Spare parts supply chains in asset-intensive industries are highly exposed to disruption because demand is intermittent, suppliers are distant, and lead times are long. Integrating on-site additive manufacturing alongside conventional suppliers is increasingly viewed as a proactive resilience capability, yet existing sourcing models rarely capture the process-specific economics of metal additive manufacturing. This paper develops a mixed-integer linear programming model for a hedged spare parts supply chain in which conventional suppliers and an on-site bound metal deposition facility are jointly optimized. The model embeds the full bound metal deposition cost structure, covering design, feedstock procurement and inventory, printing, debinding, sintering, post-processing, and part qualification, together with supplier capacities, minimum order quantities, and multimodal transportation. A twelve-month numerical study of three industrial spare parts shows the hedged design meets all demand at minimum cost, qualifies every part for additive production, deploys additive capacity exactly where conventional sourcing binds, and retains substantial reserve capacity for disruption response. A demand sensitivity analysis reveals the demand thresholds at which the additive share expands, clarifying the conditions under which the capability is most valuable. The study is thus positioned as a decision-support demonstration for spare parts sourcing, offering managerial guidance on when embedding on-site AM capacity constitutes a cost-effective resilience investment.