S21 vs M60 vs A15 Full Analysis: Will Bitcoin Mining Still Be Profitable in 2026?

The fourth Bitcoin halving in April 2024 reduced block rewards from 6.25 BTC to 3.125 BTC, and now, more than two years later, the industry is still feeling its effects. This supply-side shock, which has been widely discussed in the market, entered a structural stress-testing phase after the Bitcoin price hit a historic high of about $126,000 in October 2025. As of May 27, 2026, Gate market data shows Bitcoin trading at $75,773.2, marking a cumulative decline of over 31% from its peak in the past year. Meanwhile, hashprice has dropped from roughly $100/PH/day before the halving to about $33.30/PH/day—a decrease of more than 66%.

Against this backdrop, the question "Is Bitcoin mining still profitable after the halving?" has become a central concern for industry participants.

The Mining Reality of 2026: Triple Compression Forces

The profit squeeze following the halving continues to intensify two years on. In the first half of 2026, the Bitcoin mining industry faces three structural pressures simultaneously: the permanent halving of block rewards, competition from network hashrate operating at historic highs, and rising energy costs in major mining hubs worldwide.

Structural Industry Impacts

As of Q1 2026, the weighted average cash cost for publicly listed mining companies is about $77,000 per Bitcoin. When factoring in hardware depreciation, site maintenance, and corporate management expenses, the "total cost" rises above $100,000 per Bitcoin. JPMorgan’s estimate is relatively conservative at around $77,000. With the current market price at approximately $75,773.2, the industry as a whole is operating at a deep loss.

Hashprice—a key metric for daily miner revenue per unit of hashrate—has fallen from about $36–38/PH/day in Q4 2025 to roughly $33.30/PH/day in early May 2026, nearing the historic low of $28 set on February 23.

Within the industry, opinions about the future of mining are increasingly polarized:

• Pessimists argue that the structural inversion, with average production costs well above market prices, will force miners to continually sell their Bitcoin reserves to cover electricity and operational expenses, creating sustained selling pressure. Data shows that publicly listed miners sold over 32,000 Bitcoins in Q1 2026—more than the total for all of 2025.
• Neutral observers point out that the rapid clearing of inefficient hashrate is a necessary adjustment for network health. After hashrate exits, difficulty adjustments can improve profitability for remaining miners. The network has seen multiple downward difficulty adjustments in 2026.
• Optimists focus on the narrative of diversified business transformation, such as AI hashrate hosting, which is reshaping mining revenue streams. Statistics show that several listed mining companies have announced AI and HPC contract totals exceeding $70 billion.

Timeline and Context

Here’s a timeline of key events since the halving:

• April 2024: The fourth Bitcoin halving occurs; block rewards drop from 6.25 BTC to 3.125 BTC.
• Second half of 2025: Industrial electricity rates rise in major mining hubs, combined with increased mining hardware depreciation, pushing the industry’s breakeven line higher.
• October 2025: Bitcoin price reaches a historic high of about $126,000; network hashrate peaks at roughly 1,160 EH/s.
• December 2025: Regulatory actions intensify in China’s Xinjiang region, combined with rising winter energy costs, causing hashrate to fall.
• February 2026: On February 23, hashprice drops to $28/PH/day, setting a new all-time low.
• March 2026: Hashprice compresses further to around $29.15, then rebounds after difficulty adjustments.
• Early May 2026: Hashprice recovers slightly to about $33.30/PH/day; the network’s 30-day average hashrate is approximately 968 EH/s.

2026: A Comprehensive Comparison of Mainstream Mining Hardware

With hashprice under sustained pressure, mining hardware efficiency—measured in "joules per terahash" (J/TH)—has become the key determinant of mining profitability.

Below is a comparison of core parameters for three major SHA-256 ASIC miner series in 2026:

Note: According to industry analysis and manufacturer specifications, some models have reasonable ranges for their parameters.

Efficiency Segments: These miners can be grouped by efficiency: older models around 20 J/TH; mainstream current models around 15 J/TH (such as S21, M60S, A1466); and new flagship models around 10 J/TH (like the S23 series).

