How Do Entrepreneurship and AI-Driven Technologies Enhance Organizational Resilience? Evidence from Chinese Enterprises

Abstract

This study investigates how entrepreneurship enhances organizational resilience through AI-driven technological innovation under dynamic environments. Using panel data from 3975 Chinese listed firms (2013–2022) and a moderated mediation regression framework, we examine both direct and indirect effects. AI-driven innovation is operationalized through digital patent text analysis, capturing firms’ technological advancement. The results show that entrepreneurship positively affects organizational resilience (p < 0.01) and promotes AI-driven innovation (p < 0.01). AI-driven innovation further mediates the relationship between entrepreneurship and resilience (p < 0.05). Environmental dynamics strengthen the effects of digital innovation on resilience (p < 0.05) and the mediated pathway. These findings provide empirical evidence on how entrepreneurial capabilities and digital innovation jointly support organizational resilience and sustainable development.

1. Introduction

In recent years, enterprises have faced crises such as COVID-19, making adaptation essential for survival. Organizational resilience, as a core dynamic capability of firms [1], helps organizations cope with uncertainty and sustain long-term development [2]. Grounded in dynamic capability theory and upper echelons theory, this study explores how entrepreneurship and digital technological innovation enhance organizational resilience in dynamic environments [3]. However, most existing studies treat resilience as an outcome rather than clarifying how entrepreneurial cognition and digital transformation shape its development process. Therefore, this study investigates the mechanism linking entrepreneurship, AI-driven innovation, and resilience, while considering the moderating role of environmental dynamism.

Firms that lead their industries often have entrepreneurs who demonstrate strong entrepreneurship and key competencies such as opportunity recognition and risk management in dynamic environments [4]. Such managers can sense market changes, respond rapidly, and meet diverse demands [5]; enabling firms to achieve superior performance and maintain competitive advantage [6]. In the Chinese context, entrepreneurship helps leaders identify risks and opportunities and make timely strategic decisions that enhance adaptability and resilience. However, prior research has mainly focused on entrepreneurship’s direct impact on firm performance, while insufficient attention has been given to how entrepreneurial behavior promotes organizational resilience through technological innovation.

Entrepreneurship is widely regarded as a multidimensional construct including innovativeness, risk-taking, opportunity recognition, and entrepreneurial talent, reflecting essential behavioral and cognitive attributes in dynamic environments. Building on the existing literature, this study adopts a four-dimensional framework and measures entrepreneurship using the entropy method [7]. Similarly, organizational resilience is defined as a firm’s ability to anticipate, absorb, and adapt to shocks, reflecting robustness and adaptability [8]. Linking these concepts theoretically, entrepreneurial cognition influences resource allocation toward innovative technologies, while resilience emerges when those technologies reinforce sensing, seizing, and reconfiguring capabilities. This conceptual framing clarifies the relationship between entrepreneurship and resilience and provides a consistent theoretical basis for empirical examination, while improving alignment with international research standards.

Externally, inter-firm collaboration and industry-chain integration facilitate knowledge sharing and help identify opportunities for digital innovation [9]. Internally, executives’ experience supports digital knowledge absorption and enhances innovation capabilities, while mechanisms such as mergers and acquisitions provide access to key technologies and accelerate innovation [10]. Although digital innovation has been widely studied, several gaps remain. Most existing research relies on annual report disclosures, leading to homogeneous and indirect indicators [11]. Empirically, large-scale quantitative evidence remains limited [12]. More importantly, the distinction between general digital innovation and AI-driven innovation has not been made clear, and existing studies rarely explain why AI technologies—not generic digital tools—are particularly relevant for strengthening adaptive learning and resilience. Theoretical clarity regarding how AI-driven innovation strengthens adaptive learning and resilience is insufficient [13]. To address these gaps, this study integrates AI-related patent text analysis with panel data to examine how entrepreneurship enhances organizational resilience through AI-driven innovation, particularly in the context of emerging markets.

Organizational resilience develops through anticipating change, responding to crises, and learning from disruption [14], driven by cognitive awareness and proactive behavior [15]. Beyond immediate crisis response, resilience depends on continuous learning and innovation. In this context, AI-driven technologies increasingly strengthen adaptive capabilities and operational responsiveness, as demonstrated by leading Chinese firms applying AI in smart manufacturing and customer service. To measure AI-driven innovation objectively, this study uses AI-related patent text, enabling standardized and scalable assessment across firms and time. This approach supports a quantitative examination of how entrepreneurship fosters AI-driven innovation and resilience under dynamic environments. By integrating upper-echelons theory and dynamic capability theory, this study proposes a mechanism-based explanation: entrepreneurial cognition influences AI-driven innovation decisions, AI-driven innovation enhances dynamic capabilities, and dynamic capabilities ultimately promote resilience. Integrating technological, behavioral, and environmental perspectives, the study advances dynamic capability theory by revealing micro-mechanisms through which entrepreneurial cognition and AI-enabled transformation build adaptive and transformative resilience. China provides an ideal empirical setting, given its rapid digitalization, institutional transition, and high market dynamism.

2. Materials and Methods

2.1. Sample Selection and Data Sources

This paper selects the data of A-share listed companies in Shanghai and Shenzhen from 2013 to 2022 as the research object. In order to ensure the accuracy of the data, factors such as the stock industry are added on the basis of previous studies. Screening is carried out through the following steps: (1) excluding the financial industry, (2) excluding ST as well as delisted companies, (3) excluding IPO companies, (4) excluding missing samples. Meanwhile, this paper shrinks the continuous variables by 1% and 99% quantiles to reduce the influence of extreme values or outliers. After a series of screenings, a total of 3975 firms with a total of 39,750 annual observations were obtained. The data in this paper are derived from Cathay Pacific (CSMAR) and China Patent databases. To enhance methodological rigor, we applied entropy weighting and log transformation techniques to reduce noise and mitigate potential heteroscedasticity in the data, following established practices in quantitative empirical research.

2.2. Variables

Explained Variables: The explanatory variable selected was organizational resilience (Res). Organizational resilience can be measured in a number of ways, such as by constructing indicators from multiple dimensions according to its connotation, or indirectly by using financial data. In this paper, organizational resilience is measured from the two aspects of stability and resilience; in addition, given that the data of listed companies are more objective and available, this paper draws on the measurement method for the following secondary indicators. See

Table 1. Organizational resilience indicators and weights.

Dimension	Level 1 Indicators	Weights	Secondary Indicators	Weights
stability	efficiency	0.42	Fixed asset turnover	0.17
			Current asset turnover ratio	0.18
			gearing	0.07
variability	dexterity	0.58	Sales period expense ratio	0.21
			R&D costs	0.14
			Beta	0.23

below for details [16].

Explanatory variables: The explanatory variable is entrepreneurship, expressed by (Entre). Although there is no unified conclusion on this index in the academic world, according to several dimensions commonly used by most scholars, combined with the relevant research in this paper, the four dimensions of innovation, entrepreneurship, risk-taking, and talent are selected for the construction of the indexes, and the entropy method is used as the measurement method to calculate the evaluation index of corporate entrepreneurship [17]. The higher the index, the higher the entrepreneurship. Based on the foundation provided by the existing literature and theory, this paper constructs comprehensive indices from the following four dimensions, as seen in

Table 2. Comprehensive rating indicator construction system for entrepreneurship.

Level 1 Indicators	Secondary Indicators	Description of Indicators
creativity	Number of patents	Measured as the sum of three types of authorizations: inventions, utility models, and designs.
	Share of R&D investment	R&D capital investment/operating revenue
	Percentage of R&D staff	Number of product development personnel/Number of employees in the company
enterprising spirit	external investment	Total corporate outward investment as of year-end
	Management shareholding	Proportion of corporate shares held by management
	Two positions and one	The chairman and general manager are the same person
Adventurous (positive intent), courageous, and resourceful tendencies	Weighting of uncertain returns	(Financial assets held for trading + Available-for-sale financial assets + Investment properties + Accounts receivable)/Total assets
entrepreneurship	Revenue growth rate	Increase in operating income/previous year’s operating income
	cash flow per share	Net cash flow/number of common shares outstanding at end of period

, in which the construction system of the comprehensive rating indices of entrepreneurship is also outlined.

Mediating variables: AI-driven innovation in this study specifically refers to digital technological innovation activities that explicitly involve artificial intelligence-related technologies such as machine learning, deep learning, and natural language processing. By distinguishing AI-driven innovation from general digital innovation, we focus on the subset of digital transformation processes where AI serves as the technological foundation, enhancing firms’ adaptive and learning capabilities. Digital Technology Innovation (DigiInno): The focus of this paper is on the measurement of the level of digital technology innovation at the firm level. Generally, the indicator of the enterprise’s R&D innovation level is the number of patent applications. The existing literature identifies digital patents through textual information in patents, and uses the number of patent applications to establish digital technology innovation measurement indicators [18]. Some scholars conduct patent text analysis around artificial intelligence technology [19]. Recent empirical evidence further demonstrates that patent abstract-based keyword extraction is an effective and widely recognized approach for measuring digital and AI-related technological innovation in Chinese firms [20].

