A rare large-scale migration in history: 50% of Bitcoin miners are "offline". What are they doing based on the data?

Author | Glassnode

Editor | Colin Wu

Over the past week, a mining exodus has been underway, as an estimated 50% of Bitcoin miners have gone offline and migrated. We have assessed the extent of this exodus based on observed on-chain data:

Bitcoin volume has risen slightly this week, although it remains within the trading range established since mid-May. The price of Bitcoin has reached a high of $36,460 and a low of $32,775 this week, with volatility decreasing during the sideways price movement.

The volatility of Bitcoin prices has decreased, but the volatility of the mining industry has increased significantly. During the period of large-scale mining machine migration, the computing power of the Bitcoin network has dropped sharply. This week, the Bitcoin protocol experienced the largest difficulty adjustment in history, with the difficulty falling by 27.94%.

The Bitcoin protocol is remarkably resilient. New blocks continued to be mined this week, and transactions were settled without interruption. At the same time, approximately 50% of the mining base was shifting, and 50% of capital was reallocated to other regions. This week we will focus on a few on-chain mining metrics that help describe and measure the scale of global hashrate migration, an ongoing, incredible event.

Mining migration is expanding

This week, volatility in mining metrics has increased significantly as the impact of the Chinese mining ban has gradually unfolded. When a significant portion of the hashrate is offline, blocks are mined more slowly until the difficulty is adjusted downward. Bitcoin's mechanism combines the complexity of mining difficulty with the average hashrate observed over a 2016 block difficulty window (about 14 days).

This week, the average block time calculated on a 24-hour average soared to a short-term high of 1958 seconds (32.6 minutes), which is more than 226% longer than the average block time of 600 seconds. Fortunately, the super-long block time on June 28 was only temporary, and the average block time has since returned to an average of 800 to 900 seconds.

To estimate the size of the ongoing hashrate shift, we considered the change in hashrate since the peak observed in April-early May 2021. Note that hashrate is estimated as the derivative of block time and difficulty, and we use a 24-hour moving average here to smooth out some of the natural variation.

The Bitcoin network’s hashrate typically peaked at around 180 EH/s before dropping to a local low of 65 EH/s (coinciding with the 1,958 second average block time on June 28).

Since then, the hashrate has gradually recovered and stabilized in the range of 88 to 110 EH/s, reflecting an overall decrease in hashrate of 38% to 49%. This reflects the proportion of Bitcoin hashrate that is currently offline and affected by the Chinese ban.

Now that the protocol difficulty has adjusted down, we can see that the difficulty ribbon has inverted to its deepest point since the bear market capitulation in 2018. An inverted difficulty ribbon is an extremely uncommon event where a fast moving average of difficulty (9-day, 14-day, etc.) falls below a longer-term moving average (128-day, 200-day).

Difficulty ribbon reversals often represent miner capitulation events, often observed at the end of a bear market, or following a halving event where miner revenues are squeezed and profitability is hit. This is a result of miners shutting down machines that cost more than the benefit, and has historically been associated with strong bullish market reversals.

Miners’ income soars

Although historically, difficulty ribbon reversals are bullish events. In the current situation, the huge logistics costs incurred by Chinese miners may require them to sell their previously accumulated Bitcoin, thus generating selling pressure. However, it should be noted that the remaining approximately 50% of miners who are still operating have seen their income rise, which is somewhat of a balance.

When Bitcoin was trading in the $50,000 to $60,000 range in April, hashrate was at its peak, with total mining revenues of $50 million to $60 million per day. While prices have since fallen by about 50%, miners that are still operating have seen 38% to 49% of their competitors shut down operations in the short term.

Total mining revenue is now about $25 million to $30 million per day, but there are fewer miners sharing the revenue. The number of newly issued Bitcoins per day remains the same, but there are fewer competitors to share the spoils. This means that after the latest difficulty adjustment, the operating expenses of miners who are still in operation remain the same, but their profit margins have increased by nearly 2 times, close to the profit margin level in April this year.

The miner outflow multiple tracks miner selling relative to its annual average. From this we can see that miner selling has slowed down significantly recently, even in the midst of the mining machine migration. Historically, we have seen miner selling behavior follow the following cyclical pattern:

(Wu Blockchain reminds that overseas data websites often have errors in the data of miners' currency holding addresses, and the reference value is not high)

• Hoarding in a bear market (green), miners reach a stable selling pattern (outflow multiples are flat).

• Selling increases during a bull market (red), and miners take advantage of the bull market to accelerate the selling of Bitcoin (the outflow multiple increases).

• Selling decreases after the market peaks (blue) as miners slow their selling, perhaps due to growing conviction, or a desire to limit excessive selling pressure (outflow multiples fall).

Impressively, even during the mining machine migration, we have not seen a significant increase in miner selling behavior. The speed of hash rate recovery may provide us with more clues:

• The rapid hash rate recovery suggests that Chinese miners are successfully relocating, liquidating hardware equipment, or otherwise recovering costs, reducing the risk of selling miners' reserves.

• A slow and prolonged hash rate recovery suggests the opposite, increasing the chances that miners will incur costs and accumulate debt, thereby increasing the likelihood that miners will sell their Bitcoin.

It doesn't seem to be selling heavily.

In fact, the macro selling behavior of miners has changed dramatically since the beginning of 2020. For almost all of Bitcoin’s trading history, miners have consistently sold more Bitcoin than they accumulated, leaving the unsold supply in a structural downward trend.

The chart below shows the relationship between unsold miner supply (i.e., block transactions that have never moved) and the 365-day moving average. The structural downward trend in miner selling that began in mid-2020 appears to have not only flattened, but reversed. Unsold Bitcoin from block production is now above its annual average.

This suggests that miners have begun accumulating at a rate that the market has never seen before. Considering the timing of this event, there are likely a number of factors behind this:

• The global macro-monetary landscape supporting Bitcoin has become more pronounced in 2020, increasing miners’ conviction to hold onto the currency.

• Miners have access to a plethora of financing options, such as Bitcoin-collateralized debt, and liquid options and futures markets to hedge risk.

• As global chip manufacturing capacity constraints reduce the production of ASIC miners, as prices rise, existing hashrate becomes increasingly profitable because few new competitors can enter the market.

Bitcoin mining was likely very profitable during this halving period (and remains so) as prices rebounded in the bull market.