Game-theoretic analysis of low-carbon technology innovation and AI-augmented service pricing in supply chains

Abstract

The first study is about advancing low-carbon photovoltaic technology. Chinese photovoltaic manufacturers have solidified their global leadership in solar technology, with polysilicon serving as a critical material for solar panels. Industry pioneer GCL-Poly has achieved a breakthrough by commercializing granular silicon production via the fluidized bed reactor (FBR) deposition method, reducing carbon emissions by 74% compared to the traditional Siemens process. In the first study, we employ game-theoretic models to systematically examine innovation and cooperation dynamics in the supply chain. Under monopolistic market structures, we analyze three innovation modes with a low-carbon supplier monopoly: manufacturer's innovation, supplier's innovation, and joint innovation, along with a benchmark high-carbon monopoly. Results demonstrate that joint innovation yields win-win outcomes in low-carbon monopolies. In competitive settings (e.g., one manufacturer and two suppliers), the manufacturer' material selection depends on granular silicon's cost: low costs promote uptake, whereas high costs sustain demand for conventional silicon rods. Achieving a triple-win for the manufacturer, high-carbon supplier, and low-carbon supplier remains difficult. The high-carbon supplier suffers profit erosion under joint innovation but gains competitive advantages under manufacturer's innovation. The second study is about artificial intelligence (AI)-augmented service pricing. We examine service pricing in a two-tier supply chain where firms outsource to an AI-augmented supplier. The supplier's AI adoption creates a learning and scalability effect that improves operational efficiency and achieves near-zero marginal cost scalability, and firms demand cost-sharing arrangements. We analyze both centralized and decentralized configurations (single firm-supplier and duopoly). In the centralized case, AI's learning and scalability effect reduces optimal market prices and increases the firm's maximal profit. For the decentralized single firm-supplier setting, this effect boosts profits for both parties, with the firm benefiting disproportionately. Also, this decentralized system remains less profitable than the centralized one. While AI enhances operational efficiency, its growing influence does not fully mitigate double marginalization even as the learning and scalability effect strengthens. For the decentralized setting with two firms, we derive equilibrium solutions under both first-come-first-served (FCFS) and priority service disciplines. We find that, with identical firms, duopolies outperform monopolies when market size falls below a critical threshold. Additionally, while AI's scalability significantly reduces congestion, it does not eliminate the supplier's ability to leverage priority pricing--in fact, it amplifies the supplier's profitability from offering priority service.