Red Flags

MARA — MARA Holdings, Inc.

MARA Holdings, Inc. operates as an energy and digital infrastructure company in North America, the Middle East, Europe, and Latin America. The company leverages Bitcoin Mining and Artificial Intelligence compute to monetize excess energy and underutilized power, optimize power management across operations and support AI inference applications. The company was formerly known as Marathon Digital Holdings, Inc. and changed its name to MARA Holdings, Inc. in August 2024. MARA Holdings, Inc. was incorporated in 2010 and is based in Hallandale Beach, Florida.

**THE CORE CONCERN**

MARA’s Beneish M-Score of −1.57 breaches the −1.78 manipulation threshold. On this negative scale, the value lies closer to zero than the cutoff and therefore registers as statistically consistent with earnings manipulation in the model’s original calibration. The signal is driven primarily by a DSRI of 2.11, indicating accounts receivable expanded 111 percent faster than revenue. Revenue rose from $0.1 B to $0.9 B before easing to $0.8 B, producing a four-year CAGR of 61.7 percent; the same period shows an Altman Z-Score of 0.09 (well below the 1.81 distress boundary) and short interest at 31 percent of float. These ratios together constitute the most direct quantitative red flag.

**HISTORICAL PRECEDENT**

The M-Score model assigned comparable readings to Enron in fiscal 1997 and 1998, three years before its 2001 collapse. In that case, the score reflected aggressive revenue recognition and understated liabilities that were later restated. While the underlying business models differ, the shared pattern is an M-Score breach accompanied by rapid reported revenue growth and elevated receivables relative to sales.

**WHAT TO VERIFY**

Investigators should examine three items in the 10-K or 10-Q. First, the aging schedule and collection terms for any digital-asset or hosting receivables to determine whether the DSRI elevation exceeds normal payment cycles in the sector. Second, the reconciliation between reported hash-rate capacity additions and actual revenue recognized in each quarter, to test whether equipment financings or customer deposits are being recorded prematurely. Third, the composition of “other current assets” and any related-party Bitcoin lending or custody arrangements that could affect both accruals and cash-flow classification.

**COUNTERARGUMENTS**

MARA operates in a high-growth, capital-intensive industry where many forensic ratios are expected to trigger without manipulation. Revenue expanded at a 61.7 percent CAGR over four years; such expansion routinely produces elevated SGI and DSRI values in legitimate firms shifting to larger contracts or longer settlement cycles. The company remains pre-profitability, so negative operating cash flow and high accruals can reflect investment-phase economics rather than accounting distortion. In addition, the Altman Z-Score was calibrated on manufacturing firms and is acknowledged to be less reliable for financial-services or digital-asset businesses whose balance sheets are dominated by volatile crypto holdings. Short interest at 31 percent may simply reflect sector-specific skepticism about Bitcoin price exposure rather than accounting concerns. These factors indicate that the quantitative signals warrant examination but do not by themselves establish manipulation.

BTDR — Bitdeer Technologies Group

Bitdeer Technologies Group operates as a technology company for blockchain and high-performance computing (HPC) in Singapore, the United States, Bhutan, Norway, Finland, Ethiopia, Canada, and internationally. The company offers hash rate sharing solutions, including cloud hash rate and hash-rate subscription plans; and a one-stop mining rig hosting solution comprising deployment, maintenance, and management services for cryptocurrency mining; as well as mining cryptocurrencies for its own account. It also operates mining datacenters to generate hash rates; handles various processes involved in computing, such as equipment procurement, transport logistics, datacenter design and construction, equipment management, and daily operations; and sells mining rigs. In addition, the company offers M

**THE CORE CONCERN**

The Beneish M-Score of 2.48 lies closer to zero than the −1.78 manipulation threshold, placing the company in the statistical range the model associates with earnings manipulation. The score is driven by a DSRI of 1.84 (receivables expanding 84 percent faster than revenue) and a TATA of 0.643 (accruals equal to 64.3 percent of total assets). These inputs coincide with an Altman Z-Score of 1.71, inside the distress zone below 1.81, and short interest at 30 percent of float. Revenue has expanded from $0.3 billion to $0.8 billion over four years at a 24.9 percent CAGR, with the most recent year showing 77 percent growth (SGI 1.77). While the absence of any year in which positive net income coincided with negative operating cash flow and a five-year cumulative NI–OCF divergence of zero limit the classic earnings-quality red flags, the combination of rapid top-line expansion and elevated accrual ratios still warrants examination of revenue recognition and working-capital quality.

**HISTORICAL PRECEDENT**

The pattern most closely tracks the pre-collapse profile of Enron, where the Beneish model produced an M-Score above the −1.78 threshold in fiscal 1997–1998, three years before the 2001 bankruptcy. Enron also exhibited rising receivables relative to revenue and high total accruals during periods of reported growth, later shown to reflect aggressive contract accounting and off-balance-sheet financing rather than sustainable cash generation.

**WHAT TO VERIFY**

Investigators should examine three items in the 10-K or 10-Q. First, the aging schedule and concentration of accounts receivable, specifically whether days-sales-outstanding has lengthened materially beyond industry peers and whether any single customer or related-party balance exceeds 10 percent of total AR. Second, the reconciliation between reported revenue and cash collections from customers, with particular attention to any deferred-revenue or contract-asset movements that could indicate front-loaded recognition. Third, the composition of the $0.643 TATA figure—whether it stems from capitalized development costs, cryptocurrency inventory valuation, or changes in long-term contract accounting—and how those items affect future cash outflows.

**COUNTERARGUMENTS**

High revenue growth itself can produce the observed forensic ratios without manipulation. A 24.9 percent CAGR and 77 percent year-over-year increase naturally elevate the SGI and DSRI components; enterprise-software and infrastructure firms shifting to larger, longer-term contracts routinely report rising AR-to-revenue ratios. The zero NI–OCF divergence and near-zero accruals ratio (NI−OCF)/Assets of 0.000 indicate that reported earnings have translated into cash without material timing differences over the multi-year period. In this setting, the M-Score breach and Z-Score distress reading may largely reflect the accounting characteristics of a high-growth technology business rather than deliberate distortion of results.