SharonAI Holdings, Inc. Class A Common Stock
SHAZ has real contracted AI-infrastructure demand and a credible revenue inflection into Q4 2026, but at 277.46x sales with three straight major EPS misses and revenue still effectively pre-ramp, most of the upside remains execution-dependent rather than earned.
SHAZ is one of the stranger setups in the market because the operating story is much better than the income statement. The company has assembled ingredients that many AI infrastructure hopefuls never get: 120 MW contracted under multiyear take-or-pay agreements, 212 MW secured capacity by end-2027, a jump in contracted TCV to $8.8B YTD from $2.2B three months ago, and financing support with $2.2B raised since Dec-25. Those are not cosmetic milestones. They suggest customers are willing to commit ahead of delivery, and management is winning enough supply to matter.
The problem is that almost none of this has flowed into reported financials yet. FY2025 revenue was $0.0B, operating margin was -1204.6%, net margin was -3784.1%, and free cash flow was roughly $-0.3B. On top of that, management's execution record with the Street is poor: three straight misses, capped by 2026-08-06 EPS of -26.16 versus -0.305 expected. That is why the stock, even after falling 47.4% from the 52-week high and trading 15.6% below the 200-day average, does not yet earn a BUY. This is still a trust-me story until revenue actually starts.
The case now depends on three things. First, the company must hit management's statement that, "We expect the first material revenue to commence in the fourth quarter of 2026 as large-scale B300 and GB300 deployments come online." Second, the B300 handover and GB300 schedule must translate into sustained billing, with management targeting full billing on the cluster in Q3 2027. Third, pricing has to stay healthy enough for the future margin profile to justify today's enterprise value, and the record price over $4 per GPU-hour is the best current evidence on that front.
What the market may still be underestimating is that real contracted demand appears to exist, and the stock is no longer priced at the euphoric extreme seen near $97.48. What the market may still be overestimating is how quickly contracted demand turns into recognized, profitable revenue. That balance keeps the call at HOLD. Confirmation of the Q4 revenue start, clean hardware delivery, and signs that EBITDA losses are narrowing as utilization rises would be enough to revisit the rating upward.
Hardware arrives on time, B300 and GB300 billing ramps cleanly into 2027, and record pricing above $4 per GPU-hour proves durable.
Initial Q4 2026 revenue lands, contracted capacity keeps building, and investors give credit for the 2027 billing ramp but not full perfection.
Revenue conversion slips past Q4 2026, another funding step is needed, and the market stops capitalizing TCV until billing actually starts.
Quant call (HOLD) diverges from analyst consensus (Strong Buy: 1 buy / 0 hold / 0 sell) — resolve explicitly.
In the short run, the market is a voting machine, but in the long run, it is a weighing machine.
Forward earnings power is the whole story here because trailing metrics are almost unusable: P/E is N/A, trailing sales are effectively zero, and the stock is being valued on whether 120 MW of contracted capacity, >64,000 GPUs by mid-2027, and full GB300 billing by Q3 2027 convert into a real earnings base. The moat today is mostly access and orchestration, contracted capacity, financing, and customer relationships, but over 3-5 years that moat only hardens if SHAZ proves reliable delivery across power, data-center readiness, and high-end GPU deployment before larger balance-sheet competitors compress pricing. AI is clearly a secular tailwind because customer demand and contract duration are strong, yet this is an advantaged enabler only if management executes the ramp on time; otherwise the same AI build-out wave can bypass SHAZ in favor of better-capitalized providers. At 4.73% on the 10-year and 3.63% Fed funds, long-duration execution stories still need proof, which supports a HOLD rather than paying ahead for a 2027 outcome today.
