#USA #Palo Alto #Sovereign AI #Wand AI #STORONE

WAND AI Partners with STORONE

Wand AI is taking significant strides in the realm of AI infrastructure by integrating StorONE into its Sovereign AI portfolio. This partnership introduces StorONE's Real-Time Tiering technology, designed to optimize storage capabilities. This innovation allows AI programs to maximize their storage utilization efficiently, addressing a crucial need in national AI deployments. The collaboration promises a transformative approach to handling data, significantly lowering costs while maintaining performance.

The Sovereign AI framework heavily relies on two main resources: graphical processing units (GPUs) and storage. Both elements must operate at high efficiency to ensure that national AI initiatives are cost-effective and scalable. Wand AI has already achieved considerable improvements in compute resource utilization through shared capacity systems—boosting usage rates from approximately 25% to over 80%. However, the traditional approach to storage has been less efficient, often requiring expensive solutions to manage cold data that seldom gets used.

Leveraging StorONE’s Real-Time Tiering allows organizations to rectify inefficiencies in their storage ecosystems. Typically, AI storage is provisioned to account for peak demands, resulting in most of the data on flash memory being inactive at any given time. Consequently, organizations have to procure more storage than necessary, tying up resources in systems that do not yield immediate returns.

StorONE offers a novel approach by writing all data to flash instantly at optimal performance levels. It continuously identifies inactive data and relocates only that data to lower-cost media, all within the same storage volume. This process eliminates the need for complex policies, external tools, or necessary restore steps, ensuring immediate access to data when required. The outcome is impressive—a reported 90% reduction in flash storage costs, far surpassing the conventional 30-40% savings provided by traditional data reduction techniques.

One of the most compelling aspects of StorONE's approach is its ability to pool capacity across various workloads rather than confining it to specific ones. This mechanism allows flexibility and accessibility, simplifying how ministries and national institutions can manage vast amounts of information while reducing hardware costs. According to Gal Naor, CEO of StorONE, this technology can have profound implications at the national scale—decreasing flash storage usage translates into fewer physical racks, reduced energy consumption, and minimized data center footprints.

In light of these advancements, Cristian Felix, Wand AI’s Chief AI Architect, emphasizes the partnership's unique value. He notes that sovereign AI is fundamentally about improving economic efficiency as much as enhancing technological capabilities. By welcoming StorONE into their ecosystem, Wand AI adds a pivotal functionality that enables governments and agencies to deploy AI workloads on far less expensive flash media, ultimately conserving budget and resources to further enhance their AI operations.

How It Works

• ---

The integration of StorONE promotes increased utilization in two main areas: 1) Flash utilization: All data can be immediately written to flash, while active data remains accessible, which optimizes overall storage efficiency. 2) Capacity utilization: Consolidating all workloads across shared storage pools cuts down on individual array costs and improves overall resource allocation among departments.

Furthermore, these developments have ramifications beyond mere storage metrics. Investors and stakeholders can expect that the implementation of this innovative system will not only reduce flash requirements for equivalent workloads but will also streamline operations by limiting the physical hardware needed for data storage.

“Deciding on AI program parameters typically focuses heavily on GPU utilization, but many organizations overlook the costs associated with the underlying flash storage they need to support it,” continues Naor. He underscores how the use of his company’s technology has led one enterprise client to cut down from nine cabinets to just two, demonstrating an 80% reduction in physical space and a staggering 78% reduction in power consumption.

Wand and StorONE’s collaboration stands to revolutionize the landscape of AI infrastructure, addressing longstanding issues regarding storage efficiency, speed, and cost. Their joint reference architecture is now accessible for interested governments and enterprises, marking a significant step towards smarter, more responsive AI deployments.

About Wand AI

Wand AI is pioneering the development of dedicated infrastructure for AI labor. This advanced operating system enables a seamless integration of AI capabilities alongside human workers, ensuring that identities and operational protocols are managed effectively. Supporting pivotal sectors, Wand AI extends its services to banks, asset management firms, and other key organizations across the globe.

About StorONE

As a leader in enterprise storage solutions, StorONE maximizes the value derived from flash storage through innovative, hardware-agnostic software platforms. Their Real-Time Tiering ensures data is managed proactively, optimizing storage expenses while maintaining operational efficiency, and making it a preferred choice for diverse sectors looking to modernize their data management capabilities.