(SHAZ) SharonAI Holdings, Inc. Company Overview

SharonAI Holdings, Inc. is a Nasdaq-listed AI infrastructure company supplying high-performance computing through an Australian AI Cloud and Asia-Pacific “AI Factory” network. It combines NVIDIA accelerators, CPUs, storage, networking, orchestration software and data-center power into clusters rented for training, inference and research. The company reports one High Performance Compute Services segment in its 2025 Form 10-K.

Who buys this computing capacity?

Customers include AI laboratories, developers, enterprises, government agencies and research institutions. Sharon AI has named Canva and GMI Cloud and disclosed take-or-pay arrangements for B300-class systems. Buyers obtain configured clusters without separately sourcing scarce GPUs, power, colocation, storage and networking.

Customers pay for compute, storage, networking and related services through on-demand usage or term commitments. Take-or-pay contracts reserve capacity and establish minimum payments, supporting financing visibility. Total contract value is not recognized revenue: revenue begins as contracted service is delivered under the accounting terms.

What did the revenue mix look like before the current expansion?

By FY2025 Sharon AI was already primarily a GPU-service company, although its revenue base remained small. Expansion rests on 2026 contracts valued at about $1.25 billion, $950 million and the $1.32 billion New Zealand agreement announced July 16, 2026. Their value reaches the income statement only after systems are installed, accepted and serving customers.

The quarter ended March 31, 2026 shows a business between historical operations and a much larger buildout. Revenue remained modest while an IPO, a data-center sale, customer commitments and equipment prepayments reshaped the balance sheet. The Q1 2026 Form 10-Q is therefore a deployment baseline, not a steady-state quarter.

Why did accounting gains and losses obscure operating performance?

Q1 included a $65.920 million gain on the sale of the 50% TCDC interest, a $70.228 million adverse fair-value change on convertible notes and $13.519 million of tax expense. These items matter, but they do not measure utilization or unit profitability. Operating analysis should focus separately on revenue, cost of revenue, expenses and cash flow.

The official Q1 results announcement emphasizes contracts and deployments. Future quarters must validate that narrative through commissioning, revenue recognition, margin improvement and cash collection.

Acquisitions, a public-market transaction and a rapid shift from digital assets to AI computing created Sharon AI’s present structure. The sequence explains both its speed and complexity.

1. April 2024
   The group acquired Australian entities and digital-infrastructure assets that supplied an initial operating base, relationships and technical capability.
2. June 2024
   The acquisition of Distributed Storage Solutions added infrastructure activities and helped form the platform later consolidated under Sharon AI.
3. December 2024
   Certification as an NVIDIA Cloud Partner strengthened access to the NVIDIA ecosystem and became an important credibility signal for enterprise AI customers.
4. December 2025
   A business combination closed, the company adopted the SharonAI Holdings name, and securities began trading under SHAZ. The transaction also brought public-company reporting and governance obligations.
5. February 2026
   A Nasdaq initial public offering sold 4,166,666 Class A shares and raised approximately $125 million gross, providing capital for the AI-cloud buildout.
6. First half 2026
   Sharon AI sold its 50% TCDC interest for consideration described at about $74 million, redirected attention to Australia and Asia-Pacific, and announced major B300 and GB300 customer commitments.
7. June–July 2026
   A six-year NVIDIA compute collaboration, $1.6 billion strategic financing, and the New Zealand contract expanded planned capacity to 132MW and pushed expected deployment above 62,000 GPUs by mid-2027.

What did the public-market transition change?

The December 2025 combination and February 2026 offering turned a small private platform into a listed company able to raise large equity and convertible debt. The business-combination announcement marks the legal change; economically, access to capital for GPUs and colocation was more important.

Why was the TCDC divestment strategically significant?

Selling the Texas joint venture generated cash, a note receivable and equity consideration while concentrating strategy on Australia, New Zealand and regional sovereign AI. The trade-off is greater dependence on NEXTDC facilities and timely completion of specific projects.

