The industrial internet is an important strategic deployment to gain a competitive advantage in the new round of the industrial revolution. As stated in the Outline of the 14th Five-Year Plan (2021-2025) for National Economic and Social Development, it is necessary to promote the deep integration of the internet and the manufacturing industry, accelerate the construction of the “5G+Industrial Internet” platform project, and realize the digital transformation of the manufacturing industry. Unlike the consumer-oriented internet which focuses on demand-side data integration, the industrial internet highlights the data integration of the entire industrial chain and supply chain, and provides accurate data support for the entire value chain of the product lifecycle, thus giving a practical and effective way for the digital transformation of the manufacturing industry. This issue has attracted extensive attention and research from scholars in related fields. The previous research findings have explored the role and impact at different stages of the integrated development of industrial internet platforms with manufacturing industries, as well as the development strategies of platforms in the time dimension based on maximizing their own benefits. The underlying assumption is that industrial internet platforms are absolutely value-neutral and that value creation spillovers are universal and equitable. However, it overlooks the objective fact that most small and medium-sized manufacturing participants have their professional production data homogeneously integrated by the platform and are deeply tied up, thereby losing their autonomy, increasing their dependence and being in a vulnerable position in the supply chain. Existing research on manufacturing dependence mainly focuses on the strategy choice and optimization of the dominant party under resource dependence theory, with limited impact on the relatively weaker party in the relationship between platform and manufacturing participants. In this context, it is more appropriate to focus on the evolution of the relationship between participants as the weaker party and the platform based on the power dependence theory.In light of this, this paper considers the impact of heterogeneous features at different stages in the evolution of industrial internet platforms. On a basis of resource and power dependence theories, this paper defines the development stages of industrial internet platforms, constructs a model of the evolution of dependence between the two in the time dimension, and divides the evolutionary process of dependence between the two in the time dimension into three scenarios: complementary dependence based on digital transformation needs, asymmetric power dependence based on data services, and endogenous joint dependence based on value co-creation.Subsequently, based on the dynamic evolutionary characteristics of the dependence of the two, we construct a multi-stage single-cycle supply chain game model involving an industrial internet platform and multiple independent participants. With the objective of maximizing benefits, the paper finds out the optimal investment decisions and returns of the participants, analyses the key influencing factors of the optimal decisions, and compares the returns at the above-mentioned different stages. Further, the paper explores the impact of changes in parameters, such as platform data service prices, participant dependence coefficients, data breach risk perception and digital regulatory changes, on participants’ decisions and benefits through numerical simulations, and discusses the price range as well as effective range of dependence for manufacturing participants to establish businesses on the internet platforms.The results are as follows. During the platform growth period, whether manufacturing participants join the industrial internet platform and their decision equilibrium are influenced by the price of the platform’s data services, with an acceptable price range for participants. In the platform maturity period, ideally, participant returns show a U-shaped trend of decreasing and then increasing as the dependence coefficient increases. However, due to the asymmetrically inferior dependence relationship in practice, the cost of perceived loss of data risk rises, resulting in participant returns surviving only within the decreasing interval of the dependence coefficient. Participants are less sensitive to the price of platform data services during the maturity period compared to the growth period. Meanwhile, process-oriented manufacturing participants exhibit high data service dependence and low price sensitivity, while discrete manufacturing participants show low dependence and high price sensitivity. Besides, digital regulatory governance had a significant impact on the increase in acceptable prices and dependence intervals for different types of manufacturing participants.