This paper draws on this idea to conduct a keyword text analysis of patent application documents from listed enterprises, identifying digital patents through a comprehensive keyword dictionary. We acknowledge that this approach may involve measurement errors such as false positives/negatives. To enhance robustness, we implemented several safeguards: (1) constructing the keyword dictionary based on both the academic literature and official technical reports, (2) using patent abstracts rather than annual reports for technical specificity, (3) adopting a multi-source data integration approach. Unlike prior studies that primarily extracted keywords from annual reports, this study combines firm-level financial data with AI-related patent analysis and industry-level indicators. The patent data are crawled from the national patent database, and AI-driven innovation is measured by matching abstracts with our AI-keyword dictionary. This multi-layered design extends beyond the self-reported nature of annual report data, enhancing the objectivity and dynamic validity of our DigiInno measurement.

In summary, this paper summarizes the number of digital patents of enterprises by crawling the abstract information of enterprises’ patent applications in the patent database, introducing digital technology feature words, identifying the patents in line with digital technology innovation, and adding 1 to take the logarithm of the number of digital patents of enterprises, which is expressed by the variable DigiInno, as an indicator of enterprises’ digital technology innovation. Figure 1 shows that the number of applications regarding digital technology patents of listed enterprises in China has been increasing in the past decade. This change reflects the vigorous momentum of digital technology innovation.

2.3. Research Model and Hypotheses

2.3.1. Theoretical Background

Higher-order echelon theory posits that the decisive factor for the development of the enterprise is the enterprise leader, and the role of the leader should not be ignored in the links of business management and high-level strategic decision-making [21]. “Executive Characteristics-Strategic Choice-Firm Performance” is the research framework of the higher-order theory, because each manager has different experiences and levels of experience, thus forming different personality traits, including not only values and cognition, but also culture, age, status, and economic base. In terms of the existing literature, this theory is mostly used in teams of leaders [22], the field of technological innovation [23], and corporate strategy [24], and so on. In addition, some scholars combine the two theories with the aim of making leaders’ decision-making more scientific and rational, which contributes to the further expansion of higher-order theories. For example, the functional experience of managers can improve the adaptive capacity of enterprises [25].

The concept of dynamic capability theory was first proposed by Teece et al. to be able to stand firm in the midst of fierce market competition, with the spirit of innovation and the ability to respond quickly; they pointed out that in order to quickly adapt to the external environment, enterprises need to self-adjust and constantly integrate and reconfigure the elements of knowledge, employees, and resources [26]. This paper argues that organizational resilience has three characteristics of dynamic capabilities: (1) perception ability—high organizational resilience helps enterprises to quickly identify external crises; (2) resource integration ability—in the event of a crisis, organizational resilience prompts enterprises to integrate internal and external resources to resolve the risk; (3) learning ability—organizational resilience prompts enterprises to hone and learn in the face of adversity and gain valuable development experience, recovering to the pre-impact status or even better than before. In conclusion, as an important dynamic capability of enterprises, organizational resilience is crucial to their sustainable development.

2.3.2. Hypotheses Development

(1) Entrepreneurship and organizational resilience

Higher-order theory suggests that top managers, as the core force of an organization, have different qualities and traits that lead to different cognitive and decision-making impacts on their members [27]. Entrepreneurship is an important intangible production factor, which is a comprehensive manifestation of the personality traits and governance ability of top managers, and the sustainable and healthy development of enterprises cannot be separated from the driving role of managers [28].

Organizational resilience is viewed as a dual-dimensional construct, comprising stability (robustness) and adaptability (flexibility). Stability reflects a firm’s capacity to maintain its essential functions under shocks, while adaptability captures its ability to reconfigure resources, innovate, and learn amid uncertainty. Entrepreneurship contributes differently to these two dimensions [29]. At the cognitive level, entrepreneurs with strong psychological endurance and optimism can guide firms to remain calm and focused during crises, enhancing stability, while fostering positive mindsets that encourage exploration and learning, promoting adaptability [30]. At the behavioral level, traits such as innovation, initiative, and prudence enable entrepreneurs to make quick yet strategic responses to external shocks, strengthening both operational stability and adaptive flexibility [31]. At the contextual level, entrepreneurs’ environmental sensitivity and resource rationalization foster a culture of continuous learning and renewal, reinforcing stable operations while stimulating adaptive transformation. In summary, entrepreneurship enhances organizational resilience by integrating cognitive toughness, strategic action, and contextual learning. It simultaneously stabilizes organizational structure and improves adaptive capability, providing a comprehensive foundation for sustainable enterprise development.

H1: 

Entrepreneurship positively affects organizational resilience.

(2) Entrepreneurship and Digital Technology Innovation

Based on the higher-order theory, this study argues that high entrepreneurship can provide guidance for business strategy development, and at the innovation level, this paper expects that entrepreneurship positively promotes digital technology innovation capability. Because innovation is a decision and action made to grasp the entrepreneurial opportunity, and the goal of grasping the opportunity is to be able to have independent core competitiveness, to create more and more value for the enterprise, so as to promote the sustainable development of the enterprise, employees, driven by managers, will also establish a sense of innovation, seek breakthroughs in work content and methods, and form an organizational learning atmosphere, so as to develop innovative and forward-looking products and services, which will help enterprises achieve innovative development [32].

Specifically, at the strategic level, managers with a high level of entrepreneurship will look at problems from a long-term perspective and with a strategic overview, and will be good at formulating ambitious strategic visions to help the enterprise accomplish its set goals. At the innovation incentive level, managers are good at encouraging employees to be innovative and entrepreneurial, making innovation an assessment goal, and allowing employees to be good at identifying problems and innovating solutions. In such an environment, employees are more likely to successfully innovate and create ideas for change, contributing to the achievement of the enterprise’s innovative strategic goals. A survey found that some enterprises can survive and develop even in the case of resource shortage because they can creatively integrate existing resources as a means of solving the new problems encountered, and then look for new opportunities [33].

In addition, entrepreneurship is more likely to explore market opportunities, pay attention to the hot industry dynamics, and efficiently utilize existing resources to create conditions for implementing innovative behaviors. In summary, managers with high entrepreneurship can enhance their industry status through entrepreneurial activities, respond quickly when the enterprise faces a crisis, and formulate a reasonable program for enterprise innovation and development. Therefore, it has a positive impact on the level of digital technology innovation. In summary, this paper argues that entrepreneurship can help enterprises identify and seize market opportunities and improve the level of digital technology innovation through the integration of resources. Therefore, this paper proposes Hypothesis 2:

H2: 

Entrepreneurship positively influences digital technology innovation.

(3) Digital technology innovation and organizational resilience

The existing literature suggests that organizational resilience is characterized by both stability and strain. The two seem to be contrary to each other, but in fact, a balance between the two is needed to achieve high organizational resilience. Stability allows the enterprise to shield its original state from impact when it encounters a crisis, avoiding the degradation or loss of its original functions. Contingency allows enterprises to defend against external risks, and through learning and changing in adversity, they can then achieve new innovations and further grow and expand. This paper argues that an enterprise’s digital technology innovation enhances its organizational resilience in terms of both stability and resilience.

Firstly, digital technological innovation improves organizational stability. In a dynamic environment, due to unexpected events and force majeure, the organization’s original business model or production line and financial chain can be damaged. The use of digital technologies such as big data and artificial intelligence enables the integration of these elements, facilitates communication between the top and bottom of the organization and the inside and outside of the organization, enhances the trust between members, and improves the efficiency of the organization’s business execution [33]. When encountering an external crisis, it can not only integrate the existing resources, but also exert the production substitution effect, thus reducing the loss of the crisis on the enterprise. Therefore, digital technological innovation plays the role of a resource reserve for organizational action, and these backup resources can support its continuous operation, avoid the risks brought by external shocks, and enhance its risk-resistant ability.

Secondly, digital technological innovation enhances organizational resilience. Digital technology helps enterprises to improve their data computing ability by obtaining a large amount of more comprehensive information and then filtering and integrating useful information, helping enterprises to alleviate information asymmetry. Big data computing can help enterprises quickly perceive the internal and external environment, make predictions about future trends, and also help companies pay attention to real-time market dynamics, offering a reasonable prediction of possible risks. In addition, big data storage technology can record external environmental information as well as enterprise operation data, and these results can help enterprises analyze and reflect on the deficiencies in the process of coping with crises, so as to accumulate experience and feedback for the next step of optimization [34]. Digital technological innovation also enhances the dynamic capability of enterprises, and the process of sensing, capturing, and then transforming data facilitates the process of enterprises adapting to the external environment and gaining a more comprehensive understanding of unknown risks, which in turn enhances the competitiveness and resistance of the organization. Therefore, this paper proposes Hypothesis 3:

H3: 

Digital technology innovation positively affects organizational resilience.

(4) The mediating role of digital technology innovation

Based on the dynamic capabilities perspective, entrepreneurship, digital technological innovation, and organizational resilience are strongly correlated. Digital technological innovation has changed the production mode in the past and boosted modernization and transformation. The application of digital technologies such as big data has enhanced the information processing and processing capabilities of enterprises and become a powerful tool for enterprises to gain competitive advantages in the market.

