<p>The global aquaculture industry is increasingly constrained by resource limitations and sustainability challenges, making the transition to low-carbon digital practices imperative. Although artificial intelligence (AI) presents significant potential for enabling intelligent, low-carbon aquaculture, its adoption remains hindered by uncertainties surrounding its efficiency, cost-effectiveness, and environmental impact. This study examines the decision-making processes of aquaculture enterprises and technology providers during the industry’s transition to intelligent systems in a low-carbon context. Utilizing a differential game model, we compare three core scenarios—traditional aquaculture, basic energy-saving retrofits, and AI-driven intelligent aquaculture—to examine alternative low-carbon transition strategies. We further extend the analysis to include a fourth scenario incorporating government subsidies, in order to assess their amplifying effect. Equilibrium strategies and stakeholder outcomes are evaluated through both theoretical analysis and case-based simulations. Our findings challenge the assumption that AI guarantees improved performance because its implementation may, in some cases, increase the complexity and energy consumption. However, in most scenarios, AI significantly enhances operational efficiency and environmental outcomes, particularly when faced with uncertainties. Furthermore, government subsidies play a critical role in aligning incentives, mitigating externalities, and accelerating the industry’s transition to low-carbon practices. These insights provide valuable guidance for policymakers and industry stakeholders seeking to foster sustainable and efficient low-carbon transitions in aquaculture.</p>