← Back to article

Editorial review · 260614-008

How FLUX’s piece on The depreciation clock Goldman just started scored.

Read the article →
86/100
Solid

Solid reporting. Some issues but credible overall. The reader is well-served.

Accuracy 87
Balance 85

Accuracy

Core claims (7pp ROE drag, $770bn aggregate capex, $320bn hyperscaler capex, 44% vs 22% ROE, depreciation rising from 7% to 12% of revenue) are attributed to specific Goldman research and secondary outlets, which I treat as post-cutoff source-attributed. The Apple inclusion in the 'five core hyperscalers' list is non-standard and inconsistent with the article's own glossary (-5 minor unsourced framing). One minor deduction for the 80%-to-20% correlation figure offered without a date range or methodology (-3).

Balance

The piece is opinionated but engages Goldman's own counter-thesis (the 1.5pp EPS productivity tailwind) and names the framing tension rather than strawmanning it. Source diversity is thin: every load-bearing number traces to Goldman, with Kiplinger as the only external voice, which is acceptable for a deal-note format but worth flagging (-8). Tone stays within FLUX's neutral-analytical baseline and the bear and bull readings are both given air.

Concerns (3)

Reproducibility

Run
14 Jun 2026, 12:49 BST
Reviewer
claude-opus-4-7
Prompt SHA
48c20c719fc8
Article SHA
f3c190ab5d5a
Editor
FLUX
Published
14 June 2026
Cost
$0.0000

How this review works: read the methodology. Each published Dispatch is scored by a single primary reviewer (Claude Opus 4.7) against the public rubric. A second model (Gemini 2.5 Pro with Google Search) runs the same prompt as a variance signal and is shown above only when the two scores diverge by more than ten points.