Bitcoin Mining's Great Migration: $70B in Miner Contracts Push Hashprice to 5-Year Lows (2026)

How did the Hormuz blockade trigger a hashprice collapse?

The Strait of Hormuz carries approximately 20% of global oil supply. When naval operations closed the strait in March 2026, the International Energy Agency (IEA) classified it as the largest oil market disruption in decades. Goldman Sachs revised its oil price forecast to $147/barrel. The cascading effect on electricity markets was immediate and severe.

For Bitcoin miners, electricity represents 60–80% of operating costs. As natural gas and grid electricity prices spiked across North America, hashprice — the revenue a miner earns per unit of hash rate — fell to $30/PH/s, its lowest level in five years. This metric captures the full squeeze: revenue per terahash declining while cost per kilowatt-hour surges simultaneously.

The hardware survival hierarchy

Not all mining hardware is created equal. Efficiency, measured in joules per terahash (J/TH), determines which machines survive an energy price shock. The Hormuz crisis created a clear survival hierarchy across ASIC generations.

Table: Production cost per BTC by ASIC generation at post-Hormuz energy prices. S19 operators mining at a loss against $65–70K BTC.

The Antminer S19, which dominated hash rate through the 2021–2024 cycle, became economically unviable almost overnight. At $79,995 per BTC in production costs versus a market price of $65–70K, every block mined by an S19 operator represented a net loss. Bitmain’s newer S21 XP and S23 models, both rated below 10 J/TH, are the only machines expanding into new capacity at current energy prices.

The efficiency gap between generations is not incremental — it is existential. An S23 operating at sub-10 J/TH consumes roughly 40% less electricity per terahash than an S19 at 16 J/TH. At post-Hormuz energy prices, that difference translates directly into the gap between profit and bankruptcy.

The Foundry disconnection event

The energy crisis compounded with Winter Storm Fern in February 2026, which forced grid operators across Texas and the US Midwest to curtail industrial loads. Foundry USA, the largest Bitcoin mining pool by hash rate, reported that 200 EH/s of capacity — approximately 60% of its total pool — disconnected during the storm. This single event represented the largest coordinated hash rate drop from a single pool in Bitcoin’s history.

The disconnection was not permanent. Most operators reconnected within 48–72 hours as grid conditions stabilized. But the event exposed a structural vulnerability: the concentration of hash rate in energy-price-sensitive regions. When electricity costs spike, geographically concentrated mining pools experience correlated failures — a systemic risk that the network’s geographic migration is now addressing.

Why are Bitcoin miners pivoting to AI and HPC infrastructure?

The hashprice collapse accelerated a structural transition already underway: the conversion of Bitcoin mining facilities into AI and high-performance computing (HPC) data centers. The economics are straightforward. A megawatt of power allocated to AI GPU hosting generates higher and more predictable revenue than a megawatt allocated to Bitcoin mining at $30/PH/s hashprice levels.

Between 2025 and 2026, publicly traded mining companies announced over $70 billion in AI/HPC contracts. This is not speculative diversification. These are multi-year, contracted revenue streams backed by hyperscaler demand for GPU compute infrastructure.

Table: Major Bitcoin miner AI/HPC contracts announced 2025–2026. Core Scientific’s CoreWeave deal is the largest single infrastructure contract in Bitcoin mining history.

The Core Scientific model

Core Scientific’s $10.2 billion, 12-year contract with CoreWeave represents the template for the industry. The deal converts 590 MW of existing mining infrastructure into GPU hosting capacity for AI workloads. The contract structure provides predictable, long-term cash flows — a stark contrast to the volatility of hashprice-dependent mining revenue.

CoreWeave, backed by Nvidia and valued at over $35 billion, needs physical infrastructure faster than traditional data center developers can build it. Bitcoin miners already have three critical assets: grid interconnection agreements, power purchase contracts, and cooling infrastructure. Converting a mining facility to GPU hosting requires hardware swaps and upgraded networking, but the foundational infrastructure — power, cooling, physical security — is already in place.

The 12-year contract duration is significant. It means Core Scientific’s AI revenue is locked in across multiple Bitcoin halving cycles, energy price fluctuations, and hashprice downturns. This de-risks the company’s balance sheet in a way that pure Bitcoin mining never could.

IREN: When GPUs overtake ASICs

IREN Ltd (formerly Iris Energy) exemplifies the revenue crossover. With 10,900+ Nvidia GPUs deployed, the company projects that 71% of its 2026 revenue will come from AI/HPC services rather than Bitcoin mining. This represents a fundamental shift in identity: IREN is no longer a mining company that does some AI work. It is an AI infrastructure company that still mines Bitcoin on the side.