- $8.8B YTD contracted TCV versus $2.2B three months ago, showing demand formation is real, not theoretical
- First material revenue in Q4 2026, with full billing on the GB300 cluster in Q3 2027, making the next 12 months a timing and delivery story
- P/S of 277.46 versus peer average 2.0, leaving little room for another delay after three consecutive EPS misses, including -26.16 vs -0.305 on 2026-08-06
- Contracted capacity is 120 MW under multiyear take-or-pay agreements, and management says 92 MW remains available to sell, giving a visible path to additional bookings before revenue fully ramps
- Secured AI factory capacity rose by 80 MW to 212 MW for end-2027, supported by new Australia partner capacity, which expands the platform faster than prior expectations
- Record price over $4 per GPU-hour on recent B300 deployment suggests pricing power is holding while management says demand materially outpaces supply
- $2.2B capital raised since Dec-25, including a $1.6B oversubscribed round, reduces near-term financing pressure during the heaviest build-out phase
- FY2025 revenue was $0.0B and management still guides only that first material revenue starts in Q4 2026, so today's valuation is being asked to underwrite a business that is not yet proven at scale
- FY2025 EPS was -$2.18, operating margin was -1204.6%, and net margin was -3784.1%, showing the current economic profile remains deeply negative even before the main revenue ramp
- Execution credibility is weak: EPS missed street estimates in all three reported 2026 quarters, including -8477.0% in August, -63.4% in May, and -225.7% in February
- Leverage is not extreme on equity terms but still meaningful for a pre-revenue build, with $1.0B total debt, 0.90 debt/equity, and $-0.3B FCF
| Metric | Value | Context |
|---|---|---|
| Current price | $51.31 | Reference price for the engine target and scenario framing |
| Price target | $57.15 | 11.4% upside vs current, High confidence, basis sales_quality_blend |
| Quant call | HOLD | High conviction baseline from the quantitative engine |
| FY2025 revenue | $0.0B | 118.7% YoY, but still effectively pre-revenue at reported scale |
| FY2025 EPS | $-2.18 | 1504.5% YoY, still deeply negative |
| FY2025 operating margin | -1204.6% | Prior FY N/A, reflects build-out before revenue ramp |
| Valuation | P/S 277.46, P/B 0.76, P/E N/A | Sales multiple far above peer avg 2.0; peer avg P/E 15.8 |
| Balance sheet | $1.0B debt, 0.90 debt/equity, current ratio 1.63 | Adequate liquidity for now, but leverage matters with negative FCF |
| Free cash flow and capex | $-0.3B FCF, $0.3B capex | Heavy investment phase, no buybacks or dividends |
| Commercial backlog indicators | $8.8B YTD contracted TCV, 120 MW contracted, 212 MW secured capacity | Demand and capacity scaled sharply ahead of revenue recognition |
| Earnings track record | 3 straight misses | Latest was -26.16 vs -0.305, after -1.43 vs -0.875 and -2.85 vs -0.875 |
| Price action | -47.4% vs 52-week high, -15.6% vs 200-day average | Stock has already de-rated materially from peak optimism |
The biggest risk is that hardware delivery and data center readiness slip the ramp beyond Q4 2026, because the stock already trades at 277.46x sales despite $0.0B FY2025 revenue.
The next catalyst is the November 5, 2026 earnings report, specifically evidence that B300 handover is converting into billable revenue and that the Q4 2026 material-revenue start remains on track.
Ownership is still relatively light and event-driven, with the top three institutions holding only 11.1% and insiders at 0.0%, which limits the signal from long-only sponsorship. Short interest is elevated at 10.63% of float, or 2,536,880 shares, though days-to-cover is only 1.7, so the setup can amplify upside around a clean revenue-start update but does not constitute a deep structural squeeze. Options positioning is described as positive gamma in the front end, which can dampen extreme near-term volatility unless earnings materially reset expectations.
Current economics are pre-ramp, with first material revenue expected in Q4 2026 and monetization tied to B300 and GB300 deployments.
Multiyear take-or-pay agreements provide the clearest operating KPI until reported revenue becomes meaningful.
This is the near-term commercial upside lever if management converts current demand outpacing supply into signed contracts.
Management is targeting more than 64,000 GPUs deployed by mid-2027 across this footprint.
SHAZ scaled contracted demand and capacity sharply, with material revenue starting in Q4 2026.
