ASC 350-30 × EU AI Act Art. 53(1)(d) — One-Page Map

Data license classification in the AI training data supply chain. One inventory. Two obligations.

Prepared for Cohere. The ASC 350-30 classification and the Art. 53(1)(d) training-content disclosure are the same exercise, done once. This page maps the treatment by license type and applies it to your three most recent data events.

PREPARED FOR COHERE WWW.QUANTIFIABLE.ORG OCTOBER 2026
01 — The Rule

One question determines the classification

Framing

Under ASC 350-30, the classification of a licensed data asset turns on one question: does the license grant a finite term or a perpetual right? Term-based is expensed over the term. Perpetual capitalizes as a separate intangible. The ratio across the portfolio determines how much of the data spend is a COGS line versus a balance-sheet asset.

License Type Accounting Treatment P&L / Balance Sheet
Term-based12-mo, 3-yr, renewable Expensed over the license term COGS
PerpetualNo expiry, or renewal at nominal cost Capitalized as identifiable intangible asset BALANCE SHEET
Acquired via M&AIP assumed with the company Capitalized at fair value as of acquisition date BALANCE SHEET
Unlicensed / disputedNo executed agreement No asset recognized. Legal liability if applicable. OFF-BS RISK
Alternative future use

The "alternative future use" test (ASC 350-30-25): if purchased data can be used beyond the immediate training run (e.g., fine-tuning, RAG, product delivery), it qualifies for capitalization even if the license is term-based. In practice, most AI training data has alternative future use — which means the term/perpetual distinction is the only variable that determines the classification.

Perpetual ≠ indefinite-lived

A perpetual license may still be finite-lived if technological obsolescence or regulatory change limits its useful life. The useful life determination must be reassessed each reporting period.

02 — The Map

Common license types in the AI training data supply chain

Source Category Typical Structure ASC 350-30 Art. 53(1)(d)
Financial data APIsCapital IQ, FactSet Term-based, annual renewal, per-seat or per-API-call Expensed over term. If alt. future use met → capitalize. Disclose as "licensed data source" with category, size, date.
Proprietary domain datasetsAcquired via M&A Perpetual IP, assumed with the company Capitalized at fair value. Indefinite-lived if no foreseeable expiry. Disclose as "licensed/acquired data" with category and size.
Publisher / media contentNews, magazines Term or perpetual, negotiated per outlet. Often disputed. Term → expense. Perpetual → capitalize. No license → no asset. Disclose category and rights-reservation compliance (Art. 4(3) DSM).
Web-crawled / scrapedPublic web No license (or TDM opt-out under Art. 4(3) DSM) No asset. Compliance cost only. Disclose crawler identifiers, purposes, content types, top 10% domains.
Government / sovereignPublic sector Perpetual or term, jurisdiction-specific Capitalize if perpetual. Expense if term. Disclose as "publicly available" or "licensed."
Synthetic / user-generatedInternal or user-provided Internally generated or user-licensed No capitalization (ASC 350-30-25-3). Disclose as "synthetic" or "user-provided" category.
03 — Applied to Cohere

Four data events, four open questions

Data Event Source License Type Status Open Question
S&P Global collabJune 8, 2026 Capital IQ Financials, earnings call transcripts → North Undisclosed. No term, renewal, or perpetual language in press release. UNRESOLVED 12-month renewable term or perpetual license with annual fee?
Reliant AIMay 19, 2026 Proprietary biomedical datasets (GSK, Kyowa Kirin, Medicus Pharma) Perpetual IP (assumed via M&A) CAPITALIZED What was the allocated fair value in the PPA?
Publisher agreementsNews/Media Alliance, Feb 2025 News articles, magazine content Disputed. Suit alleges unauthorised use. Settlement terms not public. UNRESOLVED What are the settlement/licensing terms?
Aleph AlphaMerger signed Sept 16, 2026 European training data, government/defense datasets, sovereign-AI domain data Unknown. Pre-close. PRE-CLOSE What is the license structure of each dataset?
04 — The Art. 53(1)(d) Intersection

The same inventory, built once

The EU AI Act Article 53(1)(d) requires every GPAI model provider to publish a sufficiently detailed summary of training content using the mandatory AI Office template (published July 24, 2025; obligation in force since August 2, 2025).

Template requires
  • → Data source categories
  • → Indicative sizes, modalities, languages
  • → Major datasets with identifiers (>3%)
  • → Crawler specs, top 10% domains
  • → Rights-reservation compliance
ASC 350-30 requires
  • → Source name and category
  • → License type (term / perpetual)
  • → Term and renewal terms
  • → Indicative size and modality
  • → Acquisition date

The key insight: every data source you document for the Art. 53 template (category, license type, size, acquisition date) is the same data you need to determine whether it's a COGS line or a balance-sheet asset. Build the register once. Feed both obligations.

Timing

The AI Office's enforcement powers (including fines) began August 2, 2026 — already in effect. Models placed on the EU market before August 2, 2025 have until August 2, 2027. For the combined Cohere–Aleph Alpha entity, the Art. 53 disclosure will need to cover both model portfolios at close.

05 — What This Means for the Data Team

Four actions, in priority order

  1. Build the inventory once. A single data-source register that captures: source name, license type (term / perpetual / acquired / unlicensed), term (if applicable), renewal terms, indicative size, modality, and acquisition date. This register feeds both the ASC 350-30 classification and the Art. 53(1)(d) template.
  2. Resolve the S&P Global question first. It's the highest-profile data partnership in North, and the classification is currently unresolved. A 12-month renewable term and a perpetual license produce materially different balance-sheet outcomes.
  3. Flag the publisher agreements. If the News/Media Alliance matter resolves as a term license, the historical "unauthorised use" period creates a retrospective classification question — was the data ever an asset, or was it always a liability?
  4. Pre-close Aleph Alpha data audit. The purchase price allocation at close will require a fair-value allocation across Aleph Alpha's data assets. The license structure of each dataset determines whether it's a finite-lived intangible (amortized), an indefinite-lived intangible (impairment-tested), or expensed. Doing this audit pre-close avoids a measurement-period scramble.