Firstly, entrepreneurship enables entrepreneurs to respond quickly to market demand, boosting the organization to make changes and create new value [35]. The improvement of the level of digital technology innovation accelerates entrepreneurs’ adaptability to the environment, and at the same time, their competitive position in the market is improved, which effectively helps enterprises to withstand external crises, make rapid and effective strategic decisions, and optimize the structure of resources.

Secondly, in the market environment, where consumers’ demands are increasingly personalized, digital technology innovations lean more to the clients’ side through enterprises, giving rise to new business models. Digital technology innovation also reduces the cost of each link in the enterprise, so the saved resources can further improve and optimize digital technology, thus generating more and higher-quality digital technology and promoting enterprise value creation [36]. Enterprise innovation can have a significant improvement on organizational capabilities; for those suffering from crisis and risk in particular, the integration of resources is key, helping the enterprise to recover from crisis and get on track, which has a positive impact on organizational resilience. Based on the above analysis, Hypothesis 4 is proposed:

H4: 

Digital technology innovation has a mediating effect on the relationship between entrepreneurship and organizational resilience.

(5) Moderating effects of environmental dynamics

In order to seek survival and development in a dynamic environment, enterprises need to make timely adjustments and changes, and the role of digital technology innovation cannot be ignored [37]. If the enterprise always maintains agile insight and accurate judgment, such a virtuous cycle can greatly enhance the enterprise’s sustainable development ability and competitive advantage in the same industry. By continuously integrating its own resources to cope with environmental changes, it ultimately enhances the organizational resilience of the enterprise. Some enterprises have overinvested in the initial stage of digital technology innovation, resulting in the irrational use of resources and leading to a financial crisis. On the contrary, organizational resilience will be enhanced if enterprises fully grasp the external environment and their own actual situation, make full use of existing resources, and make accurate decisions, especially in the early stage, to ensure that digital technology innovation activities are carried out smoothly. Therefore, it is necessary to consider environmental dynamics when researching the impact of digital technology innovation and enterprise organizational toughness. Based on the above analysis, this paper proposes Hypotheses 5a and 5b:

H5a: 

Environmental dynamics positively moderate the impact of digital technological innovation on organizational resilience.

H5b: 

Environmental dynamism positively moderates the mediating effect of digital technological innovation between entrepreneurship and organizational resilience.

Based on the higher-order and dynamic capability theory, this paper introduces digital technological innovation and environmental dynamics to construct a research model to analyze and clarify the relationship between these four variables; this study not only broadens the research on the impact of entrepreneurship and organizational resilience, but also describes the path of entrepreneurship’s impact on organizational resilience, and further analyzes the boundary conditions of the process to provide theory and solutions for enterprises to cope with crises and challenges. The research model of this study is shown in Figure 2:

2.4. Methodology

Moderating Variables

The production and business activities of a firm cannot be separated from the environment in which it operates, and the environment determines the effectiveness of the firm to a large extent. Therefore, this study notes that environmental dynamism is an important moderator that can influence the relationship between entrepreneurship and organizational resilience. While environmental dynamism creates obstacles for firms to innovate and develop, it also generates more opportunities. Changes in firms’ adaptation and coping patterns to the external competitive environment in different environments can affect the mechanism of digital technological innovation on firms’ organizational resilience. It provides insights for this paper to explore the moderating role of environmental dynamics between entrepreneurship and digital technological innovation. Building on dynamic capability theory and contingency theory, environmental dynamism is further conceptualized as a contextual moderator influencing the entrepreneurship–innovation–resilience linkage. In highly dynamic environments, firms must constantly sense market shifts, seize emerging opportunities, and reconfigure resources to adapt to rapid technological and institutional changes. Such uncertainty amplifies the importance of entrepreneurial cognition and AI-driven innovation in fostering organizational learning and adaptive resilience. By contrast, in more stable environments, resource reconfiguration is less critical, and the marginal contribution of entrepreneurship and innovation to resilience tends to weaken. Therefore, environmental dynamism determines the extent to which entrepreneurial and technological capabilities translate into sustained organizational resilience.

Based on the theory and existing literature, this paper adopts Dess’s measure to quantify environmental dynamism (2013–2022) as the coefficient of variation (standard deviation/mean of industry main business income), with results summarized in

Table 3. Variable definitions and calculations.

Variable Properties	Variable Name	Notation	Calculation Method
explanatory variable	Organizational toughness	Res	Calculated from the weights of the organizational resilience indicators constructed by [38]
explanatory variable	entrepreneurship	Entre	Entropy value method to construct a comprehensive evaluation index of entrepreneurship
intermediary variable	Digital technology innovation indicators	DigiInno	Digital technology patents plus one take the natural logarithm
moderator variable	Environmental dynamics	EN	Standard deviation of main operating income/mean main operating income by year
control variable	Enterprise size	Size	Taking the logarithm of the total assets of the firm
	Age of business	Age	Logarithm of the difference between the year in which the enterprise was counted minus the year in which it was registered.
	Corporate profitability	ROA	Net profit of the enterprise divided by the average balance of the enterprise’s total assets
	Corporate gearing	Lev	Total liabilities/total assets per year
	Nature of business	state	Nature of enterprise code: 1—State-owned enterprises; 0—private enterprises
	shareholding concentration	top	Sum of shareholdings of the top five largest shareholders of the enterprise

[38]. A higher value indicates greater environmental uncertainty. However, we acknowledge that this industry-level proxy, while valid for capturing macro-fluctuations, may not fully reflect firm-specific environmental shifts (e.g., unique competitive pressures). This inherent limitation highlights a valuable direction for future research to employ more granular, firm-level data for a nuanced understanding.

Based on the above description of indicators and theoretical analysis, model (1), model (2), and model (3) were established to test Hypotheses 1–3; drawing on the mediation effect testing approach commonly used in previous studies, model (4) and model (5) were established to test Hypothesis 4. model (4) and model (5) were established to test Hypothesis 4; and in order to test the moderating effect, model (6) and (7) were established, and the DigiInno × EN interaction term was used to test Hypotheses 5a,5b. The econometric model of this paper is set up as follows.

Based on the theories and hypotheses presented above, this paper sets the following benchmark model:

$${\mathrm{Res}}_{\mathrm{i}, \mathrm{t}} = {\mathrm{\alpha }}_0 + {\mathrm{\alpha }}_1{\mathrm{Enter}}_{\mathrm{i}, \mathrm{t}} + \sum {\mathrm{Controls}}_{\mathrm{i}, \mathrm{t}} + \mathrm{Firm} + \mathrm{Year} + {\mathrm{\epsilon }}_{\mathrm{i}, \mathrm{t}}$$

(1)

$${\mathrm{DigiInno}}_{\mathrm{i}, \mathrm{t}} = {\mathrm{\alpha }}_0 + {\mathrm{\alpha }}_1{\mathrm{Enter}}_{\mathrm{i}, \mathrm{t}} + {\sum \mathrm{Controls}}_{\mathrm{i}, \mathrm{t}} + \mathrm{Firm} + \mathrm{Year} + {\mathrm{\epsilon }}_{\mathrm{i}, \mathrm{t}}$$

(2)

$${\mathrm{Res}}_{\mathrm{i}, \mathrm{t}} = {\mathrm{\alpha }}_0 + {\mathrm{\alpha }}_1{\mathrm{DigiInno}}_{\mathrm{i}, \mathrm{t}} + \sum {\mathrm{Controls}}_{\mathrm{i}, \mathrm{t}} + \mathrm{Firm} + \mathrm{Year} + {\mathrm{\epsilon }}_{\mathrm{i}, \mathrm{t}}$$

(3)

$${\mathrm{Mediator}}_{\mathrm{i}, \mathrm{t}} = {\mathrm{\alpha }}_0 + {\mathrm{\alpha }}_1{\mathrm{Enter}}_{\mathrm{i}, \mathrm{t}} + \sum {\mathrm{Controls}}_{\mathrm{i}, \mathrm{t}} + \mathrm{Firm} + \mathrm{Year} + {\mathrm{\epsilon }}_{\mathrm{i}, \mathrm{t}}$$

(4)

$${\mathrm{Res}}_{\mathrm{i}, \mathrm{t}} = {\mathrm{\alpha }}_0 + {\mathrm{\alpha }}_1{\mathrm{Enter}}_{\mathrm{i}, \mathrm{t}} + {\mathrm{\alpha }}_2{\mathrm{Mediator}}_{\mathrm{i}, \mathrm{t}} + \sum {\mathrm{Controls}}_{\mathrm{i}, \mathrm{t}} + \mathrm{Firm} + \mathrm{Year} + {\mathrm{\epsilon }}_{\mathrm{i}, \mathrm{t}}$$

(5)

$${\mathrm{DigiInno}}_{\mathrm{i}, \mathrm{t}} = {\mathrm{\alpha }}_0 + {\mathrm{\alpha }}_1{\mathrm{Entre}}_{\mathrm{i}, \mathrm{t}} + {\mathrm{\alpha }}_2{\mathrm{EN}}_{\mathrm{i}, \mathrm{t}} + {\mathrm{\alpha }}_3{\mathrm{Entre}}_{\mathrm{i}, \mathrm{t}} \times {\mathrm{EN}}_{\mathrm{i}, \mathrm{t}} + \sum {\mathrm{Controls}}_{\mathrm{i}, \mathrm{t}} + \mathrm{Firm} + \mathrm{Year} + {\mathrm{\epsilon }}_{\mathrm{i}, \mathrm{t}}$$