The margin differential is significant. While Bitcoin mining margins compress during hashprice downturns, AI compute contracts typically lock in fixed pricing for 1–5 years. This makes miner revenue more predictable and less correlated with cryptocurrency market cycles — precisely the characteristics institutional investors seek when evaluating infrastructure stocks.

TeraWulf: Gigawatt-scale ambition

TeraWulf’s Lake Mariner facility in upstate New York illustrates the scale of ambition. The $1.28 billion expansion targets 2.9 GW of total capacity — enough to power a small city. Located near cheap hydroelectric and nuclear power sources, Lake Mariner offers sub-$0.04/kWh electricity costs, making it competitive for both Bitcoin mining and AI workloads regardless of global energy price volatility.

The dual-use model is key. TeraWulf can dynamically allocate megawatts between Bitcoin mining and AI hosting based on real-time hashprice signals. When hashprice is high, shift capacity to ASICs. When hashprice drops, shift to GPUs. This optionality is the economic moat that traditional data centers cannot replicate, and it explains why former mining companies trade at premiums to pure-play data center operators.

Bitfarms: Full GPU-as-a-Service

Bitfarms represents the most aggressive end of the spectrum. The company is committing its entire 341 MW portfolio to a GPU-as-a-Service (GAAS) model by 2027, effectively exiting Bitcoin mining as a primary business. The rationale is margin-driven: at current hashprice levels, every megawatt generates higher returns hosting AI inference workloads than mining Bitcoin.

This is not retreat from crypto. It is economic optimization. If hashprice recovers to levels that make mining more profitable than AI hosting, Bitfarms can reallocate capacity. But the structural trend — growing AI compute demand against compressed mining margins — favors the pivot for the foreseeable future.

Which countries are becoming Bitcoin mining superpowers and why?

The Hormuz energy shock exposed the fragility of mining operations concentrated in regions dependent on fossil fuel pricing. The rational economic response is geographic migration toward stranded and renewable energy sources. Three countries are emerging as next-generation mining hubs: Paraguay, Ethiopia, and Brazil.

Table: Emerging Bitcoin mining hubs leveraging stranded energy. Paraguay leads with formalized state-level mining operations.

Paraguay: The state mining pioneer

Paraguay sits on a structural energy surplus. The Itaipú Dam, shared with Brazil, generates 14 GW of installed capacity, but Paraguay’s domestic demand consumes only a fraction. For decades, this excess was sold to Brazil at below-market rates. Bitcoin mining offers a higher-value alternative for monetizing that surplus.

In March 2026, Paraguay’s national electricity utility ANDE formalized a partnership with Morphware under Resolución 47/2026. This agreement represents one of the first state-level Bitcoin mining operations in the world. The current 700 MW of mining capacity is targeting 1 GW by 2027, powered entirely by hydroelectric surplus at rates below $0.03/kWh.

The economic logic is compelling. Rather than exporting cheap electricity to Brazil, Paraguay monetizes its energy surplus through Bitcoin mining at significantly higher margins. At sub-$0.03/kWh electricity and sub-10 J/TH hardware, Paraguay-based miners can produce Bitcoin at approximately $20,000–25,000 per BTC — well below the current market price and completely insulated from fossil fuel price volatility.

The ANDE–Morphware agreement also signals a broader shift: sovereign utilities are beginning to view Bitcoin mining as legitimate industrial load, comparable to aluminum smelting or data center operations. This regulatory normalization is critical for attracting the institutional capital needed to scale operations to gigawatt levels.

Ethiopia: Hydroelectric frontier

Ethiopia’s Grand Ethiopian Renaissance Dam (GERD) on the Blue Nile represents one of Africa’s largest infrastructure projects. With 6.45 GW of planned capacity and limited domestic industrial demand in the near term, the country faces the same stranded energy problem as Paraguay. Several mining operations have begun establishing facilities near hydroelectric sources, though the regulatory framework remains under development.

The timeline is longer than Paraguay — meaningful hash rate contribution is not expected before 2027 — but the scale potential is significant. Ethiopia’s electricity costs are among the lowest globally, and the government has signaled openness to digital asset infrastructure as a mechanism for technology transfer and foreign currency generation.

Brazil: Solar curtailment meets mining

Brazil presents a different model: mining as a solution to grid imbalance. The country’s rapid solar expansion, particularly in the northeast, has created curtailment problems — periods when solar generation exceeds grid demand and transmission capacity. Engie Brasil’s Assu Sol project targets 895 MW of solar-integrated mining infrastructure, using Bitcoin mining as a flexible load that absorbs excess generation.