We expect the first material revenue to commence in the fourth quarter of 2026 as large-scale B300 and GB300 deployments come online.
We are targeting more than 64,000 GPUs deployed by mid-2027 across our footprint in Australia and New Zealand.
The central message from the quarter is that we have materially increased each of the 3 inputs required to scale this business, AI factory capacity, contracted customer demand and capital.
we sold that capacity for record dollars per hour or price per megawatt hour, depending on both ways you're thinking about it for both B300 and GB300.
if you say what are the risk factors, it's hardware delivery and data center readiness.
The core edge is not current profitability, it is assembly of scarce inputs: power-backed AI factory capacity, access to high-end GPU deployments, and customer willingness to sign 3- to 5-year take-or-pay contracts out to 2031. That edge is real today because contracted TCV reached $8.8B YTD and management claims demand exceeds supply, but durability is still unproven because larger infrastructure players can attack pricing and financing if SHAZ stumbles on delivery. The moat gets stronger only if the company turns contracted MW into reliable, on-time billing through 2027.
Capital allocation is currently about survival and scale, not shareholder return. SHAZ has raised $2.2B since Dec-25, including a $1.6B oversubscribed round, has $0.0B buybacks and $0.0B dividends, and is spending heavily with capex of $0.3B while FCF sits at roughly $-0.3B. That is sensible for a pre-ramp AI infrastructure build, but discipline will only be proven once revenue starts and management shows it can fund growth without serial dilution or expensive debt layering.
SHAZ is a real AI infrastructure build, not a concept stock, but the reported numbers have not caught up to the narrative. Rated HOLD, High conviction because the commercial signals are strong while valuation and execution risk are still too high for a clean BUY. The stock can work if Q4 2026 truly marks the start of material revenue and if pricing above $4 per GPU-hour holds through the 2027 billing ramp. A move to BUY would require proof of on-time revenue conversion and improving unit economics. A move to SELL would require another material delay, weaker contracting, or evidence that the financing bridge is getting thinner.
≣ Full research note tap to expand
Executive Summary
HOLD, High conviction. SHAZ rates a HOLD at High conviction. The company has assembled unusually large forward demand for its size, including $8.8B YTD contracted TCV, 120 MW contracted capacity, and guidance for first material revenue in Q4 2026, but the stock still rests on assets not yet fully billing and on a build-out that has produced $-0.3B FCF and FY2025 operating margin of -1204.6%. The setup is better than the recent tape suggests, with shares 47.4% below the 52-week high, yet valuation and execution still cap the call at HOLD.
Investment Thesis
SHAZ is one of the stranger setups in the market because the operating story is much better than the income statement. The company has assembled ingredients that many AI infrastructure hopefuls never get: 120 MW contracted under multiyear take-or-pay agreements, 212 MW secured capacity by end-2027, a jump in contracted TCV to $8.8B YTD from $2.2B three months ago, and financing support with $2.2B raised since Dec-25. Those are not cosmetic milestones. They suggest customers are willing to commit ahead of delivery, and management is winning enough supply to matter.
The problem is that almost none of this has flowed into reported financials yet. FY2025 revenue was $0.0B, operating margin was -1204.6%, net margin was -3784.1%, and free cash flow was roughly $-0.3B. On top of that, management's execution record with the Street is poor: three straight misses, capped by 2026-08-06 EPS of -26.16 versus -0.305 expected. That is why the stock, even after falling 47.4% from the 52-week high and trading 15.6% below the 200-day average, does not yet earn a BUY. This is still a trust-me story until revenue actually starts.
The case now depends on three things. First, the company must hit management's statement that, "We expect the first material revenue to commence in the fourth quarter of 2026 as large-scale B300 and GB300 deployments come online." Second, the B300 handover and GB300 schedule must translate into sustained billing, with management targeting full billing on the cluster in Q3 2027. Third, pricing has to stay healthy enough for the future margin profile to justify today's enterprise value, and the record price over $4 per GPU-hour is the best current evidence on that front.
