Bitcoin mining concentration and hardware efficiency trends

Foundry USA commands 17.17% of the global hashpower with 29.82 EH/s. This concentration leaves the rest of the market to compete for the remaining 82.83% of processing power. Antpool follows with 14.78% of the hashpower, which equals 25.67 EH/s. Viabtc holds 14.57% and F2pool controls 13.26%. These top entities, along with Binance Pool at 12.17%, create a landscape where a few players hold massive influence. I observe how this concentration of power limits the leverage of smaller, independent mining operations.

Hardware efficiency and the ROI gap

The choice between efficiency and raw output determines whether a miner remains profitable. Bitmain’s Antminer S23 series provides up to 1160 TH/s with 9.5 J/TH efficiency. In contrast, the MicroBT M79S produces 1.35 PH/s but draws 20,000 W of power. While Bitmain’s Antminer S23 series delivers 1160 TH/s at 9.5 J/TH, the MicroBT M79S consumes nearly double the wattage to achieve its 1.35 PH/s output, which forces many operators to choose between raw power and electricity savings. A calculation for 1 PH/s of hashing power on S23-class machines shows a consumption of 2,640 kWh per day, whereas the M63S-class uses 5,136 kWh for the same work. This difference costs an operator $132 versus $257 every day if they pay $0.05 per kWh. At $0.10 per kWh, the cost difference grows to $264 versus $513 per day.

Bitdeer offers the SealMiner A2 Pro at $7,500, which provides 500 TH/s. This model costs $15 per TH. Newer machines require over 1,000 days to cover their acquisition costs. This timeframe creates a problem since the next halving arrives in approximately 850 days. I find that the industry leans heavily toward the S23 series because it delivers better profit margins in high electricity-cost environments. For example, the Antminer U3S23 costs $13.22 per day to run at $0.05/kWh, whereas the M79S costs $24.00 per day. Bitmain’s S21 series also remains relevant, with the S21e XP Hydro priced around $9,245 per unit. The S21 series offers lower efficiency than the S23, but its lower upfront cost makes it an option for those with cheap electricity. MicroBT’s M63S has fallen to resale prices between $3,000 and $5,000, which places it below $10 per TH. This market allows older, less efficient machines to stay online if electricity rates remain near $0.03 or $0.04 per kWh.

AI infrastructure shift in Texas

Core Scientific shifts its focus to artificial intelligence and high-performance computing. The company converts its 300 MW facility in Pecos, Texas, into an AI data center campus. This expansion aims for a capacity of 1.5 GW by 2027. The company also modifies 100 MW of its infrastructure to provide 70 MW to CoreWeave for high-performance computing operations. This new 12-year contract adds $1.225 billion in projected revenue to the more than $3.5 billion in revenue from previously announced contracts. I observe that these large-scale shifts in energy use toward AI graphics processing units pressure the available power for Bitcoin miners. The company in Austin, Texas, already delivered 16 MW of capacity to CoreWeave. These shifts in demand for energy-intensive facilities change the competition for electrical contracts in regions with constrained infrastructure.

Difficulty adjustments and margin erosion

Mining margins face intense pressure because the hashprice fell to $35/PH/s in November. This level is a significant drop from the $55/PH/s seen in the third quarter. The network difficulty also fell 19.9% in July 2026. This difficulty drop follows a period where the global network hashrate stayed above 180 EH/s. The total network hashrate grew by 163% over the five months preceding the end of the year. Bitcoin trades at $87,061. The network remains competitive despite the price movements. I see that the remaining miners face less competition for new blocks after the difficulty adjustment, but they still struggle with rising costs. The block reward remains at 3.125 BTC, but the dollar value of that reward fluctuates with the market. Can the industry manage these shrinking margins before the next major market shift?