Hyperscale Data Declares Monthly Cash Dividend of $0.2708333 per Share of 13.00% Series D Cumulative Redeemable Perpetual Preferred Stock

Hyperscale Data declared monthly cash dividends of $0.2708333 per share on its 13.00% Series D cumulative preferred stock and $0.20833 per share on its 10.00% Series E cumulative preferred stock. Both dividends have a September 30, 2026 record date and an October 13, 2026 payment date. The company reiterated its expectation to divest Ault Capital Group in 2027, after which it would focus on AI/HPC data centers, digital assets and Omnipresent Robotics.

Analysis

The stated distributions are contractual preferred obligations rather than evidence of improving operating cash generation; the relevant signal is whether coverage is supported by recurring data-center/hosting EBITDA rather than asset sales, new preferred issuance, or balance-sheet liquidity. Annualized cash requirements are approximately $3.25 per Series D share and $2.50 per Series E share, creating a high fixed-charge hurdle that can constrain capital expenditures needed to compete for AI/HPC workloads. For the common, the preferred stack is economically senior and amplifies downside if Bitcoin-linked cash flows or colocation utilization weaken.

Near term, this is unlikely to re-rate GPUS absent evidence that the next payments are covered by unrestricted cash and operating cash flow. Over 1-3 months, monitor SEC filings for preferred shares outstanding, cash interest/dividend coverage, debt maturities, related-party financing, and any arrearages; a routine declaration has little informational value without those data. The planned separation creates a 6-18 month complexity discount: execution risk, exchange participation uncertainty, and allocation of debt/preferred obligations may make the residual data-center entity harder to value than pure-play AI infrastructure peers.

The contrarian view is that the elevated coupon may attract yield-focused retail demand if the preferreds trade materially below par, but this is a credit/speculation trade—not a validation of the common-equity AI narrative. A sustained improvement in contracted power capacity, utilization, and customer concentration would falsify the cautious view; conversely, any dividend deferral, discount financing, or delay in restructuring would likely pressure both common and preferred securities disproportionately because liquidity is likely limited.

Key Decisions for Investors

• No new GPUS common position on this announcement; treat it as non-catalytic. Revisit only after the next 10-Q provides recurring operating cash flow, unrestricted cash, total preferred obligations, and a credible pro forma post-separation capital structure.
• For existing GPUS common exposure, reduce or hedge into AI/crypto-driven liquidity rallies until fixed-charge coverage is demonstrably above 1.5x on recurring EBITDA or operating cash flow. Thesis is invalidated by verified multi-quarter hosting growth and improved coverage without incremental dilutive financing.
• Place a credit watch on GPUSpD and GPUSpE rather than entering on headline yield. Consider a small, limit-order-only preferred position only if a discount to liquidation preference produces a return that compensates for deferral, liquidity, and restructuring risk; require current price, accrued dividend status, and outstanding-share data before recommending entry.
• Use more liquid AI infrastructure exposure for the theme—e.g., long EQIX or DLR—rather than GPUS until customer contracts, power economics, and separation terms are independently disclosed. The relative thesis fails if GPUS documents superior contracted returns and materially lower leverage than those alternatives.