Bitcoin Mining Difficulty Fell 19.9% From Peak: What Capitulation Means for You
Bitcoin mining difficulty has fallen 19.9% from its all-time peak and sat at 127.48T after the 8 August 2026 adjustment. CoinDesk reported on 1 August that difficulty had dropped below year-earlier levels for only the second time in history — roughly 14% below the 2026 high — as plunging revenues forced operators to pivot.
For anyone still online, this is unambiguously good news, and it is worth understanding exactly why. Difficulty is the network's mechanism for keeping blocks roughly ten minutes apart regardless of how much hardware is competing. When machines switch off, the remaining machines each win a larger share of the same 450 BTC per day.
In other words, the miners who capitulated handed their revenue to the miners who did not. This article covers what happened, what your machine earns now, and how to be on the receiving side of the next one. Run your own figures at mining calculators.
The Adjustments That Got Us Here
2026 has produced two of the largest downward adjustments of the modern ASIC era. February's -11.16% and June's -10.09% — the latter the eighth-largest decrease since 2016, per Hashrate Index — bracketed a year in which hashprice set two all-time daily lows. The 8 August adjustment was a mild +0.99%, suggesting the flush had largely worked through.
| Event | Change | Note |
|---|---|---|
| February adjustment | -11.16% | Largest of 2026 |
| 14 June adjustment | -10.09% | 8th-largest since 2016 |
| 1 August | -14% vs 2026 high | Below year-earlier levels |
| 8 August adjustment | +0.99% | To 127.48T — stabilising |
| Peak to now | -19.9% | All-time-high comparison |
Sources: Hashrate Index roundups and CoinDesk, 1 August 2026.
A difficulty drop is the only raise in mining that arrives without a negotiation and without capital expenditure.
What a 19.9% Drop Is Worth to Your Machine
Revenue per terahash moves inversely with difficulty. A 19.9% decrease in difficulty means roughly a 25% increase in BTC mined per terahash relative to the peak, holding Bitcoin's price constant. At the current 920 EH/s network, here is what that looks like across the fleet.
| # | Model | TH/s | J/TH | @ $0.036 | @ $0.0455 | @ $0.08 | @ $0.12 |
|---|---|---|---|---|---|---|---|
| 1 #1 | Antminer S23 Hyd | 580 | 9.5 | $391 | $353 | $214 | $54 |
| 2 | Antminer S21 XP Hyd | 473 | 12.0 | $288 | $249 | $106 | -$60 |
| 3 | Whatsminer M63S Hyd | 390 | 14.4 | $213 | $174 | $33 | -$131 |
| 4 | Antminer S21 XP | 270 | 13.5 | $154 | $129 | $37 | -$70 |
| 5 | Whatsminer M66S Hyd | 298 | 15.0 | $158 | $127 | $14 | -$116 |
| 6 | Antminer S21 Pro | 234 | 15.5 | $121 | $96 | $5 | -$101 |
| 7 | Antminer S21 | 200 | 17.5 | $93 | $69 | -$20 | -$122 |
| 8 | Whatsminer M60S | 186 | 18.5 | $82 | $58 | -$29 | -$129 |
| 9 | Avalon A1466I | 175 | 17.0 | $84 | $63 | -$12 | -$99 |
| 10 | Antminer S19 XP Hyd | 255 | 20.8 | $96 | $60 | -$74 | -$229 |
Basis: BTC ~$63,000, network ~920 EH/s, 450 BTC/day, 730 h/month. $0.036/kWh is the OneMiners Nigeria rate; $0.0455/kWh is the USA rate; $0.08 and $0.12 are typical industrial and residential rates. Re-run your own numbers at asicprofit.com.
Why Some Miners Capitulated and Others Did Not
It was not luck and it was not conviction. It was two numbers. CoinShares estimated in March 2026 that 15-20% of the global fleet was running at a loss, and noted that S19-generation machines cannot generate positive cash flow at current prices unless electricity is below roughly five cents per kilowatt-hour.