(6)

$${\mathrm{Res}}_{\mathrm{i}, \mathrm{t}} = {\mathrm{\alpha }}_0 + {\mathrm{\alpha }}_1{\mathrm{Entre}}_{\mathrm{i}, \mathrm{t}} + {\mathrm{\alpha }}_2{\mathrm{EN}}_{\mathrm{i}, \mathrm{t}} + {\mathrm{DigiInno}}_{\mathrm{i}, \mathrm{t}} + {\mathrm{\alpha }}_3{\mathrm{DigiInno}}_{\mathrm{i}, \mathrm{t}} \times {\mathrm{EN}}_{\mathrm{i}, \mathrm{t}} + \sum {\mathrm{Controls}}_{\mathrm{i}, \mathrm{t}} + \mathrm{Firm} + \mathrm{Year} + {\mathrm{\epsilon }}_{\mathrm{i}, \mathrm{t}}$$

(7)

where i is the firm, t is the year, and the explanatory variable Res is the organizational resilience of firm i in year t. The explanatory variable Entre represents entrepreneurship; the mediator variable DigiInno represents the level of digital technological innovation in the year; the moderator variable EN represents environmental dynamics; Controls represents a collection of control variables including firm size (Size), firm age (Age), firm profitability (ROA), firm gearing (Lev), firm nature (State), and equity concentration (Top); ${\mathrm{\alpha }}_0 {\mathrm{is} \mathrm{a} \mathrm{constant} \mathrm{term}; \mathrm{\epsilon }}_{\mathrm{i}, \mathrm{t}}$ is the random error term; and Firm, Year, and Province represent industry and year fixed effects, respectively. In order to make the statistical results more robust, this paper uses robust standard errors for estimation.

3. Results

3.1. Descriptive Statistics

Through the descriptive statistics of each variable, it can be found that there are certain differences between the variables; this paper uses Stata17 to carry out descriptive statistics analysis on the sample data of the variables of entrepreneurship and organizational resilience of A-share listed companies from 2013 to 2022. The individual values of the variables are counted, and a preliminary understanding of the characteristics of the samples is formed. The results of the above-mentioned descriptive statistics are shown in

Table 4. Descriptive statistics.

Variant	Notation	Observed Value	Average Value	(Statistics) Standard Deviation	Minimum Value	Maximum Values
Entrepreneurship	Entre	39,750	0.088	0.111	0.001	0.365
Digital technology innovation	DigiInno	39,750	0.066	0.363	0.000	2.565
Organizational toughness	Res	39,750	2.753	0.649	0.196	3.914
Environmental dynamics	EN	39,750	0.492	0.570	−0.294	2.948
Enterprise size	Size	39,750	21.620	1.509	16.857	25.922
Age of business	Age	39,750	2.738	0.403	1.386	3.467
Corporate profitability	ROA	39,750	0.087	0.142	−0.334	0.761
Corporate gearing	Lev	39,750	0.422	0.245	0.017	1.391
Nature of business	state	39,750	0.187	0.390	0.000	1.000
shareholding concentration	top	39,750	0.572	0.153	0.212	0.939

.

According to Table 4, it can be seen that the total number of valid observations in this paper is 39,750, the maximum value of entrepreneurship in listed enterprises is 0.365, the minimum value is 0.111, and the mean value is 0.088, which indicates that there are large differences in the entrepreneurship possessed by each enterprise, and that the overall entrepreneurship level is low and needs to be further improved. The minimum value of digital technological innovation is 0, the maximum value is 2.565, the mean value is 0.066, and the standard deviation is 0.363, which indicates that there is a large difference in digital technological innovation in listed enterprises; the maximum value of organizational toughness of listed enterprises is 3.914, the minimum value is 0.649, and the mean value is 2.753, which indicates that there is a large difference in the individual differences in the enterprise’s organizational toughness and that it is not sufficiently balanced. The values of the control variables are all within a reasonable range. The data are similar to the descriptive statistics given in the existing literature, which indicates that the selection of variables in this study is reasonable, the sample data is real and effective, and it has a certain research value.

Before conducting the regression analysis, this study performed correlation and multicollinearity diagnostics for the main variables. The correlation matrix indicates that all pairwise correlation coefficients are below 0.5, suggesting no severe linear dependence among variables. The variance inflation factor (VIF) test further confirms this result, with all VIF values below 10 and a mean VIF of 1.24, far lower than the commonly accepted threshold of 10, implying the absence of multicollinearity. Subsequently, the Hausman test was applied to determine the appropriate panel specification. The result (Chi²(9) = 629.61, p = 0.0000) rejects the null hypothesis at the 1% level, supporting the use of the fixed-effects model to control for unobserved individual heterogeneity and ensure the robustness of estimation results.

3.2. Regression Results

3.2.1. The Impact of Entrepreneurship on Organizational Resilience

Using Equation (1) to test H1,

Table 5. Impact of entrepreneurship on organizational resilience.

Variant	(1)	(2)	(3)
	Res	Res	Res
Entre	1.529 ***	0.956 ***	0.923 ***
	(57.997)	(33.385)	(32.146)
Size		0.108 ***	0.105 ***
		(24.200)	(23.275)
Age		0.387 ***	0.267 ***
		(23.085)	(10.205)
ROA		0.211 ***	0.217 ***
		(8.308)	(8.539)
Lev		−0.074 ***	−0.068 ***
		(−5.117)	(−4.608)
state		0.179 ***	0.166 ***
		(11.929)	(11.027)
top		0.018	0.040
		(0.499)	(1.091)
Year			YES
Firm			YES
cons	2.619 ***	−0.759 ***	−0.368 ***
	(819.132)	(−8.499)	(−3.392)
N	39,750	39,750	39,750
R2	0.086	0.161	0.168
Adj. R2	−0.016	0.067	0.075

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

reports the regression results of the effect of entrepreneurship on organizational resilience. In column (1), without control variables, the coefficient of entrepreneurship is 1.529 and significant at the 1% level, indicating a strong positive association between entrepreneurial orientation and organizational resilience. After adding control variables in column (2), the coefficient decreases to 0.956 but remains significant at the 1% level, suggesting that the positive effect of entrepreneurship is not driven by firm characteristics alone. When time and industry fixed effects are further included in column (3), the coefficient slightly declines from 0.956 to 0.923, yet remains significant at the 1% level. This reduction implies that part of the entrepreneurial effect is absorbed by firm-specific and temporal heterogeneity; however, the robustness of the relationship confirms that entrepreneurship consistently enhances resilience across contexts. These findings align with the imprinting theory, emphasizing that entrepreneurial experience fosters adaptive capability and resource integration that strengthen organizational resilience, thus supporting Hypothesis 1.

When faced with a crisis or an uncertain environment, leaders are able to correctly judge their situation and evaluate their decisions in order to propose solutions that can be effectively implemented. Entrepreneurship gives companies the ability to be perceptive and responsive, to make conscious feedback after an unexpected event such as a crisis, to learn from adversity and setbacks, and to improve their resilience.

3.2.2. The Impact of Entrepreneurship on Digital Technology Innovation

Using Equation (2) to test H2,

Table 6. Impact of entrepreneurship on digital technology innovation.

Variant	(1)	(2)	(3)
	DigiInno	DigiInno	DigiInno
Entre	0.297 ***	0.108 ***	0.083 ***
	(16.945)	(5.497)	(4.179)
Size		0.018 ***	0.015 ***
		(5.750)	(4.761)
Age		0.078 ***	−0.025
		(6.728)	(−1.396)
ROA		−0.109 ***	−0.110 ***
		(−6.241)	(−6.288)
Lev		−0.086 ***	−0.081 ***
		(−8.590)	(−7.985)
state		0.039 ***	0.028 ***
		(3.789)	(2.656)
top		−0.025	0.005
	0.297 ***	0.108 ***	0.083 ***
Year			YES
Firm			YES
		(0.630)	(1.286)
cons	0.040 ***	−0.487 ***	−0.181 **
	(18.789)	(−7.903)	(−2.421)
N	39,750	39,750	39,750
R2	0.008	0.020	0.024
Adj. R2	−0.102	−0.089	−0.084

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

shows the regression results of the impact of entrepreneurship on digital technology innovation. In column (1), without control variables, the coefficient of entrepreneurship is 0.297 and significant at the 1% level, indicating that entrepreneurial activities significantly promote firms’ digital innovation capability. In column (2), after adding control variables, the coefficient of entrepreneurship decreases to 0.108 but remains significant at the 1% level, suggesting that the positive influence of entrepreneurship is not solely driven by firm size, age, or profitability, but reflects entrepreneurs’ intrinsic innovation orientation and risk-taking spirit. In column (3), when time and industry fixed effects are included, the coefficient further declines from 0.108 to 0.083, while still significant at the 1% level. This slight reduction implies that part of the effect is absorbed by time- and industry-specific factors such as macroeconomic policy and digital infrastructure, yet the relationship remains robust. Overall, the evidence confirms that entrepreneurial behavior plays a crucial role in facilitating digital technological innovation, supporting Hypothesis 2 and aligning with the perspective that entrepreneurship acts as a catalyst for technology diffusion and digital transformation.

