American Bitcoin reported $153 million annual loss after building 5,401-BTC reserve

On February 26, 2026, American Bitcoin Corp. filed its fourth-quarter and full-year 2025 results with the U.S. Securities and Exchange Commission, reporting a $153.2 million annual net loss while ending the year with 5,401 bitcoin. The filing made the tension in its model unusually clear: mining and equity-funded purchases had built a large reserve, but fair-value accounting transmitted bitcoin’s decline directly into reported earnings.

This is a 2026 reconstruction of the dated record, not a claim that Coinburn published an article on February 26, 2026. The figures below are those available in the event-day filing and contemporaneous reporting.

A larger operation, and a larger accounting loss

American Bitcoin reported 2025 revenue of $185.164 million and a net loss of $153.171 million. Its filed statement attributed $227.064 million to a loss on digital assets, partly offset by other items including a $56.318 million gain on derivatives and an $18.115 million income-tax benefit. The company characterized the digital-asset charge as a non-cash, mark-to-market loss required by fair-value accounting.

For the fourth quarter ended December 31, 2025, the company reported $78.3 million in revenue, up from $64.2 million in the third quarter. That is a roughly 22% sequential increase, matching the company’s calculation. Reuters reported a fourth-quarter net loss of $59.45 million, compared with a $3.48 million profit a year earlier.

The annual comparison needs care. American Bitcoin said periods before March 31, 2025 reflected Hut 8’s former bitcoin-mining subsegment and were not directly comparable with the later standalone business. Revenue growth therefore did not represent a clean like-for-like history of one unchanged company.

The reserve was the strategy

The company said it held 5,401 bitcoin on December 31, including 2,776 bitcoin pledged or otherwise collateralized. From the beginning of the second quarter through year-end, it mined 1,654 bitcoin, including 783 in the fourth quarter. It said roughly one-third of year-end holdings came from mining and the remainder from strategic transactions and market purchases.

Operating scale was substantial on the company’s measurements: about 25.0 exahashes per second of installed capacity across roughly 78,000 ASIC miners, with approximately 21.9 EH/s operational at year-end and average fleet efficiency of about 16.3 joules per terahash. The distinction between installed and operating hashrate matters; installed machines do not all contribute production at a given moment.

American Bitcoin also reported $150.5 million in gross proceeds from its at-the-market stock program during the fourth quarter. That linked reserve growth to capital-market access as well as mining economics. For shareholders, the model combined bitcoin exposure, operating execution, equipment and power risk, collateral arrangements, and the possibility of dilution from new shares.

What the 53% claim did—and did not—show

Management said a 53% fourth-quarter gross margin meant the company accumulated mined bitcoin at a 53% discount to spot prices during the period. That was a company-defined interpretation of accounting gross margin, not evidence that every coin cost 53% less than one specified exchange price. The release did not identify a single spot venue, timestamp or volume-weighted benchmark for that comparison, and cost of revenue excluded depreciation and amortization shown elsewhere in the statements. It should therefore be read as management’s production-cost framing, not an independently verified all-in acquisition price.

The February 26 record supports a narrower conclusion: American Bitcoin had scaled mining and treasury holdings quickly, while its results demonstrated how a bitcoin-heavy balance sheet could dominate GAAP earnings. The filing also said holdings had since exceeded 6,000 bitcoin, but supplied no exact event-day count in the narrative. The next questions were how much of the reserve remained pledged, how efficiently installed capacity became operating hashrate, and whether additional accumulation would depend on further equity issuance.