Significant development uncertainty

What counts as significant uncertainty

The ASU names two factors, either of which is enough: the software has technological innovations or novel, unique or unproven functions or features whose uncertainty has not been resolved through coding and testing, or its significant performance requirements have not been identified or are still being substantially revised. An unproven model architecture, or a model whose target behaviour is still being redefined, fits that description. It is the US analogue of the IAS 38 technical-feasibility criterion under IFRS.

Why committed funding does not override it

Funding shows intent and resources; it does not show the model will work. Both the probable-to-complete threshold and this uncertainty gate must be cleared. A well-funded moonshot is still expensed until feasibility is demonstrable.