Bitcoin 2026 halving aftermath: miner capitulation and MARA expansion

Miner profit margins and the capitulation myth

Miner profit margins drop to 4.67 percent as Bitcoin price trends toward production costs. Capriole Investments puts production cost at $61,200 and electrical cost at $48,965. These figures mean miners are barely breaking even. The Puell Multiple averaged 0.73 over the last 30 days, placing it in the 16th percentile. Killa on X says this is a capitulation signal. The network mining difficulty is 18.3% below its November 2025 peak, which is the largest drawdown since the 2021 China ban. This difficulty drop coincides with a hash rate that hit 897 exahashes per second in August 2026 but recovered toward 925 exahashes per second. The April 2024 halving event reduced mining rewards from 900 units to 450 units per day. Miners must compensate for this 50 percent reduction in revenue by keeping costs down. Some estimates suggest mining costs could reach $40,000 after the halving. The Bitcoin price was $58,524 at quarter-end, which is down from $107,173 a year earlier. Does the recovery in hash rate prove the capitulation myth is false?

Marathon’s infrastructure pivot and debt management

MARA Holdings reported a net loss of $611 million in its second-quarter 2026 results, which missed the consensus estimate of $208.37 million by 16.1 percent following a period of significant Bitcoin price volatility and a 28 percent decline in the average price of Bitcoin. This loss follows a period where the company posted an $808 million profit in the year-earlier period. Management says $343 million of this loss comes from unrealized mark-to-market changes on digital assets. The company’s revenue fell 27 percent year-over-year to $175 million. It aims for a 4.8 gigawatt total potential capacity through sites like the 2-gigawatt Matagorda County project. The Matagorta site covers 1,200 acres in Texas and costs $600 million to purchase. Management expects this site to reach 1 gigawatt of grid capacity by October 2027 and 2 gigawatts by 2028. The Long Ridge Energy acquisition adds 1 gigawatt of potential capacity and 505 megawatts of contracted power. This $1.5 billion enterprise value acquisition uses Bitcoin-backed debt. The company’s energized hashrate increased 22 percent year-over-year to 70.3 exahash per second, and Bitcoin production rose 3 percent to 2,422 BTC. The company improved operational efficiency, cutting its cost per petahash per day by 4 percent to $27.7. Bitcoin holdings fell 29 percent to 35,577 BTC from 49,951 BTC a year earlier. Management uses these assets to fund an expansion into AI infrastructure and power generation. The company reported that 26% of its Bitcoin holdings were activated as of quarter-end, with 4,742 BTC loaned and 4,528 BTC pledged as collateral. Subsequent to the quarter, the company pledged an additional 18,750 BTC as initial collateral for two bitcoin-backed credit facilities. I see this as a large expansion into AI infrastructure and power generation.

Comparing MARA Holdings and Riot Platforms

I compare MARA Holdings and Riot Platforms using current market data.

MARA Holdings reports a higher trailing twelve-month net income of $927 million, which is above the $164 million reported by Riot Platforms. RIOT holds a higher market capitalization of $4.72 billion compared to the $3.53 billion held by MARA. MARA has a debt-to-equity ratio of 70 percent, while Riot maintains a lower 25 percent. I find the 3.63 price-to-earnings ratio for MARA more appealing than the 24.42 ratio for RIOT. RIOT shows 13.9 percent revenue growth, while MARA shows a decline of 26.7 percent. Riot Platforms operates in two segments, Bitcoin mining and engineering, and it produces electrical products for data centers and utility markets. MARA focuses on energy and digital infrastructure across four continents. You should look at the volatility if you want to trade these miners. MARA’s beta is 5.34 and RIOT’s beta is 3.83. I note that MARA’s net profit margin is -429.7 percent, which is much lower than RIOT’s -196.3 percent margin.

The shift in treasury strategies

Strategy authorized a $1.25 billion Bitcoin Monetization Program to handle its $1.5 billion annual dividend load. The company sold 3,588 Bitcoin to pay dividends in July 2026. This sale is only 0.4 percent of its 843,775 Bitcoin holdings. Strategy also paused its Bitcoin buying in late March 2026 after a 13-week streak. The company sold Bitcoin at average prices between $59,256 and $60,773. These coins were sold at a 20 percent loss against an average purchase cost of $75,476. This transaction is a small fraction of the total stack. Strategy’s holdings include 843,775 Bitcoin, which is 4.2 percent of the total supply. Strategy adopted a Digital Credit Capital Framework to manage its $2.55 billion reserve for dividends. The company says the full $1.25 billion program capacity remains untouched by the July sale. This shift in treasury management changes the behavior of all companies built on the never-sell story. The regime began in August 2020 when the company started parking cash in Bitcoin as an inflation hedge. It then moved to issuing convertible bonds and later preferred stock to fund purchases. The sale of 3,588 Bitcoin was the third disposal since a small tax-loss sale in 2022. The amount is a rounding error relative to the total treasury. The company insists the long-term thesis is untouched, but markets price regimes rather than mechanics. This marks the end of the era where accumulation was the only direction for the most influential crypto trade.