PurposeThe Physical Internet (PI) application in a supply chain is explored by automakers to achieve a digital supply chain to challenge timely delivery while maintaining high customised production at the lowest operating cost.Design/methodology/approachA bi-objective mixed integer model is formulated, where production is performed in multistage manufacturing systems (MMS) and then delivered in a two-level distribution system. Next, a hybrid iterative method algorithm is developed to solve the practical-scale problem within an admissible time. Finally, PI's benefits on production and supply chain operation are discussed through extensive computational experiments in different supply chain configurations.FindingsThree significant findings are obtained. First, PI can achieve a comparable or better service level, while the cost is always lower. Second, PI can improve the utilisation of production and transportation resources. Third, with a more complex supply chain and a higher production cost or truck fixed cost, PI's advantages over traditional supply chain become more vigorous, but the increase in orders will weaken it.Practical implicationsThe auto enterprise should adopt a PI-enabled supply chain (PI-SC), especially with the increase of network complexity and specific cost factors.Social implicationsImportance should be attached to the PI-SC to make customers better involved in the supply chain.Originality/valueFirst, the application of PI in the existing plant is described. Second, MMS production with multi-mode transportation is jointly scheduled. Third, the decision support of the PI-SC is provided for auto enterprises.