SHARON AI, an Australian Neocloud provider, has partnered with VAST Data, an AI Operating System company, to deliver scalable AI inference capabilities to enterprise and government customers.
The collaboration integrates the VAST InsightEngine, an end-to-end system designed for ingestion, embedding, indexing, and retrieval of structured, unstructured, and streaming data in real time. This system supports inference by providing low-latency, massively parallel vector and hybrid search for Retrieval Augmented Generation (RAG) and agentic workflows at scale. When integrated within the VAST AI OS, it maintains unified governance, security, and lineage, ensuring policy-based access controls, encryption, and auditability for all queries.
Ofir Zan, AI Solutions & Enterprise Lead at VAST Data, stated, “As AI systems grow more capable, the ability to reason securely over large datasets in real time will define the next generation of enterprise intelligence. Together, SHARON AI and the VAST InsightEngine orchestrate event triggers and functions connected to data pipelines that scale complex multistep retrieval and reasoning workflows — all within a sovereign environment.”
This partnership aims to enable organizations to transition from AI experimentation to production with repeatable, enterprise-grade workflows. For financial services, the combined technology powers RAG at any scale using a large native vector index to search billions of embedded records while enforcing fine-grained permissions, crucial for high throughput and low latency inference. In public safety and smart cities, the system facilitates ingesting and analyzing massive volumes of video and metadata in real time, which can reduce operational costs, enhance situational awareness, and improve incident response, all while keeping sensitive data within national borders.
Wolf Schubert, CEO of SHARON AI, commented, “By combining SHARON AI’s sovereign GPU cloud with the VAST InsightEngine, we’re creating the foundation for enterprises and government institutions to run cutting-edge AI workloads locally, securely, and without compromise. With our supercluster now live in NEXTDC’s Tier IV M3 data center in Melbourne, this milestone demonstrates our commitment to delivering sovereign, high-performance AI infrastructure for Australia.”
The first workloads on this cluster are underway, with researchers from the University of New South Wales (UNSW) collaborating with SHARON AI cloud to advance reasoning-focused AI research across multiple domains. PhD students are utilizing these resources to improve reasoning in small language models through structured reasoning, auto formalization, and novel expert-aware post-tuning of Mixture-of-Experts architectures. They are also fine-tuning and evaluating state-of-the-art Large Language Models (LLMs) such as Falcon, Llama, Qwen, and Deepseek in parallel for tasks like Question Answering, with applications to mathematics and spatio-temporal reasoning. Additionally, the research aims to accelerate global weather forecasting by training high-resolution data-driven models on large-scale ERA5 datasets for faster and more accurate predictions.
Collectively, this research explores how specialized post-tuning, fine-tuning, and GPU-accelerated model architectures can enhance AI reasoning performance, scalability, and domain-specific applications. This effort by UNSW researchers is establishing the groundwork for smaller, more efficient, and more capable reasoning models applicable across science, forecasting, and advanced AI evaluation.
