Abstract
<title>Abstract</title> <p> This study investigates the effect of broadband infrastructure and artificial intelligence (AI) adoption on labour productivity across the EU-27 from 2019 to 2025. Grounded in Romer (1990) endogenous growth theory and Rogers (2003) diffusion of innovations theory, the study employs a two-way fixed-effects panel regression stratified by initial productivity levels. The results indicate that fibre-to-the-premises (FTTP) penetration consistently drives productivity across all Member States, confirming its role as a foundational enabler. However, AI adoption presents a “productivity paradox”. In lower-income states, the standalone effect of AI is significantly negative, likely due to high implementation costs and structural disruption. Crucially, the interaction between broadband and AI reveals a significant complementary effect for these states; broadband infrastructure acts as a moderator that mitigates AI’s negative effects and facilitates productivity gains. In contrast, high-income states exhibit “infrastructure saturation”, where the interaction between broadband and AI is not statistically significant. These findings suggest that while broadband is universally beneficial, capturing the benefits of AI in catch-up economies depends critically on infrastructure readiness. Policy should therefore prioritise integrated digital strategies to prevent economic divergence within the Union. <italic> <bold>JEL-Classification</bold> </italic> : C23, L96, O32, O33, O47 </p>