Bitcoin mining's industrial economics have deteriorated to their weakest point since the April 2024 halving. Hashprice — the standard revenue metric per unit of computational power — sits near $29/PH/s/day in July 2026, a five-year low. According to CoinShares' Q1 2026 mining report, the weighted...
"Execution, not signing, becomes the next premium. Missing construction milestones risks structural de-ratings." — Griffin MacMaster and Matthew Sigel, VanEck
Bitcoin mining's industrial economics have deteriorated to their weakest point since the April 2024 halving. Hashprice — the standard revenue metric per unit of computational power — sits near $29/PH/s/day in July 2026, a five-year low. According to CoinShares' Q1 2026 mining report, the weighted average cash cost to produce one bitcoin among publicly listed miners rose to approximately $79,995 in Q4 2025, while BTC traded in the $68,000–$70,000 range, implying losses of roughly $10,000–$12,000 per coin mined for the median operator.
Simultaneously, record heat in the eastern United States forced the Department of Energy to issue emergency orders to the PJM Interconnection grid on June 30, pushing wholesale electricity prices from ~$40/MWh to over $600/MWh in Virginia and compelling miners to curtail operations. The resulting hashrate dip — estimated at 1%–3% globally — exposed the fragility of a mining industry that now competes directly with AI data centers for scarce grid capacity.
This report examines how Bitcoin mining's economic model is being squeezed from two sides: collapsing unit economics post-halving and an energy grid that increasingly treats miners as discretionary load. The industry's response — a $70 billion collective pivot toward AI infrastructure — carries its own $50 billion funding gap, according to VanEck.
On April 20, 2024, at block height 840,000, Bitcoin's block subsidy dropped from 6.25 BTC to 3.125 BTC. The network now produces approximately 450 BTC per day across ~144 blocks. At $68,000 per BTC, daily subsidy revenue across the entire network is approximately $30.6 million — split among miners deploying a combined ~900 EH/s of hashrate.
The cost structure tells the story. According to CoinShares' Q1 2026 report, electricity costs averaged $52,463 per BTC mined, with depreciation and amortization adding $58,381/BTC. Transaction fee income — once hoped to offset subsidy declines — has remained consistently below 1% of total block rewards throughout 2026.
Hardware efficiency determines survival. The current generation breakdown:
| ASIC Model | Efficiency (J/TH) | Breakeven Electricity Rate | |---|---|---| | Antminer S21 XP (2025) | 13.5 | ~$0.088/kWh | | Antminer S21 (2024) | 17.5 | ~$0.068/kWh | | Antminer S19 XP (2022) | 21.5 | ~$0.055/kWh | | Antminer S19 Pro (2020) | 30.0 | Unprofitable at most rates |
At current hashprice levels near $30/PH/s/day, any machine less efficient than an S19 XP running on power priced above $0.06/kWh is operating at a loss. CoinShares estimates this applies to 15%–20% of the global mining fleet. Industrial-scale miners with hosting rates below $0.08/kWh and sub-15 J/TH hardware report margins of 20%–50%. Everyone else is losing money or breaking even.
The shift from 98 J/TH (Antminer S9, 2017) to sub-15 J/TH represents a 7x improvement in energy efficiency over eight years. Yet efficiency gains have been offset by difficulty increases: network difficulty peaked at 155.97 trillion after a +6.31% adjustment on October 29, 2025, and currently sits near 134 trillion following six downward adjustments in 2026.
On June 30, 2026, the U.S. Department of Energy issued emergency orders to PJM Interconnection — the grid operator covering 13 eastern states and the District of Columbia — to prevent blackouts as a heat dome settled over the region. PJM forecast peak electricity demand of 166,147 MW, which would have exceeded its all-time record of 165,563 MW set in 2006.
Wholesale electricity prices in Virginia surged from approximately $40/MWh to over $600/MWh during peak afternoon hours. Bitcoin miners operating within PJM's territory — which spans from Virginia to Illinois — faced a binary choice: curtail operations or absorb electricity costs roughly 15x their normal rate.
Most curtailed. The global hashrate dipped 1%–3% during the event, consistent with prior heat-wave-induced curtailments. The pattern is now recurring: extreme weather events in ERCOT (Texas) and PJM territories force temporary mining shutdowns, creating brief but measurable impacts on network security metrics.
The National Weather Service has projected a moderate risk of extreme heat for July 14–19, 2026. Investors are monitoring PJM reserve margins, wholesale electricity pricing, and mining company curtailment disclosures ahead of the event.
The EIA's July 2026 Short-Term Energy Outlook forecasts wholesale power to average approximately $45/MWh this summer nationally. That average obscures the volatility: a single heat wave can push localized prices above $500/MWh for hours.
Bitcoin's network hashrate briefly touched the 1 zettahash-per-second (ZH/s) milestone — equivalent to 1,000 EH/s — in late May 2026, with the seven-day moving average hovering near that threshold. The milestone was short-lived.
By early June, hashrate had fallen to approximately 861 EH/s. As of early July, the network has recovered to approximately 894–924 EH/s, depending on the measurement window. The volatility reflects both the grid curtailment events described above and the broader economic pressure on marginal miners.
The 1 ZH/s figure matters symbolically but says little about network security or miner health. What matters is the relationship between hashrate, difficulty, and hashprice. Difficulty has adjusted downward six times in nine epochs during 2026, signaling that hashrate additions are not keeping pace with the upward difficulty trajectory that characterized 2024–2025.
