Why Core Scientific Stock Crumbled in March

Q4 2025 revenue was $79.8M, down 15.9% YoY from $94.9M, and adjusted EBITDA swung to a loss of ~$43M versus a $13M gain a year earlier (≈$56M deterioration). GAAP net income was $216M, largely driven by a >$330M non-cash gain from warrants and CVRs. Management plans to monetize substantially all Bitcoin holdings over 2026 to boost liquidity and fund capex, and the company secured a $1.0B, 364‑day loan facility from a JPMorgan unit. Shares saw roughly a 12% decline in March amid investor selling on the disappointing fundamentals and accelerated pivot to AI-capable data centers.

Analysis

The industry pivot from high-density Bitcoin mining to AI-capable data centers reallocates capital and alters component demand: GPUs, high-speed interconnects, and dense power-distribution gear replace ASIC inventory and bespoke immersion rigs. Over a 12–24 month window this can create a concentrated uplift in GPU procurement cycles but also a transient oversupply of power/thermal specialists and used ASIC equipment that depresses secondary markets and drags valuations of legacy-mining operators.

Bank-sourced liquidity for transition plays lowers immediate default probability but increases covenant and refinancing dependency; that tends to compress equity volatility initially while leaving credit spreads vulnerable to a single negative AI-occupancy or crypto-price shock. The most likely near-term tail is forced liquidation of liquid crypto assets by leveraged operators, which can amplify crypto market drawdowns over days-to-weeks and feed back into risk premia on small-cap miner equities and specialized equipment vendors.

Strategically, hardware winners are those with software and provisioning stacks that capture recurring revenue (driving stickiness beyond one-time rack sales) — this favors GPU + orchestration ecosystems over standalone silicon suppliers without a services layer. Counterparty banks and large dealer ECNs that distribute financing will see fee and trading flow benefits; conversely, vendors dependent on long lead-time ASIC upgrades face multi-quarter demand erosion. Monitor GPU orderbooks, resale ASIC volumes, and bank covenant amendments as the early quantitative signals of success or failure in the pivot.

Key Decisions for Investors

• Long NVDA via 3–6 month call-spread (buy ~0.6–0.8 delta, sell ~0.25–0.35 delta) to capture near-term demand reallocation into GPUs; target asymmetric payoff ~3:1 reward:risk. Enter on <5% pullback or ahead of reported large AI-capacity bookings.
• Pair trade — long NVDA / short INTC (equal notional, 6–12 month horizon): NVDA exposure to software/ecosystem should outperform INTC's more cyclical CPU exposure if AI rack conversions accelerate. Risk: execution dependence on NVDA maintaining pricing power; take profits if spread narrows <10% relative performance or widen >30%.
• Credit-sensitive trade — buy JPM equity or short-dated investment-grade CDS on regional lenders that lack balance-sheet capacity (6–12 months): banks underwriting transition financing should capture fees and spread compression; prefer bank exposure over small-cap miner creditors. Risk: macro tightening or loan losses could flip this trade quickly.
• Hedge / event short — buy put spreads on small-cap miner/operator equities or construct relative-value shorts in the miner cohort for 1–3 month windows to protect against accelerated liquidation of liquid crypto assets. Set stop-loss if miner CDS widens >150bps or if crypto indicators show rapid recovery.