General-purpose AI and FinTech require substantial complementary intangible investments to develop digital business assets, yet these remain undervalued in national accounting despite their economic significance. This study presents a theoretical model explaining how the initial underrecognition of digital assets leads to productivity mismeasurement, creating a J-curve effect: early adoption phases show depressed productivity growth as investments accumulate, followed by overestimation as benefits materialize. The model provides a framework for assessing intangible assets (R&D, software, hardware, FinTech), with empirical results confirming strong J-curve patterns for software (15.9% valuation gap in 2023) and FinTech, but weaker effects for hardware. By adjusting for intangible capital, our estimates exceed official productivity figures by 11.3% (2024) and 15.9% (2023). While AI’s current productivity contribution remains modest, the analysis reveals a consistent growth trajectory, highlighting its evolving role as a digital business asset. These findings resolve key aspects of the Solow paradox by demonstrating how unmeasured intangibles distort productivity statistics during technological transitions. The J-curve framework offers policymakers and analysts a tool to correct systemic undervaluation of digital investments in national accounts.