Phase Space AI

Valuation

Core Scientific [CORZ]

Core Scientific, Inc. [CORZ] — Valuation

As of: 2026-07-29 · Spot: $18.17 · EV (ex digital assets): $10,130m · Framework: v1.5.1 / criteria.md 2026-07-29 Cluster context: CLUSTER_Analysis.md

Two outputs are required on every name: a 12-month target and the implied-path test. Reporting only one is a defect. Both are addressed below, and where an input does not exist it is declared UNIDENTIFIED rather than substituted with a peer median.

---

Exit multiple: IDENTIFIED — the only one in the cluster

CORZ grew total revenue +108.8% YoY in Q2 2026. CoreWeave grew +111.6% YoY in Q1 2026. The comparator's growth brackets the subject's, so per valuation.md the anchor is valid and the multiple is IDENTIFIED at 6.6x EV / annualised AI revenue.

This matters beyond CORZ: it is the only name in the cluster where the growth-matched rule can actually be applied rather than declared unsatisfiable. And the result is a FAIL for the best business in the cluster — an asymmetry worth recording as a calibration observation in its own right. The rule bites hardest precisely where a comparator exists.

Note the additional propriety of this comparator: CoreWeave is CORZ's own largest customer. The anchor is not a loose sector proxy; it is the counterparty whose economics sit on the other side of the same contracts.

The implied-path test

Holding the anchor multiple, the share count and the net debt fixed:

• Required: colocation revenue of $1,535m annualised to support today's EV of $10,130m at 6.6x.
• Demonstrated: $546.7m annualised (Q2 2026 rate of $136.669m).
• Contracted upside: 395MW of 590MW is billing and produces $546.7m annualised. The remaining 195MW implies ≈$816m at the same revenue per MW → ≈$1,363m at full CoreWeave delivery.
• Margin (demonstrated − required): −$988m on today's run-rate, or −$172m even at full CoreWeave delivery. In percentage-point terms the price requires a 181% uplift and the contracted book delivers 149% — a −32pp margin.
• Implied compression: CORZ trades at 18.5x annualised AI revenue. Convergence to the 6.6x anchor with no revenue growth is −64% multiple compression.

Verdict: FAIL. The price requires materially more than the entire contracted book delivers, and the incremental argument — new customers at CoreWeave-like economics — is narrative, not evidenced. AMD is the one data point and its contract size is not disclosed; only the 30m-share warrant is quantified. criteria.md is explicit that narrative does not qualify for PASS WITH ARGUMENT.

Stated plainly: the best-converted, highest-margin, only-sequentially-growing business in this cluster FAILS its Valuation Criteria. That is not a contradiction — it is the whole point. CORZ has done the operational work; the price has already paid for more than the work delivers.

Sensitivity over the exit multiple — the highest-variance parameter

Terminal value exceeds 60% of EV, so per criteria.md this is the mandatory primary analysis and it is run over the exit multiple, never over scenario probabilities.

Read the top row. At the growth-matched anchor, CORZ must earn 13% more than its fully-delivered 590MW CoreWeave contract produces. At WYFI's multiple it needs 68% of the contracted book, which it will reach. The entire valuation question is which multiple survives, and that is precisely why valuation.md insists the sensitivity be run here and not over probabilities.

A cross-check that does not require a growth match

CORZ earns the same gross margin as WYFI on its contracted revenue and trades at 1.7x WYFI's multiple and 2.8x CoreWeave's. A defensible bull case exists — CORZ's colocation is arguably higher quality than CoreWeave's own economics, since CORZ is the landlord with power passed through rather than the operator carrying GPU depreciation. But that argument would justify a premium of some size, not a specific multiple, and valuation.md does not permit a multiple to be reverse-engineered from a qualitative preference.

12-month target: UNIDENTIFIED

Both required inputs fail:

1. No consensus. Alpha Vantage quota shared and not attempted. Consensus Criteria INDETERMINATE, blocks nothing.
2. The name's own multiple history is unusable. valuation.md requires the 12-month multiple be anchored on the name's own trading range with the percentile stated, and CORZ's history spans: emergence from Chapter 11 in January 2024, a business that went from 87% mining to 83% colocation in four quarters, five restated periodic filings, a stockholders' deficit, and a share count carrying 41.4% of warrant overhang whose fair value is the dominant line in the income statement. The historical multiple series is measuring a different company at every point.

Declared UNIDENTIFIED. No target is manufactured. Note that this is a case where a peer-median substitute would have been easy and superficially defensible — CoreWeave is right there, growth-matched, and is the customer. It is still not a 12-month target, because valuation.md establishes that a long-horizon multiple cannot be indexed to a 12-month horizon (tested 2026-07-29: best achievable correlation +0.08).

---

Why the mining leg and the AI leg cannot share an exit multiple

This is the methodological core of the cluster and it applies directly to CORZ:

A single blended exit multiple applied across these two streams is the exact defect that inflated ADMA by 51pp and INOD by 32pp. This analysis therefore ranks on EV / annualised AI-HPC revenue and EV / RPO — a price-for-what-exists and a price-for-what-is-signed — and scores the mining leg at UNIDENTIFIED with a floor at zero, because a stream with a negative gross margin at one operator and a halving margin at another has no defensible positive revenue multiple.

---

Required disclosures per valuation.md