S21 Series: Leading Efficiency, Broad Coverage

Bitmain’s Antminer S21 series is currently the most widely deployed new-generation mining hardware. The S21 Pro uses 5nm chip technology, delivering 234 TH/s at about 15.0 J/TH. The S21 XP further optimizes efficiency to 13.5 J/TH. The mass-produced S23 Hydro (9.5 J/TH) covers both air and liquid cooling approaches within the product lineup.

The S21 series stands out for its comprehensive product range and mature supply chain.

M60 Series: MicroBT’s Steady Competitor

MicroBT’s WhatsMiner M60 standard model offers about 19.9 J/TH at 172–186 TH/s; the upgraded M60S+ improves efficiency to roughly 16.5 J/TH. Industry analysis notes that the M60 series strikes a strong balance between hashrate and efficiency, continuing MicroBT’s tradition of "balanced configuration."

In 2026, the M60 series’ core value may lie less in extreme efficiency and more in supply chain stability as the "second supplier." For large-scale mining operations, avoiding over-reliance on a single manufacturer is a key risk management consideration.

A15 Series: Canaan’s Diversified Options

Canaan’s Avalon A15 series covers an efficiency range of 15.5–18.8 J/TH. The A15 Pro+ achieves about 15.5 J/TH, while the standard A15 is around 18.8 J/TH. The A15 XP reaches up to 212 TH/s, and the high-end A1566I model delivers 261 TH/s at a power draw of 4,500 W.

The A15 series prioritizes "diversified choices," offering differentiated hardware solutions across price segments rather than competing for absolute efficiency. For small and medium-sized mining farms seeking to diversify supplier risk, the A15 series provides a third path between Bitmain and MicroBT.

Choosing Between Liquid-Cooled Flagships and Standard Air-Cooled Models

New-generation liquid-cooled flagship models (like the S23 Hydro 3U, 9.5 J/TH, 1,160 TH/s) deliver superior efficiency compared to air-cooled units. However, liquid cooling requires more advanced infrastructure: water cooling loops, three-phase high-voltage power, and complex thermal management. In contrast, air-cooled flagships like the S21 XP (13.5 J/TH, 270 TH/s) can be deployed directly in standard data center racks, making them much easier to install.

Breakeven Calculations Across Electricity Price Tiers

A miner’s actual profitability is determined by a combination of hashrate, efficiency, network difficulty, and electricity price. The following calculations use network parameters as of late May 2026, with Bitcoin priced at $75,773.2 and hashprice at about $33.30/PH/day, based on Gate market data.

Core Calculation Formulas and Assumptions

• Daily BTC Revenue = Miner hashrate (PH/s) × hashprice ($/PH/day)
• Daily Power Consumption = Miner power (kW) × 24 hours

Assume hashprice is about $33.30/PH/day, excluding mining pool fees. It’s important to note that hashprice fluctuates with Bitcoin price, network hashrate, and transaction fees; these calculations are reference values based on current market conditions, and actual profitability will vary as network conditions change.

Daily Net Profit for Each Model at Different Electricity Prices

Note: Daily revenue is calculated based on hashprice of $33.30/PH/day. Daily net profit equals daily revenue minus daily electricity cost (electricity price in $/kWh). Positive values indicate profit; negative values indicate loss. Some models’ income may fluctuate due to differences in hashrate and power parameters.

With hashprice at $33.30/PH/day and Bitcoin priced at about $75,773.2, the most efficient S21 XP (13.5 J/TH) delivers a daily net profit of about $6.49 per unit at ultra-low electricity rates of $0.04/kWh. Even at a high electricity price of $0.10/kWh, this model can still generate about $1.42 in daily net profit per unit. In contrast, the Avalon A15 standard model (18.8 J/TH) turns unprofitable when electricity exceeds $0.06/kWh, losing more than $2 per day at $0.08/kWh.

The S23 Hydro 3U, as a liquid-cooled flagship, excels with 9.5 J/TH efficiency, maintaining daily net profit of about $20.34 per unit even at $0.10/kWh. However, its high power draw of 11,020 W makes it more suitable for professional mining farms with large-scale data center infrastructure.

