Core Scientific Pivots From Bitcoin Mining to AI Data Centers as Two Seas Capital Gains Sway Before March 2 Earnings

Core Scientific's Pivot To AI Data Centers: What Operations Leaders Should Do Now

Core Scientific is shifting from Bitcoin mining to AI-focused data center operations. The company ended merger talks with CoreWeave and is targeting 590MW of capacity without raising additional capital. Two Seas Capital is expanding its board presence, and Chairman Jordan Levy is stepping down-both moves point to tighter execution and a data center-first mandate.

Near term, the stock sits between $17.56 and $17.67-below the 20-day moving average, but above the 50-day and 200-day moving averages. Momentum reads (RSI, CCI) lean mildly bearish, with Stochastic RSI neutral. Analysts note $17.08 as a key level; holding it and delivering on AI milestones could improve the odds of a short-term move higher. The Q4 2025 report and March 2, 2026 earnings call will be the operational litmus test.

Why This Shift Matters For Operations

AI workloads are power-dense, thermally demanding, and supply-chain constrained. Moving from mining to AI/HPC changes almost everything: cooling strategy, rack density, network fabrics, interconnect timelines, and the commercial model (capacity pre-sold vs. speculative build). Execution is less about hash rate and more about MW energized, liquid cooling readiness, and time-to-revenue per site.

This pivot also tracks with broader data center demand trends, but it runs into the real-world limits of grid readiness and long-lead hardware. Your edge is operational clarity: what will be built, when it's energized, and how reliably it can host GPUs at 30-80kW per rack.

What To Watch On The Earnings Call (Ops Lens)

• Capacity and timing: MW energized today, MW under construction, MW with signed interconnect agreements, and the cadence to reach 590MW.
• Cooling readiness: percentage of capacity liquid-cooled (direct-to-chip or immersion), expected PUE at AI densities, and water strategy (WUE, heat reuse).
• Contract mix: committed vs. uncommitted capacity, contract terms (tenor, take-or-pay), prepayments, and cancellation protections.
• Capex control without new equity: site-level capex per MW, procurement status for transformers, switchgear, chillers/pumps, and network gear.
• GPU hosting assumptions: target rack densities, power envelopes, and network fabrics (400G/800G, RDMA readiness).
• Site-risk profile: interconnect queue positions, substation upgrades, permitting timelines, and any single-point vendor dependencies.

Governance Moves: Why Ops Should Care

Two Seas Capital increasing board influence suggests more pressure on capital discipline and measurable milestones. Leadership change at the chair level supports a more focused data center operating model. Expect tighter oversight of build schedules, vendor selection, and cash conversion. Similar strategic moves are being seen at peers like Riot Platforms.

Technical Setup: Planning Signals For Ops

Price is below the 20-day average but above the 50/200-day-short-term pressure, longer-term support intact. Indicators are mildly bearish and price sits near the top of its daily range, implying potential consolidation before a move. Translation for operations: budget and schedule with buffers. Avoid locking in timelines that depend on best-case interconnect or hardware deliveries.

Immediate Priorities For Data Center Operations

• Interconnect and power: secure queue positions, validate transformer lead times, and stage substation upgrades to phase MW online quarterly, not yearly.
• Thermal design at scale: standardize liquid cooling blocks, manifold design, and service workflows; model hot-aisle containment and redundancy at 30-80kW per rack.
• Supply chain: pre-buy long-lead equipment (transformers, switchgear, busway, chillers, pumps) with alternates approved; set dual-sourcing for high-risk items.
• GPU capacity onboarding: align power, cooling, and fabric timelines with customer GPU deliveries; dry-run installs with mock racks and spare parts on hand.
• Network fabrics: lock reference architectures for 400/800G, plan leaf-spine growth, and validate optical budgets at target rack densities.
• Reliability and maintenance: define SLAs tied to AI uptime, schedule preventive maintenance windows per cluster, and stock failure-prone spares.
• Energy strategy: evaluate PPAs and demand response; model cost per kWh under peak load and curtailment scenarios.
• Compliance and reporting: track PUE/WUE, carbon reporting, and safety procedures for high-density liquid systems.

Risk Map And Practical Mitigations

• Grid delays: maintain multiple interconnect paths and sites; modularize substations for faster energization.
• Transformer/switchgear shortages: place blanket POs, qualify secondary vendors, and keep swap-compatible specs.
• Cooling complexity: pre-commission manifolds and CDU loops; train field techs on liquid-handling SOPs and spill response.
• GPU supply variability: build flexible PODs that can run mixed densities; use phased cold-aisle buildouts to avoid stranded capacity.
• Water constraints: prefer closed-loop liquid and adiabatic minima; explore heat reuse partnerships to de-risk permits.
• Cost volatility: hedge power where possible; structure contracts with inflation clauses and energy pass-through terms.

Key Numbers From The Market Snapshot

• Trading range: ~$17.56-$17.67, below the 20-day MA and above the 50/200-day MAs.
• Indicators: RSI/CCI mildly bearish; Stochastic RSI neutral.
• Watch level: $17.08 as support; holding it plus AI delivery increases odds of a short-term uptick.
• Timeline: Q4 2025 earnings release and March 2, 2026 call-expect updates on AI data center milestones and capacity.

Investor Activity And What It Signals For Ops

Two Seas Capital increasing involvement points to stronger oversight and focus on execution quality. Paloma Partners cutting its exposure-from 8.1% to 0.49% of its U.S. equity assets-shows some institutions remain skeptical about execution speed and certainty. For operations teams, this means scrutiny on timelines and proof of progress will be high-site milestones need to be auditable and frequent.

90-Day Execution Checklist

• Finalize site-by-site MW ramp plan with interconnect dates, equipment ETAs, and handover milestones.
• Lock thermal reference designs (liquid loops, CDUs, leak detection) and train staff on install/repair SOPs.
• Issue POs for transformers/switchgear with alternates; create a swap guide to handle vendor substitutions without redesign.
• Run a full-stack simulation: power-on to GPU burn-in to network throughput; document bottlenecks and fixes.
• Define customer onboarding playbook: rack allocation, network turn-up, telemetry, and acceptance tests.
• Set weekly KPI cadence: MW energized, MW contracted, PUE at load, install cycle time, and incident rate.

Signals Of Real Progress To Communicate Externally

• Quarterly MW energized vs. plan and percentage liquid-cooled capacity ready for AI loads.
• Contracted backlog (MW and duration) and time-to-revenue from energization.
• Supply chain on-time delivery for transformers, switchgear, and cooling equipment.
• Stability metrics: PUE at target density, incident MTTR, and SLA adherence.

Additional Resources

• IEA: Data centres and data transmission networks - context on power demand and grid impact.
• AI for Operations - frameworks and playbooks for AI data center buildouts and reliability at scale.

Bottom line: the strategy is set-590MW without new capital, AI-first design, and tighter governance. The only story that matters now is steady MW online, liquid cooling ready, contracts signed, and uptime proven. Build for that, report it clearly, and keep buffers where the grid and supply chain can slip.

For information only; not investment advice.