What the market may still be underestimating is that real contracted demand appears to exist, and the stock is no longer priced at the euphoric extreme seen near $97.48. What the market may still be overestimating is how quickly contracted demand turns into recognized, profitable revenue. That balance keeps the call at HOLD. Confirmation of the Q4 revenue start, clean hardware delivery, and signs that EBITDA losses are narrowing as utilization rises would be enough to revisit the rating upward.
Key Metrics
| Metric | Value | Context |
|---|---|---|
| Current price | $51.31 | Reference price for the engine target and scenario framing |
| Price target | $57.15 | 11.4% upside vs current, High confidence, basis sales_quality_blend |
| Quant call | HOLD | High conviction baseline from the quantitative engine |
| FY2025 revenue | $0.0B | 118.7% YoY, but still effectively pre-revenue at reported scale |
| FY2025 EPS | $-2.18 | 1504.5% YoY, still deeply negative |
| FY2025 operating margin | -1204.6% | Prior FY N/A, reflects build-out before revenue ramp |
| Valuation | P/S 277.46, P/B 0.76, P/E N/A | Sales multiple far above peer avg 2.0; peer avg P/E 15.8 |
| Balance sheet | $1.0B debt, 0.90 debt/equity, current ratio 1.63 | Adequate liquidity for now, but leverage matters with negative FCF |
| Free cash flow and capex | $-0.3B FCF, $0.3B capex | Heavy investment phase, no buybacks or dividends |
| Commercial backlog indicators | $8.8B YTD contracted TCV, 120 MW contracted, 212 MW secured capacity | Demand and capacity scaled sharply ahead of revenue recognition |
| Earnings track record | 3 straight misses | Latest was -26.16 vs -0.305, after -1.43 vs -0.875 and -2.85 vs -0.875 |
| Price action | -47.4% vs 52-week high, -15.6% vs 200-day average | Stock has already de-rated materially from peak optimism |
Financial Strength
The balance sheet is better than the income statement, but it is not yet a source of comfort strong enough to offset execution risk. A 1.63 current ratio and the ability to raise $2.2B since Dec-25 show access to capital, which matters in this phase more than accounting earnings. Still, $1.0B of debt and roughly $-0.3B FCF mean the company is burning cash ahead of the ramp, not harvesting it, and margins remain extremely negative because utilization has not arrived. The key financial question is not solvency today, it is whether management can bridge the build-out to material billing without another expensive financing step.
Competitive Position
The core edge is not current profitability, it is assembly of scarce inputs: power-backed AI factory capacity, access to high-end GPU deployments, and customer willingness to sign 3- to 5-year take-or-pay contracts out to 2031. That edge is real today because contracted TCV reached $8.8B YTD and management claims demand exceeds supply, but durability is still unproven because larger infrastructure players can attack pricing and financing if SHAZ stumbles on delivery. The moat gets stronger only if the company turns contracted MW into reliable, on-time billing through 2027.
Management & Guidance
Management's operating guidance is ambitious and, on paper, improved: first material revenue in Q4 2026, more than 64,000 GPUs deployed by mid-2027, 212 MW secured capacity by end-2027, and full billing on the GB300 cluster in Q3 2027. The call itself was directionally strong, with contracted demand and capacity both stepping up, and the explicit risk disclosure helps: CEO James Manning said, "if you say what are the risk factors, it's hardware delivery and data center readiness." The issue is credibility. Reported results have missed the Street in each of the last three quarters, including an extreme August miss, so guidance cannot be taken at face value until the company demonstrates that handovers and deployments translate into recognized revenue on schedule.
Segment Analysis
- AI infrastructure and compute capacity (Not disclosed): Current economics are pre-ramp, with first material revenue expected in Q4 2026 and monetization tied to B300 and GB300 deployments.
- Contracted capacity (120 MW contracted): Multiyear take-or-pay agreements provide the clearest operating KPI until reported revenue becomes meaningful.
- Available capacity (92 MW unsold): This is the near-term commercial upside lever if management converts current demand outpacing supply into signed contracts.