Industry analysis suggests that for highly efficient mining farms with electricity below $0.05/kWh, cash mining costs can be controlled between $34,000–$43,000 per BTC, with gross margins reaching 37% to 57%. However, these profit levels are calculated before depreciation and total cost accounting.

Industry Impact Analysis: From Hashrate Scale to Capital Efficiency

Structural Implications of Hashrate Exit

After peaking at about 1,160 EH/s in October 2025, Bitcoin’s network hashrate recorded its first quarterly decline in six years in Q1 2026. According to YCharts, network hashrate was about 964 EH/s on May 15, 2026. This hashrate exit has been accompanied by multiple downward difficulty adjustments—several in 2026 alone, with a 7.76% drop in March marking the largest single reduction in a year. On May 1, difficulty fell another 2.3% to 132.47T.

This hashrate exit isn’t a pathological signal for the network; it’s a natural clearing process in a low-profit environment. Estimates suggest that 60%–70% of network hashrate is operating underwater. As these units exit, difficulty adjustments automatically improve marginal profitability for remaining miners.

Shifting Narratives: From Mining to AI Infrastructure

The mining capital market is undergoing a significant narrative shift. Several listed mining companies have announced AI and HPC contracts totaling more than $70 billion. Riot Platforms sold 3,778 Bitcoins in Q1 2026 to cover operating costs. CoinShares forecasts that by the end of 2026, up to 70% of listed miners’ revenue could come from AI/HPC.

From an industry evolution perspective, mining is experiencing an "infrastructural transformation." The core assets of mining farms—power capacity, data center facilities, cooling systems—are shifting from "dedicated assets" for mining to "general-purpose assets" for a range of high-performance computing workloads. However, there are fundamental differences in operating models between AI data centers and mining: mining is more tolerant of power interruptions, while AI inference services demand much higher latency and stability. Moreover, AI data centers require significantly higher upfront capital investment than mining hardware and compete directly with large cloud service providers. Viewing AI transformation as a "complete hedge" rather than a "limited buffer" is an oversimplified narrative.

Popular Opinions That Need Scrutiny

"The real test comes a year after the halving—2026 is the make-or-break year"

Data supports this narrative. The halving occurred in April 2024, and mining profitability deteriorated sharply at the end of 2025 and early 2026. Bitcoin’s price fell more than 31% from about $126,000, combined with reduced block rewards and sustained high hashrate, creating triple compression effects that peaked 18–24 months after the halving. Early 2026 saw hashprice drop to $28–30/PH/day, well below the threshold for sustainable industry operations. This narrative accurately describes the delayed impact of the halving on mining.

"AI transformation can fully offset mining losses—mining is no longer risky"

This claim lacks complete empirical support. While mining companies have announced substantial AI/HPC contracts (over $70 billion), the proportion that translates into stable cash flow remains uncertain. AI data center operations differ fundamentally from mining: mining tolerates power interruptions, but AI inference services require far greater latency and stability. Additionally, AI data centers demand much higher upfront capital expenditures and face direct competition from large cloud providers. Treating AI transformation as a "complete hedge" rather than a "partial buffer" is an oversimplified market narrative.

Conclusion

Two years after the 2024 halving, Bitcoin mining profitability has cycled from expansion to contraction and now to structural adjustment. With Bitcoin trading around $75,773.2 and hashprice at about $33.30/PH/day, miner efficiency has moved from being a "bonus" to a "lifeline"—efficiency below 15–16 J/TH is now the basic threshold for profitability in 2026.

Among the S21, M60, and A15 series, the S21 XP, with its 13.5 J/TH efficiency, stands out as the mainstream air-cooled miner that can remain profitable even at electricity rates up to $0.10/kWh. For most miners, keeping electricity costs below $0.06/kWh and consistently upgrading to new-generation miners with efficiency under 15 J/TH remains the key strategy for surviving this downturn.

The essence of mining has never changed: it’s a capital-intensive industry with shrinking margins, but always with structural opportunities. Every hashrate clearing widens the survival space for efficient operators; every technological iteration redraws the line between winners and losers. Is mining still profitable in 2026? The answer depends on four variables: electricity price, miner efficiency, cash flow management, and belief in Bitcoin’s long-term value.