Bitcoin's hashrate posted its first quarterly decline since Q1 2020, dropping approximately 4% year-to-date to around 1 ZH/s. The next difficulty adjustment on April 18 is projected at -14.27%, the largest single downward adjustment since 2021. Average block times have stretched to 11 minutes 39 ...
"Bitcoin miners are becoming AI companies and selling their BTC to fund the transition." — CoinDesk Markets Desk, March 27, 2026
Bitcoin's hashrate posted its first quarterly decline since Q1 2020, dropping approximately 4% year-to-date to around 1 ZH/s. The next difficulty adjustment on April 18 is projected at -14.27%, the largest single downward adjustment since 2021. Average block times have stretched to 11 minutes 39 seconds, well above the 10-minute target.
The cause is economic, not technical. With production costs at roughly $80,000-$90,000 per BTC and market price near $67,000, publicly listed miners are losing an estimated $19,000 per coin mined. The response has been a sector-wide pivot: over $70 billion in cumulative AI and high-performance computing contracts have been announced across the public mining sector. Core Scientific, the largest U.S. miner by capacity, is liquidating substantially all of its 2,537 BTC holdings to fund AI colocation. By end of 2026, listed miners could derive 70% of revenue from AI, up from 30% today.
The question is whether Bitcoin's security model can absorb a structural reallocation of mining capital without long-term degradation.
Bitcoin's 7-day average hashrate sits at approximately 871 EH/s as of April 12, 2026, according to CoinWarz data. This represents a significant retreat from the network's peak above 1,050 EH/s recorded in late 2025.
The decline triggered a cascade of difficulty adjustments:
| Adjustment | Change | Notes | |---|---|---| | January 2026 | -1.4% | First decline of 2026 | | March 2026 | -7.76% | Second-largest drop of the year | | April 18, 2026 (est.) | -14.27% | Projected; would be largest since July 2021 |
Average block time has climbed to 11 minutes 39 seconds, nearly 17% above the protocol's 10-minute target. According to CoinDesk, hashrate fell approximately 12% from its peak in what CryptoQuant described as the worst drawdown since China's mining ban in mid-2021.
The five-year trend provides context. Over the past half-decade, hashrate surged from roughly 100 EH/s to above 1 ZH/s — a 10x increase. Each prior year saw first-quarter growth exceeding 10%. Q1 2026 broke that streak for the first time since 2020.
The April 2024 halving cut block rewards from 6.25 BTC to 3.125 BTC. The protocol still produces approximately 450 BTC per day across 144 blocks, but each block now pays half what it did 24 months ago.
The cost side has not cooperated. Key metrics:
Hashprice — the standard metric measuring miner revenue per petahash per second per day — has fallen to approximately $30-$35/PH/s/day, down roughly 35% from 2025 levels near $55/PH/s/day. Only operators with sub-$0.08/kWh electricity and latest-generation ASIC hardware (sub-20 J/TH efficiency) maintain positive margins.
According to industry estimates, 15-20% of global mining rigs are now operating at a loss. Approximately 252 EH/s of computing power is offline, with much of that capacity believed to be retired on a long-term basis rather than temporarily idled.
Mining one Bitcoin now requires approximately 854,400 kWh of electricity at current network difficulty, according to BestBrokers research. The total daily electricity bill for the Bitcoin mining sector reached $20-$23 million in late 2025.
Bitcoin's price decline from $126,000 in October 2025 to approximately $65,000-$67,000 by February-April 2026 — a drop exceeding 45% — compressed margins across the entire sector.
The sector's response has been unambiguous: convert mining infrastructure into AI compute capacity. The scale of announced contracts:
| Company | AI/HPC Contract Value | AI Revenue Share (Current) | Pipeline Capacity | |---|---|---|---| | Core Scientific | $10.2B (CoreWeave deal, 12 yrs) | 39% of total revenue | 1.5 GW | | TeraWulf | $12.8B contracted HPC revenue | 27% | Not disclosed | | Hut 8 | $7B (15-year lease, River Bend) | Early stage | Not disclosed | | IREN | $3.6B GPU financing (Microsoft) | 9% | 200 MW liquid-cooled |
Revenue projections reflect the pace of transition. IREN's annual revenue is projected to jump 103% to $1.04 billion, then 175% the following year to $2.5 billion, according to analyst estimates compiled by Bankless Times. Hut 8's 2026 revenue forecast is $426 million, up 76% from an estimated $242 million in 2025.