Contracts do not create capacity by themselves. Each deployment needs powered space, cooling, networking, storage, accelerators, integration, acceptance and working capital. Sharon AI is therefore executing an industrial commissioning program. Its six-year NVIDIA collaboration contemplates 72MW and up to 40,000 GB300 GPUs under revenue-sharing and credit-support arrangements.

How fast has announced capacity expanded?

How does a contract become revenue?

Storage is another constraint. The expanded VAST Data partnership covers 600 petabytes. At management’s benchmark of 6 petabytes per 1,000 GPUs, that supports roughly 100,000 GPUs. Utilization and pricing will determine whether storage adds margin or mainly enables compute revenue.

Sharon AI’s strongest argument is regional specialization: local data residency, power access, certified infrastructure and dedicated clusters for Asia-Pacific customers. NVIDIA Cloud Partner status and enterprise hardware relationships can reduce perceived execution risk, while long-term commitments may support equipment financing.

Which competitors pressure the business?

Is the moat durable?

The moat remains unproven. Contracted power and integrated workloads can create barriers and switching costs, but hardware is purchasable, buyers can multi-source and rivals can cut prices. Sharon AI must show that regional trust and execution produce returns above its cost of capital.

At March 31, 2026, Sharon AI held $164.288 million of cash against $221.469 million of current liabilities, including $199.358 million of convertible notes at fair value. Current assets also contained a $51.014 million note receivable, $42.453 million of prepayments and $8.491 million of listed equity from the TCDC transaction. Liquidity was substantial, but not all assets were unrestricted cash.

What changed after the quarter?

The financing announcement says proceeds fund the NVIDIA collaboration and expansion. It also changes per-share analysis: 6,719,896 Class A shares and pre-funded warrants for 6,374,823 more accompanied $700 million of convertible debt. Headline cash must be matched against new shares, interest, capex and future free cash flow.

What would financial strength look like operationally?

Financial strength will require positive gross profit on mature clusters, operating cash flow that funds maintenance, and financing that does not demand repeated dilution. Free cash flow should subtract equipment and deployment investment from operating cash flow. During construction, it can remain negative even as reported earnings improve.

Sharon AI has two share classes. At July 2, 2026, 35,268,686 Class A shares and 136,341 Class B shares were outstanding. Class A has one vote per share; Class B has 160. The super-voting class gives the three executive founders far greater influence than economic ownership alone suggests, according to the 2026 proxy statement.

How different are economics and votes?

Concentrated control supports long-horizon investment but limits outside shareholders’ ability to redirect strategy if deployment economics disappoint. The board, chaired by Andrew Penn, must therefore provide credible oversight of financing, incentives and capital allocation.

The opportunity is enormous relative to historical revenue. A 132MW footprint and more than 62,000 planned GPUs could create a scaled regional provider. The same asymmetry magnifies risk: delays, financing shortfalls or weak contract margins can overwhelm the small existing operating base.

Which risk is most immediate?

Deployment sequencing links most risks. Delayed power, hardware or installation postpones revenue while financing costs and overhead continue. Low utilization or pricing weakens gross margin; delayed customer acceptance postpones cash. Filings also identify cybersecurity, connectivity, export controls, tariffs, regulation and rapid technology change.

Historical earnings provide little valuation guidance because FY2025 revenue was $1.567 million, Q1 2026 gross margin was negative and major contracts had not entered reported results. A DCF must bridge contracted capacity to commissioned capacity, revenue, gross profit, reinvestment and financing rather than capitalize headline contract values.

Which variables drive a defensible DCF?

What should students and investors monitor next?

Next periods should be judged by recognized contract revenue, commissioned megawatts, billable GPUs, gross margin, operating cash flow, equipment commitments, customer concentration and fully diluted shares. The investor-relations page centralizes new filings and operating updates.