Under ANEEL (Brazil’s National Electrical Energy Agency) regulation, mining operations connected to curtailed solar capacity receive preferential energy pricing. This creates a symbiotic relationship: solar developers get a guaranteed buyer for otherwise wasted energy, and miners access electricity costs near zero during peak solar hours. The model is similar to the flared gas mining operations in Texas and North Dakota but applied to renewable curtailment at much larger scale.

What this means for hash rate geography

The geographic migration is not replacing US mining capacity — it is supplementing it. The United States retains the largest share of global hash rate, driven by institutional-grade facilities like TeraWulf’s Lake Mariner and Core Scientific’s network. But the post-Hormuz landscape is redistributing marginal hash rate growth toward the Global South, where energy economics provide a structural advantage.

This redistribution has a security benefit. Greater geographic diversity makes the Bitcoin network more resilient to localized energy shocks, regulatory changes, or weather events. The concentration risk exposed by the Foundry disconnection during Winter Storm Fern becomes less systemic when hash rate is distributed across multiple continents and energy sources.

How does Bitcoin’s difficulty adjustment protect the network during energy shocks?

Bitcoin’s difficulty adjustment is the protocol’s built-in shock absorber. Every 2,016 blocks (approximately 14 days), the network recalculates the computational difficulty required to mine a valid block. The target is consistent: one block every 10 minutes, regardless of how much hash rate is online.

The February 2026 adjustment: A case study

When the Hormuz energy shock and Winter Storm Fern combined to force significant hash rate offline in early 2026, block times stretched beyond the 10-minute target. Blocks that normally took 10 minutes were averaging 12–14 minutes. Transaction fees spiked as the mempool backlog grew. The network was functioning — but under stress.

The February 2026 difficulty adjustment reduced mining difficulty by 16–18%. This was one of the largest single downward adjustments since the China mining ban of 2021. The effect was immediate and mechanical:

• Block times normalized: With lower difficulty, the remaining online miners could find valid blocks faster, returning block times to approximately 10 minutes within hours of the adjustment.
• Marginal miners became profitable: Lower difficulty means each unit of hash rate earns a proportionally larger share of block rewards. Some S21 operators who were borderline unprofitable before the adjustment returned to positive margins.
• New capacity was incentivized: The reduced difficulty effectively increased the reward-per-hash, attracting new mining capacity from regions unaffected by the energy shock — particularly hydroelectric-powered operations in Paraguay and Scandinavia.
• Network stabilized within two weeks: From the adjustment, the network returned to steady-state operation, with block times, fee markets, and hash rate all normalizing.

Why this makes Bitcoin anti-fragile

The difficulty adjustment mechanism means Bitcoin cannot be permanently damaged by an energy shock. In the worst case, block times slow temporarily, fees increase, and the user experience degrades for one adjustment period. But the protocol self-corrects. No central authority needs to intervene. No emergency governance vote is required. The math handles it.

Compare this to traditional financial infrastructure. When energy prices spike, stock exchanges can circuit-break, banks can freeze operations, and payment processors can experience outages that require human intervention to resolve. Bitcoin’s difficulty adjustment is algorithmic, automatic, and permissionless — precisely the qualities that make it resilient to geopolitical disruption.

The February 2026 event also demonstrated a second-order effect: difficulty adjustments create temporary profit windows. Miners who can deploy new capacity quickly — particularly those with access to cheap energy in places like Paraguay — capture outsized returns during the post-adjustment period before difficulty recovers upward. This creates a natural incentive for geographic diversification of hash rate, which in turn strengthens the network’s long-term resilience.

What are the hash rate projections for 2027–2028?

Projecting hash rate requires modeling three variables: hardware deployment cycles, energy cost trajectories, and geographic capacity additions. The post-Hormuz landscape has reshaped all three, but the directional trend is unambiguous: hash rate will continue growing through the next halving cycle.

Table: Hash rate projections through the 2028 halving. Growth driven by hardware efficiency gains and Global South capacity additions.

April 2026: Stabilization phase

The immediate outlook calls for hash rate stabilization at 1.10–1.15 ZH/s. The S19 capitulation is largely complete — operators who could not sustain $79,995/BTC production costs have already powered down or sold machines for scrap. The remaining hash rate consists primarily of S21-generation and newer hardware operating in low-cost energy regions.

The February difficulty adjustment has already normalized mining economics for surviving operators. Block times have returned to target, and hashprice has recovered modestly from its $30/PH/s low as marginal capacity comes back online. The immediate risk factor is continued energy price volatility if the Hormuz situation escalates further.

