Hash rate plummets by 145 EH/s: What does it mean that BTC is experiencing its first-ever "hash rate bear market"?

Bitcoin hash rate has contracted by approximately 14% since May 28, 2026. The total network hash rate dropped from about 1,030 EH/s at that time to 885 EH/s, representing a net outflow of 145 EH/s. This is the largest hash rate outflow event on the Bitcoin network since 2020.

What makes this event special is not the contraction itself, but its combined form: the hash rate decline occurred after the halving cycle, and was superimposed with a decline in Hashprice and changes in miners’ balance sheets. Elektron Energy CEO Rapha Zagury defines this as Bitcoin’s first “hash rate bear market”—a sustained contraction driven by market forces rather than structural weaknesses in network performance.

Is the 145 EH/s outflow due to passive shutdowns or active choices?

When hash rate drops significantly, the market usually attributes it to passive shutdowns caused by deteriorating profitability. The current outflow of 145 EH/s has a dual nature: passive shutdowns and active reallocation.

Passive shutdowns are driven by profitability. As of June 7, 2026, Hashprice—the daily mining revenue per 1 PH/s of hash rate—fell to about $28.26 / PH/s, a 26.96% decrease over the past 30 days. On-chain transaction fees account for less than 1% of miners’ rewards, block production times have deviated from the protocol’s 10-minute target, averaging around 11 minutes and 12 seconds. When unit revenue continues to weaken and fees cannot compensate, high-cost miners rationally choose to shut down equipment.

However, active reallocation is equally significant. According to Hashrate Index data, Bitcoin network hash rate had been nearly continuously growing for six years prior to this reversal. The current 145 EH/s decline is the first such sustained contraction in six years. Several publicly listed mining companies have explicitly reduced their mining operations in Q1 2026, reallocating electricity and infrastructure to AI and high-performance computing sectors. This indicates that part of the hash rate reduction is not due to losses forcing passive exit, but rather active transfer to more economically attractive uses.

Luke Gromen, founder and CEO of Forest For The Trees, summarizes this directly: “AI is sucking all the oxygen out of the room, making Bitcoin the victim.”

How much has Hashprice bottomed out, squeezing miners’ profit margins?

Hashprice is a core indicator of miners’ economic pressure. Since peaking at around $63 / PH / s / day in July 2025, it has been declining, further dropping to the $28–30 / PH / s / day range in Q1 2026. As of June 7, 2026, the level of approximately $28.26 / PH / s has fallen well below the threshold needed for most mining hardware to achieve positive cash flow.

CoinShares’ Q1 2026 mining report estimates that at a Hashprice of about $30 / PH / s / day, miners operating less efficient hardware than S19 XP and with electricity costs of 6 cents / kWh or higher are unprofitable. The report estimates that under these conditions, roughly 15–20% of global mining capacity is operating at a loss.

It’s also important to note changes in cost structure. In Q4 2025, the weighted average cash cost for listed miners to produce one Bitcoin rose to about $79,995. During the same period, Bitcoin’s price was in a retracement from a high of $124,500, creating a divergence that is directly reflected on miners’ balance sheets: listed miners have reduced their Bitcoin holdings by over 15k BTC from peak levels.

Additionally, miners’ financing structures are changing. IREN holds about $3.7 billion in convertible notes, TeraWulf’s total debt is around $5.7 billion, and Cipher Digital issued $1.7 billion in senior secured notes. Leverage amplifies the pressure during profit declines, accelerating the exit decisions of highly indebted miners.

Can an approximately 11% difficulty adjustment reverse the squeeze on miner profits?

Bitcoin’s difficulty adjustment mechanism is an automatic network balancing system: when block times persistently exceed the 10-minute target due to hash rate reductions, difficulty will automatically adjust downward in the next cycle. As of June 8, 2026, the network’s average block time was about 11.2 minutes, significantly deviating from the target. Based on current block production speed, the next difficulty adjustment is expected around mid-June 2026, reducing difficulty from the current 138.96 T to approximately 123.88 T—about an 11% decrease.

A difficulty reduction of this magnitude means that for continuing miners, the units of hash rate produce more rewards: the same hash rate will earn a larger share of block rewards after difficulty drops. For miners still active on the network, this provides a temporary profit recovery. However, two key constraints should be noted:

First, although an 11% difficulty drop is substantial, the hash rate has already declined by 145 EH/s, and in October 2025, the network’s peak hash rate approached 1,160 EH/s. The current 885 EH/s is roughly 25% below that peak. The difficulty adjustment cannot fully compensate for such a large hash rate exit in terms of revenue base contraction.