The U.S. Energy Information Administration projects national electricity consumption will reach 4,269 billion kWh in 2026 and 4,399 billion kWh in 2027 — both records — driven by AI data centers, cryptocurrency mining, and broader electrification. Bitcoin mining's estimated annual consumption of approximately 128 TWh represents roughly 3% of projected U.S. demand.
Public Bitcoin miners have collectively announced over $70 billion in AI and high-performance computing (HPC) contracts, according to CoinShares. The strategic logic is straightforward: mining facilities possess three assets that AI companies need — electrical infrastructure, cooling systems, and grid interconnection agreements. Converting mining capacity to AI colocation offers revenue stability that Bitcoin's volatile hashprice cannot match.
VanEck's June 2026 analysis, authored by Griffin MacMaster and Matthew Sigel, quantified the gap between ambition and execution. The firm estimates miners face a $50 billion near-term funding gap and as much as $221 billion in long-term capital needs to deliver on announced AI projects. Only approximately 25% of leased AI and HPC capacity had been delivered at the time of the report.
The valuation framework has shifted accordingly. VanEck identifies "energized power" — operational, grid-connected capacity — as the clearest metric. Companies with signed and delivered AI leases command valuation multiples above 10x energized power. Miners still pitching future projects trade at lower multiples. Listed miners may derive as much as 70% of their revenue from AI by the end of 2026, up from approximately 30% at the start of the year.
The pivot is creating a two-tier industry. Selected Q1 2026 results illustrate the split:
Core Scientific (CORZ): Signed a $10.2 billion, 12-year hosting agreement with CoreWeave. AI colocation revenue represented 39% of total sales in Q1 2026. Management has stated that only one or two sites will remain operational for Bitcoin mining by year-end, with the company prioritizing high-density colocation infrastructure.
Riot Platforms (RIOT): Posted $167.2 million in Q1 2026 revenue. Bitcoin mining revenue fell to $111.9 million from $142.9 million in Q1 2025. The data center business contributed $33.2 million after AMD doubled its contracted capacity from 25 MW to 50 MW. RIOT stock gained approximately 94% year-to-date despite Bitcoin's ~24% decline from January levels.
MARA Holdings (MARA): Reported a $1.26 billion net loss for Q1 2026. The company sold over $1 billion in BTC in recent months to fund AI infrastructure investments — a reversal of its previously stated HODL strategy. VanEck categorizes MARA as remaining "more closely tied to bitcoin's price performance" than to AI execution.
CleanSpark (CLSK): Has maintained a comparatively pure-mining focus with selective AI-adjacent infrastructure investments. VanEck placed CLSK alongside MARA and RIOT as names whose valuations remain primarily BTC-correlated.
VanEck identified HIVE, Bitdeer (BTDR), Keel, and IREN as potential upside names if they secure additional AI contracts. IREN, in particular, was highlighted: the company's stock performance reflects investor confidence in its AI conversion progress.
The mining industry's narrative around grid participation is under pressure. Miners have positioned themselves as flexible loads that curtail during peak demand, earning demand-response payments while stabilizing the grid.
The data partially supports this claim. In ERCOT (Texas), crypto mining electric demand reached 4,288 MW as of November 2025, with projections to surpass 5,300 MW by 2027. ERCOT's Large Flexible Load (LFL) program has approved up to 9,500 MW of flexible demand capacity. Miners participate in ERCOT's Emergency Response Service (ERS), receiving payment for availability to curtail within 10 minutes during grid emergencies.
The counterargument: AI data centers are now competing for the same grid interconnection slots, and they offer higher revenue per MW with more predictable load profiles. A July 2, 2026, Miner Weekly analysis posed the question directly — whether flexible AI load could displace bitcoin mining's grid pitch entirely.
The Congressional Research Service published a report on April 22, 2026, examining cryptocurrency mining's impact on the electricity sector. The CRS analysis notes that miners have until 2027 to prove they deserve power on America's overloaded grid, referencing pending regulatory reviews of interconnection agreements and demand-response qualification criteria.
Bitcoin mining in mid-2026 faces a structural contradiction. The economics of the core business — producing BTC via proof-of-work — have deteriorated to the point where only operators with sub-$0.06/kWh electricity and latest-generation ASICs can sustain positive margins. The halving arithmetic is unforgiving: at 3.125 BTC per block and current difficulty levels, the median public miner is producing bitcoin above market price.
The AI pivot offers a path to revenue diversification, but the gap between announced contracts and delivered infrastructure is material. VanEck's $50 billion funding gap estimate suggests that many miners will struggle to complete the transition without additional capital raises, diluting existing shareholders.
The grid question adds a third variable. Miners have spent years building a narrative as flexible demand-response assets. That narrative is now contested by AI data centers offering the same grid flexibility with higher revenue per megawatt. If regulators and grid operators begin preferring AI loads over mining loads in interconnection queues, the industry's infrastructure advantage — its core asset beyond the ASICs themselves — erodes.
The next 18 months will determine whether Bitcoin mining survives as a standalone industry or becomes a transitional phase in the buildout of AI infrastructure. The data, as of July 2026, suggests the latter outcome is increasingly probable for publicly listed operators.