Crypto Mining Stocks Rally While Bitcoin Slips: Cipher Mining Jumps 8%, MARA Climbs 6%

Crypto-mining equities rallied despite Bitcoin weakness, with Cipher Mining up 8% to $19.18, MARA up 6% to $11.96, IREN up 7%, and TeraWulf up 9%, while IBIT fell 2% and Bitcoin declined 0.7% to $78,666.21. Investors are increasingly valuing miners' power capacity and grid rights as AI/HPC infrastructure: Cipher targets roughly $793 million of annualized HPC NOI and has funded its Stingray facility through an $810 million, 6% bond offering. The divergence remains unproven, however, as MARA has yet to secure an AI lease, both companies missed Q2 2026 estimates, and a further Bitcoin selloff could re-establish the sector's historical correlation to crypto prices.

Analysis

The relevant rerating is not “AI exposure” broadly; it is the conversion of stranded/contracted power into durable, creditworthy capacity revenue. CIFR and IREN have the strongest potential to earn data-center-style valuation treatment because a greater share of their thesis can be underwritten against contracted cash flow rather than BTC price, difficulty growth, and halving risk. MARA and WULF remain more exposed to execution sequencing: power rights and conversion plans have option value, but that value should be discounted until customer commitments, financing terms, and delivery milestones are independently verified.

Near term, the divergence from IBIT is vulnerable to reversal because crypto-miner shareholder bases remain dominated by BTC-beta capital and momentum flows. A further BTC drawdown can force cross-sector deleveraging irrespective of HPC fundamentals, while elevated capex and fixed-power obligations increase downside asymmetry for uncontracted sites. Over 1-3 months, the key catalyst is evidence that rent has commenced and that utilization, customer credit, and build costs are tracking underwriting assumptions; over 6-18 months, valuation depends on whether AI demand translates into signed long-duration leases rather than merely higher land-and-power appraisals.

Consensus may be overpaying for gross MW rather than monetizable MW. Interconnection rights are scarce, but their value is highly location-specific: transmission congestion, curtailment exposure, cooling/water constraints, network connectivity, and time-to-delivery determine whether a site competes with purpose-built data-center campuses. The cleaner expression is therefore relative—long contracted-HPC execution, short uncontracted BTC sensitivity—rather than a blanket long miner basket.

Key Decisions for Investors

• Initiate a 3-6 month pair: long CIFR / short MARA, sized beta-neutral to BTC. Thesis is that CIFR’s cash-flow conversion should support relative multiple expansion while MARA still requires a signed AI lease; target 20-30% relative upside. Exit if MARA secures a credible long-duration lease with disclosed economics or CIFR misses rent/utilization milestones.
• Use IREN as the secondary long rather than adding broad miner exposure; accumulate only on BTC-driven weakness, not after momentum spikes. Require confirmation that HPC revenue/backlog is converting into operating cash flow; a sustained deterioration in contracted economics or financing spreads falsifies the thesis.
• Avoid treating WULF as a pure AI beneficiary until project financing, customer identity/credit quality, and construction economics are disclosed. Set an alert for binding lease announcements and compare implied EV per energized, usable MW against CIFR/IREN before initiating.
• Hedge residual crypto-beta in any miner-long book with IBIT puts or a partial IBIT short through the next 1-3 months. The hedge should be reduced only if miners hold relative performance through a meaningful BTC decline and subsequent earnings validate non-mining revenue contribution.