SharonAI Holdings Inc., a prominent Australian neocloud pioneer specializing in high-performance computing, and VAST Data, the trailblazing AI operating system company, have announced a major expansion of their strategic alliance. Under the terms of the broadened agreement, Sharon AI will deploy an unprecedented 600 petabytes (PB) of the VAST AI Operating System to serve as the structural data bedrock for its expanding AI cloud infrastructure.
By utilizing a industry-standard benchmark of roughly 6PB of optimized AI storage per 1,000 GPUs for intense, large-scale workloads, this massive 600PB VAST integration establishes a highly scalable data backbone capable of mirroring the massive processing demands of approximately 100,000 GPUs.
Developing the Sovereign AI Bedrock for APAC
This substantial commitment cements the VAST AI Operating System as the core data infrastructure layer for Sharon AI’s sovereign cloud environments. The upgraded system will cater directly to enterprise organizations, government entities, research institutions, and AI-native developers throughout Australia and the wider Asia-Pacific region.
Representing one of the most geographically extensive and technically advanced sovereign AI data foundations in the APAC zone, the collaboration places Sharon AI at the vanguard of Australia’s ongoing initiative to establish self-reliant, elite AI capabilities. The architecture guarantees that highly sensitive workloads, critical intellectual property, and strategic institutional data remain onshore while simultaneously delivering the raw processing capacity essential for modern training, inference, and complex agentic AI frameworks.
Also Read: Sharon AI and NVIDIA Ink Six-Year Strategic Collaboration to Deploy 72MW AI Factory in Australia
Advanced Architectural Capabilities for Regulated Sectors
In heavily regulated and sovereign ecosystems, absolute control over data residency, transit, and user access remains paramount—and compromises on execution speeds are not an option. The VAST AI Operating System was custom-engineered to overcome these exact bottlenecks. Featuring a revolutionary Disaggregated Shared Everything (DASE) architecture, the platform operates as a true parallel distributed system that completely bypasses the data traffic congestion typical of legacy systems.
Furthermore, a singular global namespace allows every processing unit to access the entire data environment instantaneously, rendering time-consuming data replication or multi-petabyte migrations obsolete. The framework’s native multi-tenancy capabilities also provide robust workload isolation, ironclad performance SLAs, and individualized administrative controls per user, offering an optimal environment for cloud providers managing multiple sovereign clients across shared hardware infrastructure.
As Sharon AI’s commercial workloads and ambitious AI factory initiatives continue their steep upward trajectory, both organizations will maintain a deeply integrated engineering collaboration to continuously optimize the platform.
Executive Commentary
“Our customers refuse to choose between keeping their data sovereign and running AI at full speed – they need both at the highest level,” said James Manning, Co-Founder and CEO of Sharon AI. “Standardizing on the VAST AI Operating System at this unprecedented 600PB scale gives us exactly that: a rock-solid, high-performance foundation we can scale confidently as demand for sovereign AI across Australia and Asia-Pacific accelerates. This is how we build Australia’s AI future – powerful, secure, and proudly local.”
“Every breakthrough Sharon AI enables – in research, medicine, industry, or national capability – runs on data, and it only moves as fast and safely as the foundation beneath it,” said Renen Hallak, Founder and CEO of VAST Data. “As the data foundation for Sharon AI’s sovereign cloud, we are proud to partner with them to support Australia’s hardest and most strategic AI ambitions. The VAST AI OS turns massive data estates into real-time intelligence assets, all while respecting sovereign boundaries. This is the future of AI infrastructure.”