Entrepreneurs with a high level of entrepreneurship are more willing to carry out digital technology innovation activities to gain a competitive advantage for their organizations. Entrepreneurs with a keen sense of insight are able to identify opportunities for digital technology innovation and thus promote digital technology innovation in their organizations. Highly entrepreneurial executives are always aware of market dynamics and are able to quickly recognize opportunities and see how these new technologies can bring disruptive changes to their businesses. Once an opportunity for digital technology innovation is identified, these entrepreneurs will not hesitate to invest in R&D and innovation. As a result, they are willing to invest a lot of time and money to form professional teams to carry out digital technology innovation activities. These innovative activities can not only help enterprises improve the quality of their products and services, but also open up new market areas and broaden their profit channels.

3.2.3. The Impact of Digital Technology Innovations on Organizational Resilience

Using Equation (3) to test H3,

Table 7. Impact of digital technology innovation on organizational resilience.

Variant	(1)	(2)	(3)
	Res	Res	Res
DigiInno	0.091 ***	0.044 ***	0.026 ***
	(11.013)	(5.275)	(3.189)
Size		0.152 ***	0.144 ***
		(64.302)	(60.356)
Age		0.048 ***	−0.008
		(5.844)	(−0.881)
ROA		0.017	0.055 **
		(0.718)	(2.352)
Lev		−0.273 ***	−0.240 ***
		(−21.722)	(−19.015)
state		−0.086 ***	−0.069 ***
		(−9.997)	(−8.016)
top		−0.529 ***	−0.493 ***
		(−25.557)	(−23.886)
Year			YES
Firm			YES
		(0.630)	(1.286)
cons	2.747 ***	−0.237 ***	0.008
	(1160.199)	(−4.494)	(0.145)
N	39,750	39,750	39,750
R2	0.003	0.144	0.156
Adj. R2	−0.107	0.144	0.156

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

presents the regression results of the impact of digital technology innovation on organizational resilience. In column (1), without control variables, the coefficient of digital technology innovation is 0.091 and significant at the 1% level, indicating that digital innovation significantly enhances firms’ capacity to adapt and recover from external shocks. In column (2), after including control variables, the coefficient decreases to 0.044 but remains significant at the 1% level, suggesting that the contribution of digital innovation to resilience is not merely attributable to firm characteristics, but stems from the intrinsic transformative effects of digital technologies on process efficiency and resource coordination. When time and industry fixed effects are further incorporated in column (3), the coefficient declines slightly from 0.044 to 0.026, yet remains significant at the 1% level. This attenuation implies that part of the observed effect is explained by time- or industry-specific digital trends, but the persistent significance demonstrates that digital innovation continues to strengthen organizational resilience. The results align with the dynamic capability theory, emphasizing that digital innovation facilitates learning, flexibility, and resource recombination that collectively enhance resilience, thereby confirming Hypothesis 3.

Digital technology innovation can help enterprises reduce production costs, visualize the production experience taught by word of mouth in the form of data-based processes, facilitate the control and improvement of the production activities process, and shorten the cost of learning time to reduce the loss of productivity brought about by personnel turnover. Deepening the application of integration of digital technology in production and operation activities can help enterprises improve business processes, so that enterprises have the ability to achieve production refinement, reduce production costs, and improve efficiency, thereby enhancing organizational resilience.

3.2.4. Tests of the Mediating Effect of Digital Technological Innovation

Using Equations (4) and (5) to test H4,

Table 8. Mediating effects of digital technology innovation.

Variant	(1)	(2)	(3)
	Res	DigiInno	Res
Entre	0.923 ***	0.083 ***	1.231 ***
	(32.146)	(4.179)	(41.880)
DigiInno			0.018 **
			(2.182)
Size	0.105 ***	0.015 ***	0.136 ***
	(23.275)	(4.761)	(57.836)
Age	0.267 ***	−0.025	−0.012
	(10.205)	(−1.396)	(−1.438)
ROA	0.217 ***	−0.110 ***	0.187 ***
	(8.539)	(−6.288)	(8.093)
Lev	−0.068 ***	−0.081 ***	−0.121 ***
	(−4.608)	(−7.985)	(−9.524)
state	0.166 ***	0.028 ***	0.012
	(11.027)	(2.656)	(1.384)
top	0.040	0.005	−0.430 ***
	(1.091)	(0.212)	(−21.241)
Year	YES	YES	YES
Firm	YES	YES	YES
		(0.630)	(1.286)
cons	−0.368 ***	−0.181 **	0.017
	(−3.392)	(−2.421)	(0.319)
N	39,750	39,750	39,750
R2	0.168	0.024	0.192
Adj. R2	0.075	−0.084	0.192

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

presents the regression results for the mediating effect of digital technology innovation. According to the stepwise testing procedure, column (1) demonstrates the relationship between entrepreneurship and organizational resilience, with a regression coefficient of 0.923, significant at the 1% level, indicating a strong total effect of entrepreneurship on organizational resilience. Column (2) reports the regression of entrepreneurship on digital technology innovation, where the coefficient of entrepreneurship is 0.083 and significant at the 1% level, confirming that entrepreneurial behavior significantly promotes digital innovation activities within firms. Column (3) includes both entrepreneurship and digital technology innovation in the model. The coefficient of entrepreneurship is 1.231 (significant at the 1% level), and that of digital technology innovation is 0.018 (significant at the 5% level). The reduction in the coefficient magnitude of entrepreneurship and the continued significance of both variables indicate a partial mediating effect. This suggests that digital technology innovation serves as an important transmission channel through which entrepreneurship enhances organizational resilience. These findings are consistent with the resource-based and dynamic capability perspectives, which argue that entrepreneurial orientation stimulates innovation investment and capability formation, thereby indirectly strengthening organizational resilience. Hence, Hypothesis 4 is supported.

In adversity, enterprises face the double pressure of “time constraints” and “resource constraints”, which gives them a certain amount of room for maneuver in reducing transaction costs and is easily obstructed by other stakeholders. In addition, when market demand changes with the environment and due to the rapid changes in the industry, the original products and services can no longer meet their needs; but companies can then go beyond this constraint by tapping into new needs and coming up with new value propositions or value creation with new partners. Companies with organizational resilience tend to have developed a change-oriented culture, a sense of innovation, and a habit of learning to improve, and tend to actively discover new growth poles when they encounter crises, while entrepreneurship is the guarantee of the implementation and success rate of digital technology innovations, which in turn improves the resilience of the organization.

3.2.5. Tests on the Moderating Effect of Environmental Dynamics

As shown in

Table 9. Moderated effects regression results.

Variant	(1)
	DigiInno
EN	0.031 ***
	(4.819)
Entre × EN	0.257 ***
	(6.020)
Entre	0.075 ***
	(3.774)
Controls	Yes
Firm	Yes
Year	Yes
cons	0.200 ***
	(2.665)
N	39,750
R2	0.026
Adj. R2	−0.083

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

, using Equation (6) to test H5a, the results of the regression of the moderating effect of environmental dynamics on entrepreneurship and digital technological innovation are shown in column (1) of Table 9. In addition, a diagram of the moderating effect of environmental dynamics is also produced (see Figure 3), and the results show that the regression coefficient of entrepreneurship in the regression of column (1) is 0.075 and is significant at the 1% confidence level, indicating that entrepreneurship has a significant positive effect on digital technological innovation. In order to explore the moderating effect of environmental dynamics, the interaction term between entrepreneurship and environmental dynamics was constructed. The results show that the coefficient of the interaction term between entrepreneurship and environmental dynamics is 0.257 and significant at a 1% confidence level, indicating that high environmental dynamics have a stronger positive promotion effect on both. In summary, it can be seen that environmental dynamism has a significant strengthening effect on the relationship between entrepreneurship and digital technological innovation, and there is a significant positive moderating effect, meaning that H5a is verified.

Environmental dynamism is the most important backup resource for enterprises, which can provide adequate resource support for enterprises to carry out innovative activities, integrate and enhance digital platforms, etc., and help to form organizational resilience. At the same time, the dynamic change in the external environment itself is also conducive to the improvement of the enterprise’s resilience, which can mitigate the risks brought about by the transformation of the enterprise and enhance the enterprise’s adaptability to the external environment. Therefore, when the environment is more dynamic, the impact of entrepreneurship, organizational resilience, and digital technology innovation is greater.

3.2.6. Moderated Mediation Effect Test

The empirical results displayed above show that environmental dynamism has a significant positive moderating effect on entrepreneurship and organizational resilience. According to the test steps of Wen Zhonglin’s moderated mediator variable, we can see that when testing the moderated mediator, we should test the mediator first and then test the moderation. In the following section, we will refer to this method to further investigate whether environmental dynamics moderate the “mediating effect of digital technological innovation between entrepreneurship and organizational resilience”.

The regression results for the moderating effect of environmental dynamics are presented in

Table 10. Moderated mediated effects regression results.

