Sharonai Holdings — Business Overview

What does SharonAI do?

SharonAI is an Australian "neocloud" — a cloud computing provider built specifically to rent out high-powered GPU computing resources for artificial intelligence and high-performance computing (HPC) workloads. Rather than offering general-purpose cloud services like Amazon or Microsoft, SharonAI focuses entirely on the dense, specialized computing power that AI model training, inference (running a trained model), and scientific simulation require. Its customers — enterprises, government agencies, research institutions, and AI-native companies — pay to use SharonAI's GPU clusters without having to buy or manage the hardware themselves.

SharonAI offers three core services:

As of December 31, 2025, the company had 411 GPUs deployed and generating revenue, with a further 2,024 NVIDIA B200 and B300 GPUs purchased and expected to be deployed in the first half of 2026 — nearly a six-fold increase in total GPU count.

How does SharonAI make money?

SharonAI charges customers for access to its GPU infrastructure, primarily through two pricing models: usage-based on-demand consumption and longer-term "take-or-pay" contracts. Historically, revenue has come largely from on-demand workloads accessed through marketplace aggregators (platforms that pool GPU capacity from multiple providers). As the company scales, it is shifting toward multi-year term contracts — typically 36 to 60 months — where customers commit to paying for reserved GPU capacity whether they use it fully or not. This provides more predictable revenue and helps SharonAI secure financing to purchase new GPU hardware.

The contract lifecycle is designed to maximize asset returns over time. A customer pays a 10–20% deposit at contract signing, which SharonAI uses to fund hardware procurement. Revenue recognition begins around month five, once infrastructure is live. After the initial contract ends, SharonAI can re-contract the same hardware at lower rates, sell capacity on spot markets at higher hourly rates, or redeploy older GPUs for token-based AI inference through its proprietary software platform — extending the useful economic life of hardware to approximately six years.

What market does SharonAI operate in?

SharonAI operates at the intersection of cloud computing, GPU-as-a-Service, and sovereign AI infrastructure — a market experiencing rapid structural growth driven by the explosion in demand for AI workloads. The broader trend is straightforward: AI model training and inference require massive amounts of parallel computing power that traditional CPU-based cloud platforms cannot efficiently provide. This has created a specialized market for GPU cloud platforms, or "neoclouds." The neocloud segment is expected to grow at a compound annual growth rate (CAGR) of 23% from 2024 to 2030, according to the filing. NVIDIA itself reported that its GPU installed base was fully utilized as of late 2025, with compute demand growing 56% year-over-year and Blackwell-series chips sold out through 2026.

A particularly important secular trend for SharonAI is the push for "sovereign AI" — the requirement that sensitive government and enterprise data be processed within national borders. Australia, like many governments, is increasingly mandating that regulated industries and public agencies keep data on-shore. This creates a structural demand for locally-hosted AI infrastructure that global hyperscalers (Amazon, Microsoft, Google) hosting data internationally cannot fully satisfy. SharonAI positions itself as the domestic alternative, operating exclusively within Australian data centers certified for government workloads.

Who are SharonAI's main competitors?

The GPU cloud market includes everyone from the world's largest technology companies to small specialized neoclouds, making it both intensely competitive and somewhat fragmented at the specialized end. SharonAI's named competitors range from hyperscalers (Amazon, Microsoft, Alphabet/Google) to purpose-built GPU cloud platforms (CoreWeave, DigitalOcean, Nebius Group) to companies that have pivoted from cryptocurrency mining into AI compute (Hive Digital, IREN, Hut 8, Bit Digital, Core Scientific, Applied Digital, BitDeer). SharonAI acknowledges that many competitors are larger, have greater access to energy infrastructure, and can more easily raise capital.

SharonAI's claimed competitive advantages center on three things: NVIDIA Cloud Partner (NCP) status, sovereign Australian hosting, and its proprietary software platform. Being one of only three NCPs in Australia provides preferential access to NVIDIA's latest GPU generations, customer referrals via NVIDIA's consumption desk, and a degree of supply chain visibility unavailable to non-certified operators. The sovereign angle — hosting within NEXTDC's government-certified Tier III and IV Australian data centers — is a genuine differentiator against global hyperscalers for regulated customers. The SharonAI Studio platform and proprietary inference engine add software differentiation by extending the economic life of older GPUs and reducing customer setup complexity.

Where does SharonAI operate?

SharonAI is primarily an Australian business, with its core GPU infrastructure hosted in data centers in Melbourne and Sydney. Its flagship deployment is the 1,016-GPU Supercluster co-located at NEXTDC's M3 data center in Melbourne. SharonAI has secured up to 54MW of total data center capacity across four NEXTDC facilities: M3 (Melbourne, ~2MW, currently operational), M2 (Melbourne, 40MW, planned), S6 (Sydney, 10.75MW, planned), and S3 (Sydney, 2MW, planned). Additional data center relationships exist with DigiCo and GreenSquareDC across Sydney, Brisbane, Adelaide, and other Australian cities.

The company has a small U.S. corporate and operational presence but has divested its only material U.S. infrastructure asset. SharonAI's principal executive offices are at 745 Fifth Avenue in New York. It had formed a 50/50 joint venture (Texas Critical Data Center LLC) to develop a data center in the Permian Basin of Texas, but sold its 50% stake in January 2026 for US$70 million, explicitly transitioning to a pure-play Australian neocloud model for the near term. U.S. subsidiaries (SharonAI Operations LLC and SharonAI Hosting LLC) remain in place for future expansion.

The company's ambitions extend to the broader Asia-Pacific region, supported by NEXTDC's connectivity to Singapore (81ms latency), Hong Kong (130ms), US West Coast (165ms), and Europe (265ms). The Megaport partnership also gives SharonAI's customers private network access to over 1,000 data centers across 26 countries. However, as of the filing date, all revenue-generating GPU infrastructure is located in Australia.