Hashrate is the number of hash calculations a mining machine, or a whole proof-of-work network, performs each second while trying to find a valid block. Each hash is one guess. On Bitcoin the network's combined guessing speed is now about one zettahash per second (10^21 hashes per second). On 16 October 2026 mining difficulty rose 5.92% to a record of about 140.58 trillion at block 971,712, and mempool.space put the network hashrate at roughly 1,068 EH/s. The figure matters because it shows how much real hardware and energy an attacker would have to beat, and how hard miners are competing for each block.
What one "hash" actually is
To mine a Bitcoin block, a miner builds an 80-byte block header and runs it through SHA-256 twice, written as SHA256(SHA256(header)). The header has six fields:
| Field | Size | What it holds |
|---|---|---|
| Version | 4 bytes | Block version |
| hashPrevBlock | 32 bytes | Hash of the previous block |
| hashMerkleRoot | 32 bytes | Summary of all transactions in the block |
| Time | 4 bytes | Timestamp |
| Bits | 4 bytes | The current target in compact form |
| Nonce | 4 bytes | A 32-bit counter the miner changes |
The result counts only if it is below the network's target. There is no shortcut: the miner changes the nonce, hashes again, and repeats. The nonce is only 32 bits, so fast machines run through every value often. When that happens the miner changes the extraNonce field in the generation transaction. That changes the Merkle root and gives a fresh set of headers to try.
Hashrate is simply how fast that loop runs.
Units: from H/s to ZH/s
Each step up is a factor of 1,000.
| Unit | Hashes per second |
|---|---|
| H/s | 1 |
| kH/s | 10^3 |
| MH/s | 10^6 |
| GH/s | 10^9 |
| TH/s | 10^12 |
| PH/s | 10^15 |
| EH/s | 10^18 |
| ZH/s | 10^21 |
Individual Bitcoin mining machines are rated in TH/s. The Bitcoin network is reported in EH/s, and since 2026 it has sometimes been reported in ZH/s.
Hashrates are only comparable within one algorithm. Bitcoin uses SHA-256. Litecoin uses scrypt. Monero uses RandomX, which its developers designed to run efficiently on ordinary CPUs and make specialised ASIC hardware less useful. A TH/s on one algorithm is not the same amount of work as a TH/s on another. Ethereum has had no hashrate since the Merge on 15 September 2022, when it replaced proof-of-work with proof-of-stake.
Network hashrate is estimated, not counted
Nobody can count every hash on the network. Explorers infer hashrate from two things everyone can see: the current difficulty and how quickly blocks are actually being found.
The Bitcoin wiki gives the rule of thumb: finding a block at difficulty D takes about D × 2^32 hashes on average. With a ten-minute (600-second) target, network hashrate ≈ D × 2^32 / 600.
Using the 16 October 2026 difficulty of about 140.58 trillion:
- 140.58 × 10^12 × 2^32 ≈ 6.04 × 10^23 hashes per block
- divided by 600 seconds ≈ 1.0 × 10^21 H/s, or about 1,000 EH/s
Mempool.space's figure of about 1,068 EH/s is higher because blocks have recently arrived faster than every ten minutes. That is why the next adjustment is expected to raise difficulty again.
Two practical consequences:
- Daily figures are noisy. Block discovery is random, so a few lucky or unlucky hours move the estimate. Trackers that use a seven-day average show the trend more reliably than a single day.
- Sources disagree. In October 2026, published estimates ranged from about 950 EH/s to 1,101 EH/s depending on the window and method. Check which averaging period a chart uses before comparing numbers.
Hashrate and difficulty
Bitcoin keeps its ten-minute rhythm by retargeting difficulty every 2,016 blocks. That is about two weeks if blocks arrive on schedule. The protocol compares how long the last period actually took with the ideal 1,209,600 seconds (two weeks):
- Blocks came too fast (hashrate rose) → difficulty goes up.
- Blocks came too slowly (hashrate fell) → difficulty goes down.
Each retarget is capped. Difficulty can rise by up to 300% or fall by up to 75% in one step. The Bitcoin whitepaper describes the same idea: difficulty follows "a moving average targeting an average number of blocks per hour."
The two numbers are easy to mix up. Hashrate is what miners actually supply. Difficulty is the protocol's response, and it lags hashrate by up to two weeks. The 16 October 2026 adjustment of +5.92% came after the 3 October adjustment of -0.03%, and it was the tenth increase of 2026.
How long would one machine wait for a block solo?