Variant	(1)	(2)	(3)	(4)
	Res	DigiInno	Res	Res
Entre	0.882 ***	0.078 ***	1.001 ***	0.836 ***
	(30.645)	(3.897)	(34.922)	(5.491)
EN	0.114 ***	0.015 **	0.141 ***	0.079
	(13.432)	(2.482)	(16.554)	(0.918)
DigiInno			0.014 *	−0.020
			(1.832)	(−0.611)
DigiInno × EN				0.071 *
				(1.671)
Controls	Yes	Yes	Yes	Yes
Firm	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
cons	−0.342 ***	−0.178 **	0.931 ***	1.131 ***
	(−3.168)	(−2.378)	(17.726)	(3.008)
N	39,750	39,750	39,750	39,750
R2	0.172	0.025	0.154	0.250
Adj. R2	0.080	−0.084	0.059	−0.166

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

. Using Equation (7) to test H5b, the regression results of the moderating role of environmental dynamics in the mediating effect of digital technological innovation on entrepreneurship and organizational resilience are shown in Table 10. In the regression of entrepreneurship and organizational resilience in column (1), the regression coefficient of entrepreneurship is 0.882 and is significant at the 1% confidence level; in the regression of entrepreneurship and digital technological innovation in column (2), the regression coefficient is 0.078 and significant at 1% confidence level; in column (3), digital technology innovation is added, and the mediator variable regression coefficient is 0.014 and significant at 10% confidence level; and in column (4), regression of entrepreneurship and organizational resilience, the interaction term of digital technology innovation and environmental dynamics is added. The results show that the coefficient of the interaction term between digital technological innovation and environmental dynamics is 0.071 and significant at the 10% confidence level. In summary, it can be seen that the moderating role of environmental dynamics in the mediating effect of digital technological innovation on entrepreneurship and organizational resilience has a more significant reinforcing role, and there is a significant positive moderating effect, and H5b is verified.

While organizational resilience can withstand the usual environmental changes and major shocks, it is even more valuable when mediated by digital technological innovations. This is mainly because, in highly dynamic environments, there is more room for innovation, even generating disruptive innovation opportunities, and highly entrepreneurial firms are able to identify and take advantage of opportunities earlier, a timing of transcendence that is more conducive to organizational resilience than being confined to the original business.

The empirical tests and analyses shown above found that the hypothesis tests in this paper are significant and the research hypotheses are valid, and a summary of the results of the tests of the research hypotheses in this paper is listed in

Table 11. Summary of results of research hypothesis testing.

Serial Number	Element	in the End
H1	Entrepreneurship positively influences organizational resilience	set up
H2	Entrepreneurship positively influences digital technology innovation	set up
H3	Digital technology innovation positively affects organizational resilience	be tenable
H4	Digital technology innovation has a significant mediating effect on the relationship between entrepreneurship and organizational resilience	be tenable
H5a	Environmental dynamics positively moderate the impact of digital technology innovation on organizational resilience	be tenable
H5b	Environmental dynamics positively moderate the mediating effect of digital technology innovation between entrepreneurship and organizational resilience	be tenable

.

3.3. Robustness Checks

After the regression analysis of entrepreneurship and organizational resilience, in order to ensure the robustness of the above findings, a robustness test was performed by replacing the explanatory and interpreted variables, shortening the time window, and introducing new control variables while ensuring that other variables remain unchanged.

3.3.1. Substitution of Variables

In the robustness test of replacing the explanatory variables, this study puts the two aspects of the enterprise’s risk resistance and adaptability into an in-depth study. In the construction of evaluation indices, this paper is based on the research results of Zhang Shaofeng and other scholars. Taking long-term growth and financial volatility as the measurement criteria, the sustainable business performance of the company was measured using three-year net sales growth accumulation, and the financial volatility of the company was measured by stock return, which was recorded as (Res2). The paper continues the regression analysis using a fixed effects model, and the regression results are shown in column (1) of

Table 12. Replacement variable regression.

Variant	(1)	(2)
	Res2	Res
Entre	0.075 ***
	(12.716)
Entre2		0.444 ***
		(33.333)
Size	0.001	0.085 ***
	(0.967)	(18.263)
Age	−0.008	0.268 ***
	(−1.403)	(10.253)
ROA	0.096 ***	0.010
	(18.523)	(0.387)
Lev	−0.027 ***	−0.075 ***
	(−8.933)	(−5.125)
state	0.005	0.136 ***
	(1.521)	(9.073)
top	0.017 **	0.004
	(2.229)	(0.120)
Firm	Yes	Yes
Year	Yes	Yes
cons	0.249 ***	0.213 *
	(11.249)	(1.909)
N	39,750	39,750
R2	0.202	0.170
Adj. R2	0.112	0.077

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

. As can be seen from the data in the table, the regression coefficient of entrepreneurship and organizational resilience after replacing the variables is still positive and significant at the 1% level, which is basically consistent with the findings of the above study, indicating that the findings of this paper are robust.

In the robustness test of replacing explanatory variables, this paper draws on Teng Haili’s study and adopts principal component analysis to measure entrepreneurship from the four dimensions of innovation, entrepreneurship, management, and responsibility, which is denoted as Entre2. This paper continues to conduct a regression analysis by using a fixed-effects model, and the regression results are shown in column (2) of Table 12. As can be seen from the data in the table, the regression coefficient of entrepreneurship and organizational resilience after replacing the variables is still positive and significant at the 1% level, which is basically consistent with the findings of the above study, indicating that the findings of this paper are robust.

3.3.2. Shortening the Sample Time Window Versus Adding Control Variables

In order to verify the robustness of the research conclusions, this study follows the approach of relevant experts and scholars by narrowing the sample time window and conducting a regression analysis of the research model. Considering that organizational resilience may have been affected by the COVID-19 pandemic after 2020, the sample period was shortened to 2013–2017 for regression analysis, and the specific results are presented in

Table 13. Shortened sample time window regression.

Variant	(1)	(2)	(3)	(4)
	Res	DigiInno	Res	Res
Entre	1.202 ***	0.039 **		0.443 ***
	(22.546)	(2.324)		(12.524)
DigiInno			0.047 **	0.096 ***
			(2.335)	(21.320)
Size	0.073 ***	−0.001	0.090 ***	0.252 ***
	(7.581)	(−1.123)	(9.352)	(9.701)
Age	0.068	0.012 ***	0.277 ***	0.269 ***
	(1.025)	(3.171)	(6.810)	(10.639)
ROA	0.125 **	−0.046 ***	0.033	−0.024 *
	(2.127)	(−4.733)	(0.555)	(−1.660)
Lev	0.091 ***	−0.009	0.034	0.165 ***
	(3.156)	(−1.552)	(1.185)	(10.949)
state	0.290 ***	0.021 ***	0.313 ***	−0.023
	(7.867)	(4.291)	(8.300)	(−0.627)
top	−0.440 ***	−0.005	−0.425 ***	0.004 ***
	(−6.093)	(−0.481)	(−5.770)	(17.562)
MngMD				−0.036
				(−1.188)
MaleRatio				0.089 ***
				(14.011)
MngFB				0.443 ***
				(12.524)
Firm	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
cons	1.004 ***	−0.009	0.175	−0.129
	(4.199)	(−0.258)	(0.977)	(−1.171)
N	19,875	19,875	19,875	39,750
R2	0.078	0.034	0.036	0.182
Adj. R2	−0.154	0.030	−0.206	0.091

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

:

According to columns (1), (2), and (3) of Table 13, the fixed-effects regression models for entrepreneurship on organizational resilience, entrepreneurship on digital technological innovation, and digital technological innovation on organizational resilience after shortening the sample time window are all significantly positive at the 1% level, consistent with the benchmark regression results.

Considering that variable bias may be omitted in the model, this paper adds a series of control variables to re-test, mainly including the level of management’s equity ratio (MngMD) affects the degree of management’s responsibility for the enterprise, and it is found that responsible executives are able to enhance the enterprise’s operational capability and promote the enterprise’s organizational resilience. Regarding the MaleRatio (MngRatio) of corporate directors and supervisors, it is generally believed that men are more willing to take more risks and are better able to withstand risks compared to women, which is more conducive to firms to improve their organizational resilience, and whether or not the directors and supervisors have a financial background (MngFB) may affect the firm’s resource conditions in times of distress, and in this paper, we chose to incorporate the control. The results are shown in column (4) of Table 13, which shows that the regression coefficient of entrepreneurship is still significant at the 1% level, suggesting that the original conclusion is robust.

3.4. Endogeneity Test

The above-shown robustness test takes into account the issue of different ways of characterizing explanatory and interpreted variables, and the robustness of the findings is verified by replacing the core variables, but there may still be endogeneity issues, in view of which, this paper utilizes instrumental variables and lagged one-period methods to test for possible endogeneity issues.

The data in this paper are all from the Cathay Pacific database and major official websites, which can effectively reduce the endogeneity problem caused by measurement error; by introducing relevant control variables, the problem of omitted variables can be solved to a certain extent. Therefore, only the endogeneity problem caused by the mutual causality of entrepreneurship and organizational resilience should be considered in this section. The possibility that the two are mutually causal needs to be taken into account. An increase in organizational toughness may also nurture and enhance entrepreneurship. Therefore, to address the potential endogeneity problem of the model, this study follows the approach of previous experts and scholars, using the annual–industry–province–region mean of entrepreneurship (Entre_Mean) as an instrumental variable. A two-stage instrumental variable regression was then conducted, and the results are reported in columns (1) and (2) of

Table 14. Instrumental variables and lagged one-period regressions.

