Ault & Company Has Invested $55 Million in Capital to Hyperscale Data

Ault & Company invested $55 million in Hyperscale Data convertible preferred stock and, including affiliates, now beneficially owns approximately 62% of the company; it also holds rights to purchase up to another $96 million of Series H preferred shares. The investment thesis centers on Hyperscale's Michigan AI data center, where a 20 MW customer agreement could generate more than $1.2 billion over 20 years, with a potential 52 MW expansion implying more than $3.0 billion in contract revenue. Management estimates the campus could be worth $750 million-$1.25 billion, though this is not independently appraised and expansion remains conditional on financing, approvals, utilities and customer demand.

Analysis

The relevant equity question is not the headline asset-value range but whether the contracted load converts into financeable, recurring EBITDA after power, network, cooling, debt service and build-out capex. The implied initial contract run-rate is roughly $3m per MW annually, a potentially attractive benchmark versus mature wholesale colocation pricing, but the undisclosed counterparty, service-level obligations, commencement date, customer credit and whether revenue is pass-through-heavy are decisive. Until those terms are independently disclosed, a campus valuation based on gross lifetime revenue is not a reliable NAV anchor.

The preferred investment improves near-term liquidity but creates a material common-equity overhang: a controlling holder can fund development through additional preferred issuance, potentially subordinating or diluting public holders before any asset monetization. The same control structure also makes a related-party asset sale, refinancing, or conversion-price amendment a governance risk rather than a clean catalyst. The mixed operating-asset structure and Bitcoin exposure further complicate a pure AI-infrastructure multiple; GPUS is likely to trade on financing credibility and float/liquidity more than on a conventional data-center NAV framework.

Over days, the announcement can support a scarcity/AI-capacity narrative in a thinly traded security, but that is not sufficient for institutional underwriting. The 1-3 month catalyst path is evidence of energized capacity, customer identity/credit quality, project-level financing, and a disclosed economic model for the initial deployment. Over 6-18 months, value realization requires utility interconnection and construction execution; failure to secure non-dilutive capital or a delay in service commencement would likely compress the implied asset valuation sharply.

The contrarian view is that the market may appropriately discount management's appraisal rather than overlook an obvious arbitrage: a control-holder's internal valuation is not a third-party bid, and a large nominal contract can have limited equity value if capex and financing costs absorb cash flow. Conversely, confirmation that the initial customer is investment-grade and that the project reaches service with limited incremental common dilution could force a substantial rerating because a separately financed, contracted asset would be easier to value than the current conglomerate structure.

Key Decisions for Investors

• No core long in GPUS before reviewing the preferred-stock terms, conversion mechanics, liquidation preference, accrued dividends and any right to issue the remaining preferred capacity. Treat these as gating data; adverse conversion/reset provisions would invalidate an NAV-based common-equity thesis.
• Place GPUS on a catalyst watchlist for the next 1-3 months: initiate only after disclosure of customer credit, commercial-operation timing, initial 20-MW capex, power cost and project-level financing. A position should be sized as a microcap special situation, with entry contingent on verification rather than the press-release valuation range.
• If the stock rallies materially ahead of those disclosures, consider a tactical short only where borrow and liquidity permit; the downside catalyst is financing or interconnection slippage. Cover on independently verified service commencement, named creditworthy customer disclosure, or a third-party asset transaction near the stated valuation floor.
• For AI-infrastructure exposure while GPUS remains unverified, prefer liquid beneficiaries with clearer contractual economics—EQIX, DLR, VRT or ETN—rather than using GPUS as a proxy. Reassess relative exposure if GPUS demonstrates contracted EBITDA and non-dilutive financing.
• Monitor Bitcoin/cash disclosures separately from data-center execution. A sharp BTC decline, restricted-cash usage, or cash burn that necessitates further affiliated preferred funding would be a near-term signal to avoid or reduce GPUS exposure.