The Capital Intensity of Intelligence
Capital formation, depreciation schedules and the balance-sheet treatment of AI infrastructure.
Abstract
The construction of AI infrastructure represents one of the largest concentrated capital formation episodes in modern industrial history. This brief examines how that capital is being raised, held and depreciated, and what the accounting treatment implies for the reported economics of the firms undertaking it. We document a steady migration of capacity off balance sheet, a lengthening of assumed useful life for accelerator hardware, and a growing divergence between reported capital expenditure and physical capacity deployment. We argue that analysis relying on disclosed capex alone will increasingly understate deployment, and set out the alternative measurement approach we use.
Key findings
- 01The share of AI infrastructure capacity held off balance sheet rose from 19 per cent in 2023 to an estimated 42 per cent in 2026.
- 02Assumed useful life for accelerator hardware has been extended by the majority of large operators, in several cases from three years to six.
- 03A one-year extension of assumed useful life across the sector moves reported operating margin by a margin that is material to reported profitability.
- 04Reported capital expenditure has become a progressively weaker proxy for physical deployment.
How an asset is financed and how quickly it is written down are accounting choices. At current scale, those choices move reported profitability by amounts that are not small.
Two developments in the financial treatment of AI infrastructure deserve more attention than they receive. The first is the migration of capacity off balance sheet. The second is the extension of assumed useful life for accelerator hardware. Both are defensible individually. Together they make the published financial record a progressively weaker guide to physical reality.
Off balance sheet
In 2023, most large operators owned both the facility and the hardware inside it, and both appeared on the balance sheet. By 2026, a substantial share of capacity is held through leases, colocation agreements with embedded capacity commitments, and special-purpose vehicles in which the operator holds a minority economic interest alongside financial partners.
- Owned, on balance sheet58
- Leased facility, owned equipment21
- Colocation with commitment13
- Special-purpose vehicle8
Frontier Labs analysis of filings and infrastructure dataset.
The economic substance in many of these arrangements is close to ownership: long tenors, capacity commitments that approximate the asset's useful life, and residual value risk borne by the operator. The accounting treatment differs from ownership nonetheless, and the difference flows through to reported leverage, capital efficiency and return metrics.
Useful life
Accelerator hardware was widely depreciated over three years in 2022. Several large operators now use five or six. The stated justification — that hardware remains economically productive for inference well after it is displaced from frontier training — is empirically reasonable. Older accelerators do continue to earn revenue.
The effect on reported figures is nonetheless large. Depreciation is a substantial share of the cost base for these operations. Extending useful life reduces annual depreciation, raises reported operating income, and does so without any change in cash flow or in physical reality. When several large firms make the same change in the same period, the aggregate effect on sector-level reported profitability is material and is not always foregrounded in commentary.
None of this is an allegation of impropriety. Lease accounting and useful-life estimation are ordinary areas of management judgement, and the disclosures are in the filings for anyone who reads them. The point is narrower: the headline numbers that circulate in public discussion of this sector's economics are increasingly sensitive to accounting choices that most readers of those numbers are unaware of.
Sources and references
- [1]Annual and quarterly filings, 18 operators, 2022–2026.
- [2]Accounting policy disclosures on useful life assumptions.
- [3]Frontier Labs infrastructure dataset, release 2026.2.
Source assessments, provenance records and the full data appendix are described in the methodology. Corrections to this publication are recorded on the corrections page.
Keywords
- capital
- depreciation
- capex
- accounting
- disclosure
Full edition
Read the complete report.
The full edition is released to named readers. Tell us who you are and it opens straight away.
- Extent
- 28 pages
- Price
- Free
Cite this publication
Economics Desk (2026). The Capital Intensity of Intelligence. Frontier Labs. https://doi.org/10.5281/frontierlabs.2026.0003
Research infrastructure powered by AGI-S.