- Geographic footprint (Australia and New Zealand): Management is targeting more than 64,000 GPUs deployed by mid-2027 across this footprint.
Capital Allocation
Capital allocation is currently about survival and scale, not shareholder return. SHAZ has raised $2.2B since Dec-25, including a $1.6B oversubscribed round, has $0.0B buybacks and $0.0B dividends, and is spending heavily with capex of $0.3B while FCF sits at roughly $-0.3B. That is sensible for a pre-ramp AI infrastructure build, but discipline will only be proven once revenue starts and management shows it can fund growth without serial dilution or expensive debt layering.
Positioning & Flows
Ownership is still relatively light and event-driven, with the top three institutions holding only 11.1% and insiders at 0.0%, which limits the signal from long-only sponsorship. Short interest is elevated at 10.63% of float, or 2,536,880 shares, though days-to-cover is only 1.7, so the setup can amplify upside around a clean revenue-start update but does not constitute a deep structural squeeze. Options positioning is described as positive gamma in the front end, which can dampen extreme near-term volatility unless earnings materially reset expectations.
Bull Case
- Contracted capacity is 120 MW under multiyear take-or-pay agreements, and management says 92 MW remains available to sell, giving a visible path to additional bookings before revenue fully ramps
- Secured AI factory capacity rose by 80 MW to 212 MW for end-2027, supported by new Australia partner capacity, which expands the platform faster than prior expectations
- Record price over $4 per GPU-hour on recent B300 deployment suggests pricing power is holding while management says demand materially outpaces supply
- $2.2B capital raised since Dec-25, including a $1.6B oversubscribed round, reduces near-term financing pressure during the heaviest build-out phase
Bear Case
- FY2025 revenue was $0.0B and management still guides only that first material revenue starts in Q4 2026, so today's valuation is being asked to underwrite a business that is not yet proven at scale
- FY2025 EPS was -$2.18, operating margin was -1204.6%, and net margin was -3784.1%, showing the current economic profile remains deeply negative even before the main revenue ramp
- Execution credibility is weak: EPS missed street estimates in all three reported 2026 quarters, including -8477.0% in August, -63.4% in May, and -225.7% in February
- Leverage is not extreme on equity terms but still meaningful for a pre-revenue build, with $1.0B total debt, 0.90 debt/equity, and $-0.3B FCF
Valuation & Price Target
Engine price target $57.15 (+11.4% vs current), High confidence.
- P/S Multiple: $0.08 (1% weight)
- Quality-adjusted: $58.7 (20% weight)
The rating aligns with the quant baseline of HOLD (High conviction). The engine's $57.15 price target implies 11.4% upside from $51.31, but that upside is not enough to overcome the fact pattern of 277.46x sales, $0.0B FY2025 revenue, $-0.3B FCF, and three straight large EPS misses while the business remains pre-ramp. The low-signal sell-side consensus of Strong Buy from a single firm is noted, but it does not outweigh the execution gap between contracted demand and recognized revenue. If SHAZ confirms material Q4 revenue and preserves pricing, the quant target can prove conservative, but that proof is not in reported numbers yet.
Risk Assessment
Risk score 32/100 (Low). The biggest risk is that hardware delivery and data center readiness slip the ramp beyond Q4 2026, because the stock already trades at 277.46x sales despite $0.0B FY2025 revenue.
The Bottom Line
SHAZ is a real AI infrastructure build, not a concept stock, but the reported numbers have not caught up to the narrative. Rated HOLD, High conviction because the commercial signals are strong while valuation and execution risk are still too high for a clean BUY. The stock can work if Q4 2026 truly marks the start of material revenue and if pricing above $4 per GPU-hour holds through the 2027 billing ramp. A move to BUY would require proof of on-time revenue conversion and improving unit economics. A move to SELL would require another material delay, weaker contracting, or evidence that the financing bridge is getting thinner.
Go deeper in the app
Live charts, options analytics, the full earnings-call breakdown, price alerts, and every other stock — updated in real time.
Get Tickr