CORZ

Moat Evaluation

Core Scientific’s documents from November 2024 through October 2025 point to an emerging moat centered on efficient scale and switching costs in high-density AI/HPC colocation. The most recent development is the October 30, 2025 termination of the planned CoreWeave merger, which leaves Core Scientific independent but still linked to CoreWeave through sizable long-term hosting agreements; the filings do not clarify any changes to those commercial agreements, so implications are uncertain. Evidence throughout 2025 highlights rapid buildout of power-dense capacity, multi-year contracted revenue with a major hyperscale customer, and municipal approvals that expand power access—factors consistent with cost advantage and efficient scale. At the same time, customer concentration and transitional financials temper durability until the customer mix broadens and the new footprint is fully monetized.

Emerging efficient-scale and switching-cost moat in AI/HPC colocation, tempered by customer concentration

Core Scientific is repurposing a large, hard-to-replicate power and real estate footprint into high-density, application-specific data centers for AI workloads, supported by long-duration contracts. In Q4 2024 and early 2025, the company reported a 1.3 GW powered infrastructure portfolio (Feb 26, 2025) and city-approved expansions that increase Denton’s access to power to 394 MW (Nov 20, 2024), with Denton alone moving to ~260 MW of critical IT load under a February 26, 2025 agreement. Additional scale projects—Muskogee (100 MW groundbreaking on Nov 18, 2024) and Auburn, AL (Feb 19, 2025)—reinforce an efficient-scale position where few sites can secure and deliver comparable high-density power within near-term timelines. These sites, contracts, and delivery milestones—such as a 16 MW tranche delivered 30 days ahead of schedule (Feb 26, 2025 earnings call)—indicate operational capabilities that can reduce unit costs and shorten time-to-revenue relative to peers.

Switching costs appear meaningful: multi-year terms (12-year hosting agreements with options; cited Feb 26, 2025) for customized, GPU-ready facilities imply embedded hardware, interconnection, and latency-sensitive workloads that are difficult and costly to relocate. Q1 2025 and Q2 2025 MD&A emphasize that colocation should provide more stable, less volatile revenue than Bitcoin mining (Feb 27, 2025; Aug 7, 2025), and Q1 2025 guidance targeted ~$360 million annualized colocation revenue entering 2026 (May 7, 2025). However, CoreWeave is both the anchor customer and a key capex funder (e.g., $196.4 million of Q3 2025 capex funded by CoreWeave; Oct 24, 2025), and the October 30, 2025 merger termination injects uncertainty around long-term strategic alignment even if contracts remain in force. Near-term financials reflect transition costs and mix shift (e.g., Q3 2025 total revenue down year over year; Aug 8 and Oct 24, 2025), suggesting the moat is still being built and will depend on diversifying customers—as management has targeted over time—to avoid overreliance on a single counterparty.

Top 3 Patterns Identified

1: Pivot from Bitcoin mining to AI/HPC colocation with long-dated, site-specific commitments

• Recent Evidence: On October 24, 2025, the company reported HDC revenue of $15.0 million in Q3 2025 (up from $10.3 million in Q3 2024) alongside continued expansion of colocation operations. Earlier in 2025, management reiterated multi-year, contract-backed growth with Denton’s expansion to ~260 MW and total contracted HPC infrastructure of ~590 MW, lifting projected revenue under CoreWeave agreements to ~$10.2 billion over 12 years (Feb 26, 2025). Q1 2025 materials forecast entering 2026 with ~$360 million annualized colocation revenue (May 7, 2025).
• Contextual Trends: MD&A documents from Feb 27, 2025 and Aug 7, 2025 emphasize the shift toward more predictable, contract-based revenue and lower exposure to Bitcoin price volatility. Execution is visible across 2024–2025 via municipal approvals (Nov 20, 2024), construction milestones (Nov 18, 2024), and early deliveries (Feb 26, 2025), suggesting a multi-quarter buildout that continues into 2026.

2: Deep dependency on CoreWeave—now without the tie-up

• Recent Evidence: The merger with CoreWeave was terminated on October 30, 2025 due to insufficient votes, leaving Core Scientific as a standalone public company. Q3 2025 showed significant capex funding from CoreWeave ($196.4 million of the quarter’s $244.5 million; Oct 24, 2025), and multiple 12-year hosting contracts remain a backbone of projected revenue (Feb 26, 2025). Management has stated an intention to diversify such that CoreWeave represents less than 50% of billable capacity by end of 2028 (May 7, 2025), but the current mix is still concentrated.
• Contextual Trends: The relationship intensified through 2024–early 2025 (e.g., $8.7 billion in 12-year contracted potential revenue as of Feb 26, 2025 earnings call; later increased to ~$10.2 billion on Feb 26, 2025 press release), culminating in a July 7, 2025 acquisition agreement that was not ultimately approved. The termination preserves independence but increases the importance of contract durability and customer diversification to mitigate counterparty risk.

3: Scale, power access, and delivery speed as potential cost/efficient-scale advantages

• Recent Evidence: By YE 2024, Core Scientific reported ~1.3 GW of powered infrastructure (Feb 26, 2025). The Denton site’s lease and power expansion to 394 MW (Nov 20, 2024), a 100 MW Muskogee facility groundbreaking (Nov 18, 2024), and the Auburn, AL site launch (Feb 19, 2025) expand a multi-state footprint geared to high-density AI workloads. Management also highlighted early delivery of a 16 MW tranche (Feb 26, 2025 earnings call) and ongoing power curtailment programs that demonstrate grid integration at scale.
• Contextual Trends: Convertible notes in December 2024/January 2025 provided additional funding capacity for buildouts (Dec 3, 2024; Jan 27, 2025). Despite rapid infrastructure growth, near-term financials show transition-related pressure (e.g., revenue declines and net losses in Q2/Q3 2025 on Aug 8 and Oct 24, 2025), suggesting the cost/scale advantage is still being translated into steady cash flows. Over time, if utilization, diversification, and operating efficiency improve, the footprint could support a more durable moat.