Snapshot
- Ticker: CORZ (Nasdaq)
- Bucket: Data-Center Operators (HPC Pivots)
- Q4 2025 fund position: $418.7 M common (28.76 M sh) — 7.59 % of 13F, #6 holding
- Shares outstanding (9/30/25): 308.4 M
- HQ: Austin, TX
Business Overview
Core Scientific operates large-scale data-center campuses across Texas (Denton, Dallas), Oklahoma, Georgia, North Dakota, and North Carolina. The portfolio includes ~350 MW currently energized with a contracted build-out path to >2 GW driven by the CoreWeave colocation contracts. Three operating segments are visible in the Q3 2025 release:
- Self-mining — operates Bitcoin ASICs at company-owned sites; the largest revenue contributor still as of Q3 2025.
- Hosted mining — runs third-party miner hardware under capacity agreements; declining as colocation HPC absorbs site capacity.
- HPC colocation — leases site infrastructure (power, building, cooling) to CoreWeave under multi-year take-or-pay contracts; the strategic growth segment.
The company emerged from Chapter 11 bankruptcy in January 2024 with a clean balance sheet, ~$300+ M in NOL carryforwards, and ownership of physical assets that the AI buildout subsequently revalued sharply. The 2024 CoreWeave hosting contracts moved Core Scientific from a distressed miner into a thinly-disguised data-center developer — and prompted CoreWeave’s 2025 attempt to acquire CORZ outright, which failed at the stockholder vote later in 2025. The hosting contracts remain in force despite the failed merger.
Financial Trajectory
| Metric | FY2024 | FY2025 9-mo | Q3 2025 | Q3 2024 |
|---|---|---|---|---|
| Revenue | $510 M | $239.3 M | $81.1 M | $95.4 M |
| YoY Δ | — | — | -15 % | — |
| Self-mining | — | — | $57.4 M | — |
| Hosted | — | — | $8.7 M | — |
| HPC colocation | — | — | $15.0 M (+45 %) | — |
| Gross margin | — | — | ~5 % | ~0 % |
| Adj. EBITDA | — | — | −$2.4 M | +$10.1 M |
| Net loss | — | — | −$146.7 M | — |
| Net loss YTD | — | −$502.8 M | — | — |
The Q3 2025 print captures the transition mid-flight: BTC mining contribution shrank with the post-halving economics and partial site repurposing, while HPC colocation revenue grew 45 % YoY but is still less than 20 % of total revenue. The path to “majority HPC” runs through the CoreWeave contract ramp (see below) rather than current-quarter financials.
Balance Sheet (9/30/25)
| Item | $M |
|---|---|
| Cash | 453.4 |
| BTC holdings (mark) | 241.4 |
| Total liquidity | 694.8 |
| Total debt (convertibles) | ~1,060 |
| Shares outstanding | 308.4 M |
The company has been an active issuer of convertible notes — both at emergence in January 2024 and again in 2025 — to fund the CoreWeave build-out. Liquidity is comfortable but the convertibles dilute meaningfully if-converted.
Key Operational KPIs
- Energized capacity: ~350 MW
- Contracted build-out to CoreWeave: ~590 MW by early 2027
- Hashrate / miner counts: Core Scientific stopped monthly production updates in mid-2025 after the CoreWeave merger process began.
- Sites: Denton TX, Dalton GA, Calvert City KY, Cottonwood TX, Pecos TX, Marble ND, others
CoreWeave Hosting Contracts (the central asset)
The CoreWeave contracts are the reason Core Scientific is the bucket’s anchor:
- June 2024: initial 200 MW colocation contract for ~$3.5 B over 12 years across two sites (Denton TX and Cottonwood TX), with options for expansion.
- August 2024: expansion adding ~118 MW for ~$2.0 B.
- By end-2024 / early 2025: further expansions brought the cumulative committed capacity to ~590 MW with aggregate contracted revenue >$8.7 B over 12 years.
- 2025: CoreWeave tendered to acquire Core Scientific outright in a stock deal; the merger failed at the CORZ stockholder vote, but the underlying hosting contracts remain in place. The vote outcome turned on whether shareholders preferred the standalone capture of operating margin (which the leases do not include) versus full integration into CoreWeave’s stack.
The economic value of the contracts at full ramp implies a re-rating from miner multiples (~5–10× EBITDA in good years, much less in bad) to data-center multiples (15–25× EBITDA on stable contracted revenue) on the converted MW base.
Why It Fits the Thesis
CORZ is the anchor of the data-centers bucket and arguably the model the fund expects other miners to follow. The CoreWeave contracts demonstrate the re-rating mechanism: stranded power capacity, valued at miner multiples, gets repriced at data-center multiples once contracted to an AI tenant. The fund grew the position from ~6 M sh on entry to ~28.8 M sh by Q4 2025 — a ~5× size increase that signals it as the highest-conviction name in the bucket.
Position History in the Fund
| Quarter | Position |
|---|---|
| Q1 2025 | New, ~6.7 M sh |
| Q2 2025 | Increased |
| Q3 2025 | 20.18 M sh |
| Q4 2025 | 28.76 M sh |
Risks
- CoreWeave counterparty risk — single customer for the entire HPC growth thesis.
- HPC retrofit execution. Mining sites typically lack tier-III/IV uptime, redundant fiber, and liquid cooling. Conversion costs and timelines have slipped at peers.
- Bitcoin price downside in the residual mining segment, which still produces meaningful revenue and EBITDA.
- Convertible dilution if equity rallies.
- Power-contract renegotiation risk if existing PPAs were written for mining-only use.
- Failed CoreWeave merger signals shareholder skepticism about take-out optionality at current prices.