American Bitcoin's $126 Million Profit Is Really a $197 Million Mark-to-Market Entry

I opened the American Bitcoin release on Wednesday expecting a mining story. What I got was a mark-to-market story wearing a hard hat.

The headline is $126 million in third-quarter net income. That's a swing from a $57.2 million loss in the second quarter. Sounds like a turnaround. It isn't one, at least not in the way the headline implies.

The specification is as follows

American Bitcoin Corp. (Nasdaq: ABTC) booked a $197.3 million gain on its digital assets in Q3. In Q2, that same line was a $71.2 million loss. So the entire swing, and then some, came from revaluing coins the company already held. Mining gross profit was $25.8 million.

Read those two numbers side by side. A $25.8 million gross profit line sitting under a $126 million net income figure. Everything in between is bitcoin price movement, plus a $64.3 million loss on derivatives booked in other expense. Operating income came in at about $191.8 million, which tells you exactly where the gain landed.

That's fair value accounting doing what it was designed to do. It's also the least useful number in the release for anyone trying to figure out whether this company is good at mining.

So look at the actual mining. ABTC produced about 891 bitcoin in the quarter, down from roughly 932 in Q2. Revenue per coin mined was about $71,100, a hair below the $71,900 from the prior quarter. Cost per coin went the other way, from about $36,500 to about $42,100. That's a 15% increase in production cost against a 1% decline in realized revenue per coin. Mining revenue fell to $63.4 million from $67.0 million. Gross margin compressed to 41% from 49%.

Seasonal power costs, per the company. Fine. That's a real explanation and it's also a recurring one. Power is the input cost that never goes away, and every miner in North America deals with summer pricing.

Here's the part most coverage will skip. Of the 8,427 BTC on the balance sheet at September 30, about 3,090 coins are pledged as collateral for miner purchases under agreements with Bitmain. That's roughly 37% of the reserve encumbered. Not sold, not gone, but committed. In a quarter where the profit exists only because of the coin price, the fact that more than a third of the pile is collateral matters. Anyone modeling this as a clean treasury holding is modeling it wrong.

The reserve did grow. Holdings rose about 425 BTC, or 5%, from roughly 8,002 at the end of June. And the number management actually wants you to watch, satoshis per share, climbed to about 11,572 from about 10,989. That's the metric that survives a price collapse, because it measures coin accumulation per unit of equity rather than dollar revaluation.

What the sats number actually proves

Chief executive Mike Ho made a point of saying the company raised sats per share about 5% without issuing shares through its at-the-market program and without buying bitcoin on the open market. That's the strongest line in the whole release, and it's the one nobody will put in a headline.

Why does it matter? Because a lot of treasury companies grow coin-per-share by printing equity into a premium and swapping paper for bitcoin. That works until the premium closes. ABTC grew the metric through production alone. Miners were running, coins came out of the ground, some production got sold to cover operating costs, and the net position still went up.

Backward compatibility is maintained except where noted below, and here's the note. The company effects a 1-for-15 reverse stock split on July 2 of this year. Share counts before and after that date aren't comparable, and any per-share figure you pull from a screener needs to be checked against the split-adjusted base. This is the kind of detail that quietly breaks a spreadsheet model.

The owned fleet stood at about 89,242 machines with roughly 28.1 EH/s of capacity. Hut 8 remains the majority owner and confirmed on the same day that it's been selling bitcoin to fund buybacks while pushing mining operations into the American Bitcoin vehicle. The Trump family stake has been reported near 20%, with Eric Trump serving as co-founder and chief strategy officer.

The market already did the math

Bitcoin traded above $80,500 while this was being written. ABTC shares were down more than 5.5% on the day.

Sit with that. A company reports $126 million of net income and $221 million of Adjusted EBITDA, against an Adjusted EBITDA loss of $45.0 million a quarter earlier, and the stock falls 5.5%. The market isn't confused. It's pricing the miner, not the accounting entry.

So here's my first strong opinion. Any analyst who quotes that $221 million Adjusted EBITDA figure without immediately quoting the $197.3 million digital asset gain in the same breath is doing marketing, not analysis. Adjusted EBITDA that includes unrealized coin appreciation isn't a measure of operations. It's a measure of the balance sheet with a flattering name attached.

And the second. The treasury company model got popular because it let public vehicles sell a story about coin accumulation at a premium. That premium only holds when the coin price is climbing and capital is cheap. With bitcoin at $80,500 and production costs rising 15% quarter over quarter, the arbitrage is thinner than the pitch deck assumed. ABTC is fine. The structure is under strain.

Would you rather own a miner generating $25.8 million of gross profit, or a bitcoin fund that happens to run ASICs? Those are different businesses with different risk profiles, and the income statement doesn't distinguish between them.

What to do with this

Ignore net income. It's a price feed with a tax adjustment. Track cost per coin mined, which went from $36,500 to $42,100 and will keep climbing if difficulty does. Track sats per share, which went up 5% honestly. And track how much of the reserve is pledged, because 3,090 coins tied up in hardware financing is a use fact, not a footnote.

What happens to a $197 million gain when the price moves the other way? You get Q2, which is exactly what this company reported three months ago. That's the whole cycle in two quarters. The mining operation is real and the satoshi accumulation is real. The profit is a translation of price into dollars, and it reverses.

Developers should note the breaking change in the return type. For everyone else, the takeaway is simpler. When a miner reports a nine-figure profit in a quarter where its gross margin fell eight points, the profit isn't the story. The margin is.