December 2026: The S21 deployment wave

The second half of 2026 is expected to see significant hash rate growth, driven by two factors. First, Bitmain’s S21 and S23 production is ramping to meet demand from operators replacing retired S19 fleets. The efficiency gains are substantial: an S23 operating at sub-10 J/TH produces 40% more hash rate per watt than an S19 at 16 J/TH.

Second, state-level mining operations in Paraguay are expected to reach operational scale. The ANDE–Morphware agreement targets first production in Q3 2026, with ramp-up to 700 MW by year end. At sub-$0.03/kWh power costs, these operations will be among the most profitable in the world, adding significant hash rate that is insulated from fossil fuel price volatility.

2027: Geographic integration

By March 2027, the projected 2.00 ZH/s reflects the integration of multiple new mining regions. Paraguay’s target of 1 GW mining capacity would represent roughly 50–70 EH/s depending on hardware mix. Ethiopian mining operations are expected to begin contributing measurable hash rate. Brazil’s solar curtailment mining model should reach commercial scale under the Engie partnership.

The 2027 hash rate landscape will be meaningfully more geographically diverse than today. While the US will likely retain the largest single-country share, the combined Global South contribution could reach 15–20% of total hash rate — up from less than 5% in 2024. This geographic diversification reduces systemic risk from any single energy market or regulatory jurisdiction.

2028: Pre-halving dynamics

The next Bitcoin halving (expected April 2028) will reduce the block subsidy from 3.125 BTC to 1.5625 BTC. Historically, the 12–18 months before a halving see aggressive hash rate deployment as miners maximize revenue at the current subsidy level before it drops.

The 2.30+ ZH/s projection for pre-halving 2028 accounts for several converging factors: next-generation ASIC deployment (potentially sub-7 J/TH efficiency), continued expansion of Global South mining capacity, and a possible “AI crossover” effect where dual-use facilities contribute hash rate from capacity that is primarily justified by AI revenue. While institutional Bitcoin accumulation may also influence miner economics through price effects, the hash rate growth trajectory is primarily driven by hardware efficiency improvements and energy access expansion.

What should mining operators do now?

The post-Hormuz landscape rewards three characteristics: energy cost discipline, hardware efficiency, and revenue diversification. Operators who possess all three will thrive through the next halving cycle. Those with only one or two face existential risk at the next energy shock or subsidy reduction.

• Retire S19 hardware immediately: At $79,995/BTC production cost, continuing to operate S19 machines is burning capital. Even at $0.05/kWh, the math does not work. Salvage value and resale to operators in ultra-low-cost regions is the rational exit strategy.
• Evaluate AI/HPC conversion: Not every mining facility can host GPUs — it requires upgraded networking, cooling modifications, and customer contracts. But facilities with grid interconnection agreements and power purchase contracts below $0.05/kWh should actively pursue AI/HPC revenue as a hedge against hashprice volatility.
• Explore Global South partnerships: The ANDE–Morphware model in Paraguay demonstrates that state-level energy partnerships can provide sub-$0.03/kWh access. Mining operators with operational expertise and hardware supply chains can joint-venture with sovereign utilities to deploy capacity at costs unreachable in North America or Europe.
• Build for the halving: Every investment decision from now through April 2028 should be evaluated against the halving math. A machine that is profitable today at 3.125 BTC block subsidy needs to be profitable at 1.5625 BTC — or it needs to pay for itself before the halving arrives.

Is Bitcoin mining dying or evolving?

The narrative of “Bitcoin mining is dying” appears after every energy shock, every hashprice collapse, and every halving event. It has been wrong every time. What is dying is the model of mining as a single-revenue-stream business dependent on cheap fossil fuel electricity and commodity ASIC hardware.

What is replacing it is something more resilient: a hybrid infrastructure model where the same megawatt of power and the same physical facility can dynamically serve Bitcoin mining, AI compute, and general-purpose HPC workloads. This model is energy-source-agnostic (hydro, solar, nuclear, natural gas), geographically distributed, and revenue-diversified.

The Hormuz blockade didn’t break Bitcoin mining. It accelerated a transformation that was already in motion. The companies and countries positioned to benefit — Core Scientific with its $10.2B CoreWeave contract, Paraguay with its hydroelectric surplus, TeraWulf with its dual-use infrastructure — are building the mining industry of the next decade.

For the broader Bitcoin network, the outcome is unambiguously positive. Hash rate will recover and grow past 2.30 ZH/s by 2028. Geographic distribution will improve as the Global South reaches 15–20% of total hash rate. And the difficulty adjustment will continue doing what it has always done: protect the network, absorb the shock, and reset the economics — no matter what happens in the world above.

For macro implications of the Hormuz crisis on cryptocurrency markets, see our Iran–Crypto Market Outlook: April 2026.

Related reading