SHAZ (SharonAI Holdings) Deep Dive
Executive Summary
•Pre revenue AI neocloud: TTM revenue of just $3.1m against a consensus forward ramp to $100m NTM and $1.36b FY+2, a 1207% forward growth rate that demands flawless GPU deployment execution
•Stock has broken down hard: -53% over 1 month and -62% over 3 months, yet sell side consensus PT sits at $95 versus $26.97 spot, a roughly 250% gap that signals either deep mispricing or stale analyst models
•Valuation frameworks are effectively unusable: single analyst coverage, n=1 historical multiple samples, and 13F ownership reported at 770% of shares outstanding all point to severe data quality problems
•The entire bull case rests on NVIDIA GB300 allocation (up to 40,000 GPUs across a 72MW Australian AI factory) against a balance sheet carrying 89.7% debt/equity and deeply negative operating margins
Financial Snapshot
•Revenue TTM: $3.1m
•Revenue growth: Not disclosed (pre scale)
•Gross margin: 33.9%
•Operating margin: -624.8%
•Net margin: deeply negative (Q4 net loss of -$0.43b, likely includes non cash items)
•ROE: -84.3%
•Debt/Equity: 89.7%
•EPS TTM: Not disclosed
•Market cap: reported $113.3m conflicts with the roughly $950m implied by 9.27x forward P/S on $100m NTM revenue and the $952m figure in recent market data
Forward Outlook
•NTM revenue: $0.10b ($100m), +1207% YoY implied
•NTM EPS: -$11.72 (still loss making)
•FY+2 revenue: $1.36b
•FY+2 EPS: -$2.75 (loss narrowing but not breakeven)
•NTM EPS revisions improving: -$4.49 now versus -$6.17 one month ago (+27.3%)
•Margin trajectory: gross margin 33.9% today, operating leverage must scale dramatically to close a -624.8% operating margin gap, and FY+2 remains loss making
•Coverage thin: 1 analyst, so the forward curve is a single model, not a consensus
Valuation Analysis
•NTM P/S: 9.27x (forward)
•NTM P/E: Not meaningful (negative earnings)
•NTM EV/EBIT: Not meaningful (negative EBIT)
•Trailing EV/Revenue: 0.1x, distorted by the revenue ramp
•LF P/BV: 0.9x versus 2 year average 39.7x (n=5, -1.3 SD), implying the market now values book at a fraction of history
•Historical EV/Rev average of 277.5x is based on n=1 and is statistically meaningless
•Implied prices at historical average multiples ($8982 on EV/Rev, $1253 on P/BV) are artifacts of tiny samples and should be ignored
•Sell side consensus PT: $95 (median $95, range $95 to $95, 1 Buy / 0 Hold / 0 Sell)
•Prior initiations: $50 at Compass Point and Lucid Capital (April 2026), $95 at Northland (September 2026)
•CRWV (CoreWeave): ~8 to 10x NTM P/S, GPU neocloud leader, larger scale and contracted backlog
•NBIS (Nebius): ~10 to 12x NTM P/S, European AI cloud, stronger balance sheet
•IREN: ~7 to 9x NTM P/S, bitcoin miner pivoting to AI compute
•APLD (Applied Digital): ~8 to 10x NTM P/S, data centre buildout with hyperscale leases
•SHAZ at 9.27x NTM P/S screens in line with peers on multiple but with materially less revenue, no proven backlog, and far weaker margins
Key Risks & Catalysts
Risks
•Execution risk: deploying 40,000 GB300 GPUs across 72MW requires capital, power, and supply chain delivery the company has not yet demonstrated at scale
•Financing and dilution risk: deeply negative cash flow against 89.7% debt/equity leaves little room, and equity raises would be highly dilutive at current prices
•Accounting and governance red flags: two "Change in Accountant" 8-K filings within 48 hours (23 and 24 September 2026)
•Customer concentration: a revenue base of $3.1m implies a handful of contracts, any loss is material
•Coverage risk: a single analyst drives the entire forward estimate set, so revisions are volatile and unreliable
•Data integrity risk: 13F ownership reported at 770% of shares outstanding and conflicting market cap figures undermine confidence in reported metrics
•Rate and capex sensitivity: GPU economics depend on financing costs and residual values, both of which compress in a higher rate environment
Catalysts
•NVIDIA GB300 deployment milestones and any confirmed GPU delivery schedule
•Revenue inflection from $3.1m TTM toward the $100m NTM run rate
•New enterprise or hyperscale customer wins
•Additional analyst initiations, which could broaden the estimate base
•Institutional accumulation: 135 holders (+106 QoQ), $1.6b 13F value, with Millennium, Citadel, Two Seas, and Diameter building positions
•Continued improvement in NTM EPS revisions (-$4.49 versus -$6.17 a month ago)
•Potential index inclusion or liquidity events as market cap scales
Investment Thesis
SHAZ is a binary, early stage AI infrastructure bet dressed up in a valuation that no conventional framework can support. The revenue trajectory on paper is extraordinary: $3.1m TTM to $100m NTM to $1.36b FY+2. But that curve is the output of a single analyst model, not a contracted backlog, and it sits against a -624.8% operating margin, 89.7% debt/equity, and two accountant changes in two days. The NVIDIA GB300 collaboration is a genuine differentiator, giving access to scarce cutting edge silicon in an underserved geography, but access to allocation is not the same as funded, profitable deployment.
The stock has already de rated violently, down 62% in three months and 53% in one month, with forward P/S compressing to roughly 9x, in line with better capitalised peers. The sell side PT of $95 looks stale and unsupported given the drawdown and the single analyst base. The bull case requires the company to convert GPU allocation into revenue and to fund a capital intensive buildout without crippling dilution, neither of which is yet proven.
Recommendation: avoid, or for risk tolerant mandates only, a small speculative position. The risk/reward is genuinely binary and the data quality is too poor to underwrite a base case. We would want to see a funded deployment schedule, contracted revenue disclosure, and a stable audit relationship before paying 9x forward sales. Fair value is unanchored by current fundamentals; on a peer cross check the stock is fairly valued at best, and the $95 consensus target should be discounted heavily until execution evidence emerges.