Every operator that switched off in 2026 was on the wrong side of either efficiency or rate, usually both. Every operator that stayed online had at least one of them locked in.
How to be on the surviving side next time
- Buy J/TH, not TH/s. Efficiency is what determines your break-even rate.
- Fix your electricity rate contractually, so a hashprice dip does not also become a power-price problem.
- Protect uptime — a downed machine misses the exact window when difficulty is low and revenue per terahash is highest.
- Keep underclocking as a tool, not a last resort.
The Capacity That Left Is Not Coming Straight Back
Much of it did not go to a competitor. It went to AI and HPC workloads. Forecasts cited by CoinShares put AI at up to 70% of listed-miner revenue by the end of 2026 for firms executing on signed contracts, up from about 30% in late 2025. Megawatts that move to a multi-year inference contract do not swing back on the next hashprice tick.
That is structurally different from previous capitulations and it means the current low-difficulty window may last longer than the post-halving flushes did. The way to hold it is a fixed rate. OneMiners runs 2,163 MW across 20 hosting sites at an average $0.0480/kWh, fixed for seven years, with per-miner uptime reporting and a 95% uptime compensation floor — sites at hosting centers, mechanics at how hosting works.
FAQ: Difficulty and Capitulation
1. What is mining difficulty in simple terms?
A dial the network turns every 2,016 blocks to keep block times near ten minutes. More competing hardware turns it up; less turns it down. Full explanation at btcfq.com.
2. Does falling difficulty mean Bitcoin is failing?
No. It means the adjustment mechanism is doing its job. Difficulty has fallen after every major revenue shock in Bitcoin's history and recovered every time.
3. How much more do I earn after a 19.9% drop?
Roughly 25% more BTC per terahash than at the peak, all else equal. The dollar value depends on Bitcoin's price, which fell over the same window.
4. Is now a good time to switch machines on?
Low difficulty plus cheap hardware is the most favourable entry the cycle offers. The condition is a power rate below your machine's break-even, which for most current hardware means under $0.08/kWh.
5. When will difficulty rise again?
When hashprice recovers enough to make idle machines cash-positive. The +0.99% adjustment on 8 August suggests the bottom of the flush.
6. What is a difficulty ribbon compression?
A technical pattern traders watch when short and long difficulty averages converge after capitulation. It is a market-analysis tool, not a mining decision input — your break-even rate is.
7. Should I sell my S19s?
If you cannot get them under about five cents a kilowatt-hour, they will not clear cash flow. Hosting them at hosting centers rates or trading up to efficient hardware in ASIC miners are the two real options.
8. Where do I watch difficulty live?
Hashrate Index and Mempool.space both publish it, along with the projected next adjustment. Both are named in the sources below.
Capitulation is a transfer, not a loss. OneMiners makes sure you are on the receiving end.
Sources
Every figure above is traceable to a named publication. Verify before you commit capital.
- Hashrate Index — Roundup, 17 August 2026 (920 EH/s 7-day SMA; BTC ~$63,000 on 16 August; difficulty 127.48T on 8 August).
- Hashrate Index / Luxor Technologies — Hashrate Lookback, June 2026 (record daily hashprice low of $27.74/PH/s/day on 6 June 2026; record monthly average of $30.37).
- CoinDesk — "Bitcoin mining difficulty shrinks 14% from this year's high as plunging revenues force operators to pivot", 1 August 2026.
- CoinShares — mining industry update, March 2026 (15-20% of the global fleet running at a loss).
- Rising Trends — "Top Crypto Trends in 2026 (What the Search Data Shows)", 13 July 2026 (ASIC 1,000,000 searches/mo +22%; IREN +1,933% YoY; Cipher Mining +513%; TeraWulf +175%).
- crypto.news — "Bitcoin mining capitulation: difficulty falls 19.9% as miners pivot to AI", August 2026.
- Mempool.space — live difficulty and projected adjustment data.