Variant	(1)	(2)	(3)
	First Stage	Second Stage	L. Res
Entre		5.559 ***	0.338 ***
		(24.993)	(10.943)
Entre_Mean	0.774 ***
	(41.980)
Size	0.009 ***	0.106 ***	0.106 ***
	(25.395)	(20.620)	(28.785)
Age	0.007 ***	0.447 ***	−0.028 ***
	(4.95)	(22.860)	(−2.745)
ROA	−0.101 ***	0.100 ***	0.652 ***
	(−31.825)	(3.497)	(18.775)
Lev	0.091 ***	−0.035 **	0.299 ***
	(−45.59)	(−2.104)	(11.534)
state	−0.055 ***	0.063 ***	0.297 ***
	(−46.552)	(3.864)	(16.952)
top (of a list, book, TV program, etc.)	−0.033 ***	−0.176 ***	−0.209 ***
	(−9.951)	(−4.151)	(−7.561)
Firm	Yes	Yes	Yes
Year	Yes	Yes	Yes
cons	0.309 ***	0.045	−0.745 ***
	(3.818)	(0.595)	(−7.195)
N	39,750	39,750	35,775
R2	0.226	0.726	0.104
Adj. R2	0.226	0.728	−0.008

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

. The findings indicate that the conclusions remain robust after controlling for possible endogeneity.

To further reduce the impact of endogeneity, this paper chooses lagged one-period organizational toughness as an instrumental variable, denoted as (L.Res), while controlling for year and industry fixed effects, and the regression results show that the two are still positively correlated and significant at the 1% level, which reveals that, after a variety of robustness and endogeneity tests, the obtained conclusions are still in very high consistency with the findings of the above-outlined study.

3.5. Heterogeneity Analysis

3.5.1. Regional Heterogeneity

The level of economic development in different geographic locations is also different. The main effect of regional heterogeneity on this study, according to the relevant provisions of the state, in addition to those of Hong Kong, Macao, Taiwan, and Tibet, is in accordance with the eastern region, the central region, and the western region in the division, as shown in

Table 15. Division of the country into regions.

Suffix of the City Name Means the Prefecture or County	Province (City, District)
Eastern Part	Beijing, Guangdong, Hainan Tianjin, Fujian, Shandong, Hebei, Liaoning, Jiangsu, Zhejiang, Shanghai
Central Region	Shanxi, Jilin, Heilongjiang, Anhui, Hubei, Hunan, Guangxi, Inner Mongolia, Jiangxi, Henan
Western Region	Chongqing, Qinghai, Ningxia, Guizhou, Yunnan, Sichuan, Shaanxi, Gansu, Xinjiang

.

As seen in the regional heterogeneity test results in

Table 16. Regression of regional heterogeneity.

Variant	(1)	(2)	(3)
	Eastern Part	Central Region	Western Region
Entre	0.980 ***	0.686 ***	0.717 ***
	(31.268)	(7.868)	(5.772)
Size	0.106 ***	0.100 ***	0.105 ***
	(20.889)	(8.046)	(6.486)
Age	0.229 ***	0.507 ***	0.220 **
	(7.955)	(6.770)	(2.096)
ROA	0.209 ***	0.186 **	0.274 ***
	(7.530)	(2.482)	(2.709)
Lev	−0.109 ***	−0.059	−0.199 ***
	(−5.857)	(−1.319)	(−2.974)
state	−0.015	0.423 ***	−0.306 **
	(−0.344)	(4.577)	(−2.469)
top (of a list, book, TV program, etc.)	0.980 ***	0.686 ***	0.717 ***
	(31.268)	(7.868)	(5.772)
Firm	Yes	Yes	Yes
Year	Yes	Yes	Yes
cons	−0.191	−1.060 ***	0.033
	(−1.559)	(−3.529)	(0.083)
N	29,610	6580	3470
R2	0.185	0.125	0.132
Adj. R2	0.094	0.026	0.030

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

for China’s eastern, central, and western regions, regional heterogeneity has a significant role in. From the regression coefficient point of view, that of the eastern region is the largest, at 0.980, while the coefficients of the central and western regions are relatively small, less than 0.6; the main possible reason for this could be that the eastern region had an early start in the development of the market environment and now has a set of relatively developed businesses, and that, on the contrary, the central and western regions have relatively underdeveloped infrastructures and a lack of large-scale enterprises to promote. On the contrary, the relatively underdeveloped infrastructure in the central and western regions lacks the drive of large-scale enterprises, and most enterprises are concerned about short-term interests, resulting in a relatively weak ability to cope with crises. Therefore, the central and western regions still need the help of the eastern regions, and at the same time, further investment is needed in the areas of infrastructure and technology.

3.5.2. Heterogeneity in Firm Size

The effect of entrepreneurship on organizational resilience is next analyzed for size heterogeneity. The regression results in

Table 17. Regression of size heterogeneity.

Variant	Major Industry	Small- or Medium-Sized Enterprise (SME)
	(1)	(2)
	Res	Res
Entre	0.797 ***	0.512 ***
	(16.722)	(22.971)
Age	0.304 ***	0.771 ***
	(8.364)	(14.657)
ROA	0.125 ***	0.127 ***
	(2.642)	(3.575)
Lev	−0.092 ***	−0.103 ***
	(−4.089)	(−4.529)
state	0.183 ***	0.118 ***
	(9.838)	(4.188)
top (of a list, book, TV program, etc.)	0.141 ***	0.008
	(2.653)	(0.115)
Firm	Yes	Yes
Year	Yes	Yes
cons	1.800 ***	0.740 ***
	(17.460)	(5.393)
N	20,418	19,332
R2	0.098	0.147
Adj. R2	−0.049	−0.001

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

show that the regression coefficient of entrepreneurship for large-scale firms in column (1) is 0.797 and significant at the 1% confidence level, while the regression coefficient of entrepreneurship for small and medium-sized firms in column (2) is 0.512 and significant at the 1% confidence level, which suggests that the entrepreneurship of large-scale firms has a more significant impact on organizational resilience compared to that of small and medium-sized firms.

The effects of entrepreneurship implementation are bound to vary for enterprises of different sizes. On the premise that the enterprise has a certain scale, it can carry out relevant activities and has a higher organizational resilience to withstand the risks that entrepreneurship may bring. Therefore, if there is sufficient financial support with a high risk tolerance, the enterprise has a high degree of voice in the market competition. On the other hand, SMEs with insufficient control of resources and in a marginal position in the economic network have a lot of limitations on their practice of entrepreneurship, and are not fast enough at obtaining resources to convert entrepreneurial achievements into commercial value, which is a great obstacle to the development of entrepreneurship in the large-scale promotion of entrepreneurship. In addition, when the entrepreneurship for small and medium-sized enterprises becomes unaffordable, it cannot continue to support the enterprise to carry out further entrepreneurship, which requires investment. This is likely to cause a lack of effective entrepreneurship, which greatly constrains the ability of entrepreneurship to help the development of the enterprise’s financial and organizational resilience. In comparison, large enterprises are much more likely to benefit from practicing entrepreneurship than SMEs.

3.5.3. Industry Heterogeneity

The following is an industry heterogeneity analysis of the relationship between entrepreneurship and organizational toughness.

Table 18. Regression of industry heterogeneity.

Variant	High-Tech Enterprises	Non-High-Tech Enterprises
	(1)	(2)
	Res	Res
Entre	0.943 ***	0.415 ***
	(33.472)	(2.939)
Age	0.108 ***	0.059 ***
	(24.144)	(2.682)
ROA	0.157 ***	0.532 ***
	(4.652)	(7.795)
Lev	0.204 ***	0.480 ***
	(8.075)	(4.116)
state	−0.108 ***	0.071
	(−6.947)	(1.592)
top	0.144 ***	0.275 ***
	(9.515)	(4.537)
Firm	Yes	Yes
Year	Yes	Yes
cons	−0.078	−0.394
	(−0.638)	(−0.847)
N	33,490	6260
R2	0.197	0.092
Adj. R2	0.097	−0.078

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

shows the regression results of entrepreneurship on organizational toughness in high-tech enterprises and non-high-tech enterprises, and it is not difficult to see that the regression coefficient of entrepreneurship in the high-tech industry group is 0.943 and significant at the 1% level, and the regression coefficient of entrepreneurship in the non-high-tech enterprise group is 0.415 and significant at the 1% level, which indicates that high-tech enterprises themselves represent a powerful driving force of entrepreneurship. This indicates that high-tech enterprises themselves represent a strong level of technological innovation and financial strength, which has become the intrinsic motivation for their development. On the other hand, the government’s support for enterprise innovation will likely be stronger for high-tech enterprises, which will in turn have a certain advantage.

The frequency of technological renewal and the output of innovation performance in high-tech industries are relatively higher than that of non-high-tech enterprises; the entrepreneurship of high-tech enterprises is more likely to have a focus on developing insight into the opportunities of market development, making timely adjustments to the strategies and products, and seizing the market to enhance competitive advantage. At the same time, high-tech enterprises will also attract the favor of investors, as it is easier to occupy a larger proportion of the market by seeking cooperation. This requires the enterprises to have good operational efficiency, but also a strong organizational toughness.