Second, the magnitude of profit recovery is fundamentally limited by Bitcoin’s price. At the current Hashprice level of about $28 / PH / s / day, an 11% difficulty reduction could raise unit revenue to about $31 / PH / s / day. While this alleviates some profit pressure, it remains far below the $63 / PH / s / day level seen in July 2025. Historically, when Hashprice and network difficulty deviate from their long-term averages simultaneously, a single difficulty adjustment alone cannot fundamentally change miners’ economic structure.

Why is the acceleration of AI data center transformation contributing to structural hash rate loss?

The shift of listed miners toward AI data centers has been a prominent trend from 2025 to 2026. The economic logic is that the core assets miners possess—cheap electricity access, substation infrastructure, industrial parks, and long-term power purchase agreements—are also scarce resources needed for AI data center expansion.

Industry data shows that by early 2026, listed miners had announced over $70 billion in AI and high-performance computing contracts. Hut 8 signed a 15-year, $9.8 billion lease agreement to build a 352 MW data center in Texas. TeraWulf secured $12.8 billion in AI contract revenue. Core Scientific raised $208 million in Q1 2026 by selling Bitcoin to accelerate its transition into AI and HPC infrastructure; it also launched a $3.3 billion debt financing plan for a full shift to AI data centers. IREN completed a $3 billion convertible note issuance in May 2026 to transition from pure Bitcoin mining to AI data center operations.

This transformation has a structural impact on hash rate: AI data center contracts typically involve long-term leases of 10–15 years, with tenants demanding stable power supply and service level agreements. This is fundamentally different from Bitcoin mining, which is interruptible and flexible. Once power capacity is locked into AI hosting, it becomes economically unreasonable to reallocate that capacity back to mining—even if Bitcoin prices rise again. As a result, a significant portion of the current 145 EH/s may be permanently detached from the Bitcoin network, rather than participating in a “shutdown—difficulty adjustment—restart” cycle.

It’s also worth noting that AI transition carries risks. The construction costs of AI data centers are much higher—around $8–15 million per MW—compared to $0.7–1 million per MW for mining infrastructure. Relying on debt financing means that the repayment risk persists; the shift does not eliminate risk but merely transfers it from mining to hosting. Profit margins in AI hosting are also influenced by supply-demand dynamics and regulatory factors.

Does the “hash rate bear market” mark the end of the assumption that hash rate only goes up?

The long-standing empirical assumption in Bitcoin mining has been that hash rate only increases over time. This is based on two main reasons: long-term bullish expectations for Bitcoin’s price driving more hash rate into the network, and the belief that each four-year halving cycle will push hash rate to new highs.

The outflow of 145 EH/s challenges this assumption. Zagury notes that current hash rate is about 25% below the September 2025 high. He also believes Bitcoin’s security remains robust because the capital cost to execute a 51% attack is still very high. However, the more significant long-term risk lies in fee market stagnation—transaction fees currently account for less than 1% of miners’ rewards, and block subsidies halve every four years.

From a macro perspective, whether the “hash rate bear market” ends the “hash rate only goes up” hypothesis depends on two long-term variables: whether AI demand continues to absorb miners’ existing electricity and hash rate supply; and whether, as Bitcoin’s market cap growth slows, the long-term trend of unit hash rate revenue shifts from growth to stability or decline. These questions cannot be answered definitively with current market data, but the 14% contraction already forces the market to reassess the validity of this assumption.

Notably, the stock performance of listed miners has diverged from Bitcoin’s price trend. According to 10X Research, a basket of mining stocks gained about 56% in the first five months of 2026, while Bitcoin itself declined roughly 17%. The capital markets are re-pricing miners’ valuation logic—no longer solely as a beta proxy for Bitcoin’s upside, but as “electricity and infrastructure operators.” This shift in valuation framework itself offers a new perspective challenging the “hash rate only goes up” assumption.

What are the implications of the 145 EH/s contraction for Bitcoin’s security and market cap structure?