Enter a network hashrate and your machine's hashrate to see the expected number of days before it finds a block on its own.
| Bitcoin network hashrate | Your machine's hashrate | Expected days until your machine finds a block on its own |
|---|---|---|
| 1,068 EH/s | 200 TH/s | 37,083 days |
| 950 EH/s | 200 TH/s | 32,986 days |
| 1,101 EH/s | 200 TH/s | 38,229 days |
- It assumes 144 blocks a day, the ten-minute target, and that your share of blocks equals your share of the network hashrate.
- The result is an average. Block discovery is random, so the actual wait can be much shorter or much longer.
This shows the expected wait for a solo miner only. It does not cover pool payouts, electricity or hardware costs, hashprice, or the effect of future difficulty changes.
Why hashrate means security
Bitcoin's whitepaper puts the security model in one line: the honest chain wins "if a majority of CPU power is controlled by honest nodes." The valid chain is the one with the most proof-of-work behind it. To rewrite recent history, an attacker needs more hashing power than everyone else combined, and must keep it running long enough to overtake the honest chain. This is a 51% attack.
The bigger the network's hashrate, the more hardware and electricity that takes. The danger is real for small chains. Between 31 July and 1 August 2020, an attacker on Ethereum Classic mined more than 3,500 blocks privately in about 12 hours and then published them, reorganising the chain. Blockchain analytics firm Bitquery traced 807,260 ETC, about $5.6 million at the time, as double-spent. According to Bitquery, the hashing power was rented for about 17.5 BTC (around $192,000). Days later another attack on the same chain reorganised more than 4,000 blocks. Ethereum Classic's developers responded by urging services to require more confirmations.
The lesson is that security depends on how much hashrate an attacker could get, rented or bought, compared with the honest network. A chain whose total hashrate can be rented for a few days is exposed. A chain where even renting a fraction of its hashrate is impractical is far harder to attack.
What hashrate means for one miner
A single machine's hashrate decides its share of the network's guesses, and so its expected share of blocks. A worked example at about 1,068 EH/s:
- A 200 TH/s machine has a share of 200 × 10^12 / 1.068 × 10^21 ≈ 0.0000187%.
- At 144 blocks a day, it would expect one block roughly every 37,000 days, or about 100 years.
This is why nearly all small miners join pools. A pool combines participants' hashrate, finds blocks far more often, and pays each member according to the work they contributed. The expected reward is the same before fees, but income arrives steadily instead of possibly never.
Efficiency matters as much as raw speed. Machines are rated in joules per terahash (J/TH). Power draw equals hashrate times efficiency. A 200 TH/s machine at 20 J/TH draws 4,000 W, or 96 kWh a day. When difficulty rises and revenue per terahash falls, the machines that stay above their electricity cost are the ones using the fewest joules per terahash.
Hashprice: the link to miner revenue
Hashprice is the expected daily revenue from 1 TH/s of hashing power. The term comes from the Luxor Hashrate Index. It depends on four things: Bitcoin's price, network difficulty, the block subsidy (3.125 BTC per block since the 2024 halving) and transaction fees. When network hashrate rises faster than price, each terahash earns less. Hashprice is gross revenue. It does not deduct power, hosting, pool fees or hardware costs.
The next halving cuts the subsidy to 1.5625 BTC at block 1,050,000. That is about 78,000 blocks after the October 2026 height, or roughly a year and a half at ten-minute blocks, which points to around 2028. The exact date depends on how fast blocks are found.
When hashrate falls
The clearest real-world shock came in 2021. The Cambridge Centre for Alternative Finance (CCAF) estimated that China's share of Bitcoin mining had already fallen from 75.5% in September 2019 to 46% in April 2021. After the government ordered mining to stop in June 2021, the CCAF wrote that this "effectively led to all of China's hashrate disappearing overnight." Blocks slowed, difficulty fell at the following retargets, and mining became more profitable for the miners still running, which is how the network is meant to respond.
Common reasons for a sustained drop:
- Electricity costs jump in a major mining region.
- Revenue falls: a lower price, a halving or low fees push hashprice below some machines' running costs.
- Regulation forces miners to shut down or relocate.
- Old hardware retires faster than new machines are installed.
A drop also shows up in the next difficulty adjustment. A short dip that the next retarget corrects is normal variance. Several downward adjustments in a row mean miners are switching off for a structural reason.
How to track it
- Network level: a block explorer such as mempool.space shows the current estimate, the difficulty history and a forecast for the next adjustment. Use seven-day or longer averages for trends.
- Miner economics: a hashprice index shows revenue per TH/s per day, which explains why hashrate is rising or falling.
- Your own machine: your pool's dashboard shows the hashrate the pool has credited from your submitted shares. That is the figure that determines payouts, not the rating on the box.
Limitations: network hashrate figures here are estimates as of 16–25 October 2026 and change with every block; regional mining shares come from CCAF pool-based estimates.
