Phase Space AI

Trade Construction

Core Scientific [CORZ]

Core Scientific, Inc. [CORZ] — Trade Construction

As of: 2026-07-29 · Spot: $18.17 (2026-07-29 close) · Framework: v1.5.1 / criteria.md 2026-07-29

This memo issues no position verdict — no Long, Short, Watchlist or Avoid. It scores Criteria and outputs an analysis. Whether this analysis justifies a position is a question about a particular book, and the book decides. What follows are the constraints any position would face, not a recommendation.

1. Liquidity Criteria (BINDING)

Value
Shares traded 2026-07-29 2,949,279
Dollar volume $53.58m
Verdict PASS

2026-07-29 was a high-volume down day across the entire complex, so this figure is generous relative to a normal session. Liquidity constrains position sizing, never admission (criteria.md).

2. Options — NO STRUCTURE PROPOSED

No options chain was pulled for CORZ. Liquidity Criteria is explicit: "Any proposed options structure requires the actual chain pulled first — open interest and quoted size for the specific strikes and expiry… A vehicle that cannot be filled is not a vehicle." The HCA precedent (maximum open interest of 18 contracts across an entire March-2027 chain) is why this is a hard rule rather than a preference.

Accordingly no strikes, expiries, Greeks or structures are proposed. Doing so without the chain would manufacture an uninvestable vehicle. Pulling the CORZ chain from data.alpaca.markets/v1beta1/options/snapshots is a prerequisite to any options expression.

3. Vehicle constraint if a position were taken

Common stock only, until a chain is pulled and open interest and quoted size are verified at specific strikes.

4. Sizing constraint — inverse volatility

criteria.md names inverse-volatility sizing as the active protection on the Downside Criteria, "because a fat-left-tail name is almost always a high-volatility name and is sized down automatically."

Value
252-day realised volatility (screen record) 64.7% (scan_v3 vol_252d_pct) — the lowest in the cluster, roughly 40% below the group
Implication among the highest in any book; inverse-vol sizing cuts this hard
One-day move, 2026-07-29 −12.4%

5. The correlation constraint that dominates everything else

On 2026-07-29 every name in this complex fell together:

HIVE HUT IREN CORZ BTDR WYFI SHAZ CRWV APLD NBIS WULF CIFR RIOT MARA CLSK
−11.1% −12.8% −13.5% −12.4% −13.1% −12.7% −10.9% −9.6% −12.7% −13.7% −12.8% −13.3% −14.0% −11.6% −10.8%

Fifteen names, one factor, a 4.4pp range. Any two positions in this cluster are effectively one position. This is a book-level exposure and correlation-limit question, not a single-name question, and it must be resolved by portfolio-book before any CORZ position is sized. Treating cluster members as independent names would breach the correlation limit while appearing diversified.

6. Invalidation — the observable that would refute the thesis

Confirming observable: the remaining ~195MW of the 590MW CoreWeave contract commencing billing on schedule, at revenue per MW consistent with the ≈$1.38m/MW/year the current 395MW implies. That is ≈$816m of incremental annualised revenue and it is the difference between $546.7m and ≈$1,363m. Second: a named, quantified second customer. AMD is named but its contract size is undisclosed, and the FAIL verdict rests specifically on the absence of evidenced incremental demand.

Refuting observable, in priority order: 1. Any CoreWeave renegotiation, deferral or termination. 83.2% of revenue and the SPV financing both sit on that one counterparty. CORZ has already booked a $41.9m loss on contract termination in Q2 2026, so this is a demonstrated risk, not a hypothetical. 2. Colocation gross margin falling below ~50%. It is 59%; the thesis depends on power staying a pass-through. 3. A sixth restated filing, or disclosure of a material weakness. Five periodic filings were amended in sixteen days and the cause is unexamined. 4. Colocation revenue failing to grow sequentially. CORZ is the only name in the cluster growing sequentially and that is the core of its Quality PASS.

7. What must be resolved before any position

  1. Investigate the restatements. A 10-K/A for FY2024, a 10-K/A for FY2025 and 10-Q/As for all three 2025 quarters, filed 2026-03-02 and 2026-03-18, in a company whose reported loss is dominated by warrant fair-value accounting and which reports a stockholders' deficit. This is the largest unexamined risk in the file and it must be resolved before any position, not after.
  2. Obtain the AMD contract's size and term. The Valuation FAIL rests on the incremental demand argument being narrative. AMD is the one candidate to make it evidenced, and only its warrant is quantified.
  3. Obtain the CoreWeave billing schedule for the remaining 195MW. No dated schedule is disclosed; the entire required-path analysis turns on it.
  4. Understand the SPV financing structure. Certain CoreWeave licences were assigned to it and debt rose $3,238m in six months. Recourse, covenants and what happens on a CoreWeave credit event are all unknown here.
  5. Establish the site-level power contracts behind the pass-through — tenor and indexation are not disclosed. The pass-through protects the margin only for as long as the contracts permit it.
  6. Pull the options chain before contemplating any non-linear expression.
  7. Resolve the cluster correlation limit with portfolio-book. CORZ has the lowest realised volatility in the cluster (64.7% vs 97–127% elsewhere), so inverse-volatility sizing will allocate it the largest position of the seven — which makes the correlation question sharper, not softer.