrastructure.”

AI Authority Trend: Hut 8 Announces AI Infrastructure Partnership with Anthropic and Fluidstack

AI-Stack achieves the optimal balance between enterprise-level computing power scheduling, operational performance, and cost.

Built on a native Kubernetes architecture, it integrates GPU partitioning, aggregation, and cross-node computing capabilities, fully supporting NVIDIA and AMD GPUs. It can integrate traditional GPU servers and Phison aiDAPTIV+ nodes into a single platform for unified management. Combined with multi-tenant access control, automated scheduling, centralized monitoring, and billing mechanisms, it helps enterprises significantly reduce the complexity of AI infrastructure governance and operation. Through its integrated hardware and software architecture, enterprises can complete the training and inference deployment of large-scale language models without purchasing high-end HBM GPUs. With aiDAPTIV+ memory expansion and AI-Stack’s intelligent scheduling capabilities, enterprises can significantly improve ROI with more flexible hardware configurations, accelerating the large-scale deployment of AI projects.

The collaboration between Digital Unlimited and Phison Electronics , aiming to define new standards for AI storage and computing power, is not merely a technology integration but also a joint investment in the development of infrastructure for the AI ​​era. Moving forward, both parties will continue to deepen their cooperation in AI, smart storage, and cloud operation architecture, promoting efficient and scalable data infrastructure to help enterprises smoothly navigate the rapid evolution of the AI ​​era.

AI Authority Trend: Eclypsium Joins NVIDIA Inception to Strengthen Security Across AI Infrastructure

Source – PR Newswire

To share your insights, please write to us at info@intentamplify.com