The useful life of compute.

The question behind the study

An accounting depreciation schedule is not a forecast of economic usefulness. This proposed study would examine the relationship between hardware performance, realised compute prices and replacement expenditure. It would seek comparable observations across hardware generations rather than infer a useful life from a single vendor claim.

Proposed analytical approach

• Assembly of dated hardware and compute-price observations with explicit treatment of configuration, availability and sample selection.
• Statistical estimation of price and performance relationships where the observations support it. Causal claims would require a suitable identification design.
• Asset cash flows under alternative utilisation, pricing and replacement assumptions, with uncertainty intervals and sensitivity testing.

Intended outputs

• A documented panel dataset and reproducible analytical code.
• A hardware-life and replacement-capex model.
• An empirical methods paper and scenario report.

Decision relevance

Support for testing useful-life assumptions in compute-intensive businesses. The analytical distinction is between accounting treatment, technical capability and cash generation.

Evidence requirements and limits

Quoted prices can differ from realised prices, and public samples may not represent private enterprise contracts. Results would be bounded by the coverage and comparability of the collected observations.