Standards-anchored reference to primary IFRS and FASB text. Not accounting advice.
AI Capitalisation
Illustrative scenario

A proprietary data-labelling pipeline

Head note

An entity builds an in-house annotation pipeline to label data for its models. This scenario separates two assets: the pipeline software itself, which can be internal-use software or a development-phase intangible, and the labels it produces, which follow the build they serve. Every figure here is an illustrative example, not client data.

1

Two assets, assessed separately

The pipeline software is assessed as internal-use software under ASC 350-40 or a development-phase intangible under IAS 38: its scoped, funded, feasible build can be capitalised ASC 350-40350-40-25. The labels the pipeline outputs are a cost of preparing whichever model they train, and follow that model's phase or stage.

2

What expenses

Exploratory work deciding how to build the pipeline is preliminary or research and is expensed. Running and maintaining the pipeline once live is a period cost, as is labelling done to test unproven use cases IAS 38§54.

3
Illustrative outcomeIllustrative example, not client data

The pipeline build, once funded and feasible, is capitalised as internal-use software and amortised over its useful life. Labels produced for a qualifying model build are capitalised into that model; labels produced to test an unproven idea are expensed. Ongoing pipeline operation is expensed. All figures are illustrative.

Sources of record
  1. S1IAS 38 Intangible Assets, IFRS Foundation (IFRS). https://www.ifrs.org/issued-standards/list-of-standards/ias-38-intangible-assets/
  2. S2Handbook: Software and website costs (ASC 350-40 internal-use software), KPMG (US GAAP). https://kpmg.com/us/en/frv/reference-library/2026/handbook-software-website-costs.html
Standards-anchored reference · not accounting adviceRevised 2026-07-24