AI-pivoted mining firms now trade at nearly double the valuation per megawatt of power capacity compared to their Bitcoin-heavy peers, according to January 2026 data. The market is pricing AI compute infrastructure at a significant premium over Bitcoin mining capacity.
AI hosting contracts offer 80-90% margins with long-term stable income, compared to the volatile and currently negative margins on Bitcoin mining. The economic logic is straightforward: same power infrastructure, dramatically different revenue profile.
The capital for this transition is coming from two sources: Bitcoin treasury liquidation and debt issuance.
Bitcoin Sales:
Debt Issuance:
These are infrastructure-scale debt loads. The implicit bet: AI revenue will materialize fast enough to service obligations that would be unsustainable on mining economics alone. If AI demand slows or contracts are renegotiated, the leverage becomes a significant risk factor.
According to CoinDesk reporting from March 27, miners are "funding pivot to AI with debt while selling BTC to stay liquid." The sector has moved from accumulating Bitcoin as a treasury asset to treating it as a liquidation source.
A secondary factor compounding the hashrate decline: Iran's mining sector has effectively collapsed. According to data reported by multiple outlets on April 9, Iran's hashrate fell 77% in Q1 2026, from approximately 9 EH/s to roughly 2 EH/s — a loss of 7 EH/s.
The proximate cause is geopolitical. U.S. and Israeli strikes on Iran in February 2026 led to infrastructure damage, energy disruptions, and heightened operational risk for mining facilities. The hashrate loss represents approximately 0.7% of the global network but signals broader risks for mining operations in conflict zones.
Neighboring UAE and Oman mining operations showed no impact, suggesting the disruption is Iran-specific rather than regional contagion. Bitcoin's protocol adjusted difficulty downward to absorb the loss without sustained disruption to block production.
The structural question is whether Bitcoin's security model functions when mining becomes a secondary business for major operators.
The self-correcting difficulty adjustment mechanism is designed to handle hashrate fluctuations. When miners leave, difficulty drops, margins improve for remaining participants, and equilibrium is restored. This has functioned for 15 years.
However, the current situation has a novel element. Previous hashrate declines — China's 2021 ban, prior bear markets — resulted from miners being forced offline by external factors. The 2026 decline is driven by miners voluntarily reallocating capital to higher-return activities. If AI contracts continue to offer 80-90% margins versus negative margins on Bitcoin mining, the economic incentive to re-invest in hashrate is structurally weaker.
CoinShares forecasts hashrate recovery to approximately 1.8 ZH/s by end of 2026, conditional on Bitcoin recovering toward $100,000. If that price recovery fails to materialize, the incentive to rebuild hashrate remains weak.
A lower difficulty level, while self-correcting, reduces the cost of mounting a 51% attack during the adjustment period. With block times running 17% above target, the network is operating in a degraded state — functional, but slower than designed.
The longer-term risk is not a single attack but a gradual decline in network security budget relative to the value being secured. If miners allocate marginal capital to AI rather than Bitcoin, the hashrate-to-price ratio may settle at a structurally lower level than historical norms.
Bitcoin mining is undergoing a structural transformation that goes beyond a cyclical downturn. The April 2024 halving compressed revenues. The subsequent price decline from $126,000 to $67,000 made mining unprofitable for most operators. And the availability of AI compute contracts offering 80-90% margins created a pull factor that simple mining economics cannot match.
The result is a sector where the largest operators are explicitly transitioning away from their core business. Core Scientific is not hedging — it is liquidating its entire Bitcoin treasury. TeraWulf and IREN are taking on billions in debt against AI revenue projections, not mining cash flows.
Bitcoin's difficulty adjustment mechanism will likely restore block times to near-normal after the April 18 adjustment. The protocol functions as designed. But the underlying economic question remains: in a world where the same power infrastructure generates substantially higher returns running AI workloads, what level of hashrate does Bitcoin's security model converge to? The answer depends on BTC price recovery. Without it, the network operates with a thinner security margin than at any point since the 2021 China ban — not because of a policy shock, but because the math points miners elsewhere.