3.5.4. Analysis of Secondary Indicators of Entrepreneurship Impacting Organizational Resilience

This paper measures entrepreneurship in terms of four dimensions—innovation, entrepreneurship, risk-taking, and talent—and explores the impact of entrepreneurship on organizational resilience in terms of these four dimensions in order to provide insights into the positive driving effect of entrepreneurship on organizational resilience.

Table 19. Regression results of secondary indicators of entrepreneurship affecting organizational resilience.

Variant	(1)	(2)	(3)	(4)
	Res	Res	Res	Res
ChuangXin	6.800 ***
	(31.054)
ChuangYe		0.817 ***
		(26.359)
MaoXian			7.460 ***
			(29.497)
CaiNeng				17.323 ***
				(11.446)
Size	0.112 ***	0.111 ***	0.111 ***	0.119 ***
	(24.856)	(24.522)	(24.675)	(26.034)
Age	0.289 ***	0.277 ***	0.263 ***	0.322 ***
	(11.023)	(10.516)	(10.030)	(12.146)
ROA	0.148 ***	0.187 ***	0.180 ***	0.110 ***
	(5.891)	(7.336)	(7.112)	(4.314)
Lev	−0.110 ***	−0.092 ***	−0.089 ***	−0.136 ***
	(−7.550)	(−6.259)	(−6.040)	(−9.262)
state	0.123 ***	0.166 ***	0.109 ***	0.126 ***
	(8.196)	(10.965)	(7.231)	(8.318)
top	0.100 ***	0.032	0.075 **	0.009
	(2.686)	(0.857)	(2.007)	(0.235)
Firm	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
cons	−0.526 ***	−0.482 ***	−0.489 ***	−0.686 ***
	(−4.872)	(−4.438)	(−4.512)	(−6.275)
N	39,750	39,750	39,750	39,750
R2	0.166	0.160	0.164	0.147
Adj. R2	0.073	0.067	0.071	0.052

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

reports the regression of the dimensions of entrepreneurship on organizational resilience

The regression results show that innovation, entrepreneurship, risk-taking, and talent all significantly affect organizational resilience. From the regression coefficients, it can be seen that entrepreneurial talent has the strongest facilitating effect, followed by risk-taking, then innovation and entrepreneurship.

3.5.5. Analysis of the Role of Environmental Dynamics on Secondary Indicators of Entrepreneurship Affecting Organizational Resilience

To further analyze the moderating effect of environmental dynamics, this study explores the moderating effect of environmental dynamics by categorizing entrepreneurship into the four elements of innovation, entrepreneurship, risk-taking, and talent.

Table 20. Analysis of the effects of environmental dynamics on the secondary indicator of entrepreneurship affecting organizational resilience.

Variant	(1)	(2)	(3)	(4)
	Res	Res	Res	Res
Entre	0.762 ***	1.015 ***	0.743 ***	0.898 ***
	(18.500)	(17.342)	(17.249)	(30.540)
EN	0.093 ***	0.118 ***	0.099 ***	0.114 ***
	(6.977)	(8.921)	(7.412)	(13.294)
CX × EN	4.798 ***
	(10.359)
CY × EN		−0.299 ***
		(−2.753)
MX × EN			4.286 ***
			(7.261)
CN × EN				3.618 **
				(2.472)
Size	0.093 ***	0.098 ***	0.095 ***	0.100 ***
	(11.655)	(12.169)	(11.634)	(22.170)
Age	0.337 ***	0.292 ***	0.318 ***	0.410 ***
	(7.940)	(6.811)	(7.448)	(24.415)
ROA	0.149 ***	0.097 **	0.137 ***	0.102 ***
	(3.507)	(2.297)	(3.226)	(3.814)
Lev	−0.068 ***	−0.093 ***	−0.074 ***	−0.093 ***
	(−2.695)	(−3.717)	(−2.923)	(−6.356)
state	0.157 ***	0.165 ***	0.156 ***	0.178 ***
	(5.830)	(6.097)	(5.794)	(11.844)
top	0.034	0.054	0.024	0.018
	(0.543)	(0.858)	(0.371)	(0.477)
Firm	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
cons	−0.308 *	−0.326 *	−0.314 *	−0.676 ***
	(−1.701)	(−1.779)	(−1.692)	(−7.558)
N	39,750	39,750	39,750	39,750
R2	0.178	0.173	0.176	0.165
Adj. R2	0.178	0.172	0.175	0.072

Note: ***, **, and * indicate significance at the 1%, 5%, and 10% levels, respectively, with t-values in parentheses.

shows that environmental dynamics significantly positively moderate the effects of innovation, risk-taking, and entrepreneurial talent on organizational resilience. However, the coefficient of the interaction term between entrepreneurship and environmental dynamics is significantly negative, indicating that environmental dynamics inhibit the effects of entrepreneurship on organizational resilience, possibly because when the environment faced is suddenly unpromising, with serious and imminent consequences, it requires energy to learn about the environment and find resources and methods to cope with this challenge, meaning that employees within the organization will exacerbate the undesirable state of social inertia, and that leadership innovation incentives in this environment will not produce a more significant effect. When compared to the behavioral changes that can lead to positive results, it can be suggested that people are more likely to change their behavior in order to avoid the emergence of negative results and to take conservative measures because of the reduced flexibility of the loss of opportunities for development, which is detrimental to the organizational resilience of the enterprise.

4. Discussion

To cultivate and apply entrepreneurship more effectively, firms should identify differences in entrepreneurial traits and design targeted development programs that strengthen opportunity recognition, risk management, and innovation motivation. Building supportive institutional conditions—such as tax incentives, IP protection, and rewards for technological breakthroughs—can further stimulate entrepreneurial vitality. Enterprises in emerging markets should also leverage AI-driven technologies to enhance resilience by improving decision-making accuracy, risk prediction, supply chain transparency, and crisis-response capabilities. Establishing digital training systems, AI data platforms, and experience-sharing mechanisms can help embed entrepreneurial culture and promote sustainable growth. Given that entrepreneurship shows stronger resilience effects in eastern, large-scale, and high-tech firms, policies and managerial efforts should especially support SMEs and firms in central and western regions to narrow capability gaps and strengthen long-term sustainable competitiveness under dynamic environments.

Fostering organizational resilience is vital for long-term sustainable development. Strengthening resilience enables firms to withstand shocks and convert crises into opportunities. Firms should optimize resource allocation to enhance flexibility and build AI-enabled digital response systems that support rapid adaptation. In dynamic environments, uncertainty should be viewed as a catalyst for learning; by monitoring external signals, predicting risks, and reallocating resources with data-driven tools, firms can shift from passive reaction to proactive adaptation. This requires clarifying strategic positioning and institutionalizing continuous learning and reflective routines. We appreciate the reviewer’s suggestion to enrich the study with qualitative insights; future research will incorporate case studies of firms undergoing AI-driven transformation to deepen understanding of resilience-building mechanisms. Together, these strategies reinforce resilience’s dual attributes—robustness and adaptability—supporting long-term sustainable growth.

5. Conclusions

In summary, the empirical results provide strong support for all six hypotheses. Entrepreneurship significantly enhances organizational resilience and promotes digital technological innovation, while AI-driven digital innovation serves as a key mechanism translating entrepreneurial cognition into adaptive capacity, contributing to both robustness and flexibility. Environmental dynamism further amplifies these effects, strengthening both the direct and indirect pathways. These findings highlight that entrepreneurship and AI-enabled innovation jointly foster long-term organizational resilience, a core capability for sustainable development in dynamic environments, thereby linking entrepreneurial actions, technological upgrading, and resilience to broader sustainability outcomes.

At present, China is undergoing rapid development, requiring entrepreneurial support to promote high-quality growth and industrial upgrading. Entrepreneurship acts as a key driver of digital technological innovation, and AI-related innovation enhances firms’ adaptive responses to external shocks, thereby strengthening organizational resilience. Through this mechanism, entrepreneurship indirectly improves resilience via digital innovation, highlighting a clear mediating pathway. Moreover, dynamic environments further stimulate digital innovation and amplify resilience-building effects. This process not only enhances short-term adaptability but also promotes long-term sustainable development by enabling firms to continuously learn, evolve, and innovate under uncertainty.

Compared with smaller firms, large enterprises benefit more from entrepreneurship-driven resilience, especially in high-tech sectors where innovation capabilities are crucial. Higher levels of innovativeness, risk-taking, and entrepreneurial talent are consistently associated with stronger organizational resilience, and environmental dynamism further amplifies these effects, though excessive volatility may weaken marginal gains. To support resilience and long-term sustainable development, firms should cultivate entrepreneurial culture and strengthen AI-enabled digital innovation capacity. Meanwhile, policymakers can enhance firms’ adaptive capabilities by providing innovation subsidies, AI infrastructure, and targeted SME support, jointly promoting organizational resilience and sustainable growth.

This study relies on panel data from listed firms, which may not fully represent SMEs or private enterprises, thus limiting external validity. Potential endogeneity may also exist—for example, entrepreneurs with strong entrepreneurial spirit may inherently possess higher resilience, raising concerns of reverse causality. In addition, industry-level environmental dynamism may not capture firm-level variations. Future research may use qualitative methods or causal designs to deepen understanding.