Summary

  • Hyperscale Data announced a 20,000,000-share Class B distribution across about 491,795,085 shares of eligible capital, a current payment ratio of approximately 0.04066734 per eligible share.
  • Applying that ratio to the disclosed 165,267,650 Class A shares yields about 6,720,996 Class B shares. The press release instead says Class A holders will receive 7,194,786, leaving roughly 473,790 shares unexplained by the published inputs.
  • The security carries ten times the voting power of Class A but brings the company no construction cash except fractional-share settlement. That ledger is separate from a Michigan deployment expected to require more than $100 million over time.

A dividend announcement normally invites one quick question: how much does each holder receive? Hyperscale Data's 4 September filing makes that question unusually difficult because its own aggregate figures do not reconcile.

The Form 8-K says the company plans to distribute 20,000,000 Class B shares to holders of common stock, four series of convertible preferred stock and a convertible note, with the latter instruments counted on an as-converted basis. The record date is 15 September. Payment is set for 6 October, subject to adjustment.

The event is still prospective. More importantly, the current allocation table needs a bridge.

Five published numbers do not close one ledger

The press-release exhibit lists 165,267,650 Class A shares outstanding, 4,774,348 Class B shares and approximately 321,753,087 Class A equivalents from the preferred stock and note. Those components add exactly to the disclosed eligible-capital total of 491,795,085.

Dividing the fixed 20,000,000-share pool by that denominator produces 0.0406673442 Class B share per eligible share, consistent with the issuer's rounded ratio of 0.04066734.

The next line breaks the reconciliation. A uniform pro-rata allocation to 165,267,650 Class A shares produces about 6,720,996.4 Class B shares. Multiplying the Class A count by the issuer's rounded ratio gives about 6,720,995.7. Yet the release separately says 7,194,786 of the distributed shares will go to Class A holders.

The difference from the exact pro-rata result is approximately 473,789.6 shares, or 473,790 when rounded to a whole share. At the disclosed uniform ratio, the larger allocation would correspond to roughly 176.9 million eligible units—about 11.65 million more than the disclosed Class A count.

This is a reconciliation question, not proof of misconduct or of any holder's final entitlement. There may be a classification, timing or transfer-agent convention not described in the release. Cash in lieu of fractions will also depend on holder-level positions. But none of those mechanics is supplied as the missing bridge, and the published aggregate cannot be reproduced from the published inputs.

Future issuance is not an answer to the current table

Hyperscale Data warns that it expects more eligible capital stock to be issued before the record date. If that occurs while the pool remains fixed, the payment ratio will fall. That is a useful dilution warning.

It does not, on its face, explain the present difference. The share counts are introduced as of 4 September, the eligible total is calculated from them, the ratio is described as current, and the Class A allocation appears in the same calculation. A future larger denominator would reduce—not increase—the allocation for a fixed Class A count unless another relevant count also changes.

The next disclosure therefore needs more than a revised ratio. It needs a table that names every eligible class, the units attributed to each, the treatment of conversion limits, the allocation before fractional settlement and the sum back to 20,000,000.

A voting instrument is not operating liquidity

The distributed Class B stock is economically unusual. Each share carries ten times the voting power of one Class A share. It can be converted one-for-one into Class A at any time after the payment date. There is currently no public trading market for Class B, and the issuer says only that it may seek an NYSE American listing, with no assurance that or when one will occur.

Those terms create a choice between voting weight, convertibility and uncertain market liquidity. They do not create cash for Hyperscale Data. The only announced cash element is payment in lieu of fractional Class B shares, which is a settlement detail rather than project financing.

That distinction is material because the company's latest Form 10-Q describes a separate capital constraint. At 30 June, Hyperscale Data reported about $36.8 million of cash and $28.3 million of restricted cash against approximately $201.7 million of current liabilities, including about $99.3 million of current notes and related borrowings. It recorded a six-month net loss of about $49.1 million and used about $9.9 million in operating cash.

Management expects to invest more than $100 million over time to construct, equip and commission service areas and power modules for a phased deployment of approximately 20 MW of critical capacity at the Michigan data centre. The filing says existing cash and other presently available liquidity are not expected to cover the next twelve months of anticipated requirements, obligations and planned capital expenditure, and that substantial doubt about going concern has not been alleviated.

The dividend does not worsen or solve that conclusion by itself. It belongs to the ownership-and-vote ledger. Equity sales, debt, asset monetisation, customer receipts and actual construction spending belong to the cash ledger. Calling both “capital” would hide the fact that one changes securities while the other must pay invoices.

Sources