The core safeguard of Bitcoin’s security is not the absolute hash rate number but the marginal cost of attacking the network. The decline from 1,030 EH/s to 885 EH/s means the total amount of hash equipment needed to execute a 51% attack has decreased, but the absolute cost threshold remains very high. From a security margin perspective, the current level does not pose systemic risk. Zagury also believes that executing a 51% attack remains economically infeasible due to the high capital costs.

However, ongoing hash rate outflows are changing two key dimensions of Bitcoin’s market cap structure:

First, supply-side pressure. The 15k BTC reduced by listed miners has been absorbed or is being absorbed by the market. Compared to total Bitcoin supply, this volume alone is insufficient to drive price trends, but its marginal impact on market psychology and liquidity should not be ignored—especially during price pressure periods, where forced or active selling by miners could reinforce downward price dynamics.

Second, based on Gate行情 data, as of June 11, 2026, Bitcoin’s price has fallen back into a lower range since February. The hash rate contraction in this price environment is both a result of price (miners exiting due to losses) and a potential amplifier (less hash rate → difficulty adjustment → lower unit costs → some profit relief). Yet, the easing effect is limited—at the current Hashprice of about $28 / PH / s / day, even with an 11% difficulty adjustment, miners’ unit revenue remains well below the mid-2025 average. If Hashprice cannot recover to sustainable profitability levels, further hash rate contraction cannot be ruled out.

From a network health perspective, the difficulty adjustment mechanism provides an automatic brake on hash rate decline, and three consecutive negative difficulty adjustments occurred from late 2025 to early 2026. The current 11% expected reduction is a significant protocol-level cost reset. But the long-term issue remains on the fee side: if transaction fees cannot increase substantially in the coming years to offset the ongoing decline in block subsidies, Bitcoin’s security budget will rely more on price cycles and tiered miner survival rather than network effects.

Summary

The 145 EH/s hash rate contraction from late May to early June 2026 is driven by three intertwined factors: Hashprice falling to around $28 / PH / s / day, intensifying profitability losses for high-cost miners; the approximately 11% expected difficulty adjustment providing a temporary cost buffer for remaining miners; and the long-term locking of electricity and infrastructure into AI data centers, which is altering the structural elasticity of hash rate. Zagury’s definition of a “hash rate bear market” is more about correcting the inertia of “hash rate only goes up,” rather than denying network security. The real long-term risk lies in whether transaction fee markets can gradually shoulder the security budget as block subsidies halve continuously. The accelerated shift of miners into AI businesses is reshaping this structural narrative, with potential long-term impacts far beyond the surface-level digital contraction of hash rate.

FAQ

Q: What is a “hash rate bear market”?

A: Coined by Elektron Energy CEO Rapha Zagury, it refers to a sustained hash rate contraction driven by market forces—not network mechanism issues. In the current case, Bitcoin’s hash rate has fallen about 25% from its September 2025 high, marking a structural interruption in the long-term upward trend.

Q: Does the hash rate decline imply a security risk for Bitcoin?

A: The decline reduces the total hardware needed for a 51% attack, but the absolute cost remains very high. At the current 885 EH/s level, it’s still not enough to pose systemic security risks. Zagury also believes executing a 51% attack remains economically infeasible given current costs.

Q: How much can the difficulty adjustment restore miner profitability?

A: The roughly 11% difficulty reduction expected around mid-June 2026 could increase unit hash rate revenue by a similar margin, assuming Hashprice remains unchanged. But given that Hashprice has already fallen about 55% from July 2025, the buffer from a single difficulty adjustment is limited and unlikely to fundamentally reverse the profit squeeze.

Q: Is the shift of miners into AI data centers reversible?

A: Long-term lease contracts (10–15 years) and strict power stability requirements make it economically unreasonable to reallocate capacity back to mining once committed to AI hosting. A significant portion of the current hash rate contraction is likely structural and irreversible.

Q: Are there new risks associated with miners’ transition to AI?

A: Yes. AI data centers cost much more to build—around $8–15 million per MW—compared to $0.7–1 million per MW for mining infrastructure. Relying on debt financing means the repayment risk persists; the shift does not eliminate risk but transfers it. Profit margins in AI hosting are also affected by supply-demand and regulatory factors.

Q: How to monitor the ongoing development after the hash rate bear market?

A: Focus on three indicators: whether Hashprice recovers above $35 / PH / s / day; whether consecutive difficulty adjustments show signs of hash rate stabilization; and changes in miners’ revenue composition from quarterly reports—these collectively reveal whether the contraction is a voluntary choice or forced exit.