How the Digital Transformation of Chinese Traditional Manufacturing Enterprises Drives Green Innovation: A Moderated Mediation Model

Abstract

Ecological protection and sustainable economic development are focal points of global attention. The continuous promotion of green technology during digital transformation has made green innovation essential for reducing environmental pollution and achieving sustainability. China is actively exploring new development models to mitigate environmental damage through digital transformation. This paper examines the impact mechanism of digital transformation on green innovation. It introduces ambidextrous learning as a mediating variable and innovation appropriability as a moderating variable. Additionally, it establishes a moderated mediation model grounded in organizational learning theory and innovation appropriability theory. The survey data from 155 traditional Chinese manufacturing enterprises are analyzed using hierarchical regression, bootstrap sampling, and structural equation modeling. The results indicate that digital transformation significantly boosts corporate green innovation. Both exploitative and exploratory learning act as positive partial mediators in this relationship. Furthermore, innovation appropriability further reinforces this relationship. The findings of this paper provide empirical evidence that supports the promotion of green innovation in enterprises. This evidence is essential for helping enterprises achieve sustainability and facilitate their green transformation.

1. Introduction

Enterprise green innovation is a crucial component of sustainable development. It aids in energy conservation, pollution reduction, and the transition to a low-carbon economy [1]. It not only plays a crucial role in the paradigm shift toward sustainability but also offers a new impetus for economic development [2]. Furthermore, it greatly contributes to achieving the “dual-carbon goal”. Concurrently, the real economy and the rapid advancement of digital technology have resulted in the emergence of a digital economy that is defining a “new quality of productive forces”. This development affects the green innovation efforts of traditional manufacturing enterprises [3]. The “2024 Implementation Guidelines for Digital Greening and Collaborative Transformation and Development” outlined by the Chinese government, stresses the necessity of promoting sustainability. These guidelines highlight the importance of green transformation and the acceleration of cultivating new quality productive forces. This involves two main directions: First, it focuses on promoting the green and low-carbon development of digital industries. Second, it aims to accelerate the green transformation of industries empowered by digital technologies. In this context, green development has become the new normal for Chinese enterprises [4], while digital transformation significantly supports market-oriented green enterprises [5]. Consequently, effective strategies for attaining the “dual carbon” objective and fostering sustainable economic and social development include several key actions. These actions involve clarifying how digital transformation impacts green innovation.

Digital transformation functions as both an economic engine and a tool for balancing economic gains with environmental performance. It enhances the total productivity factor by upgrading information technology and data processing capabilities [6]. Additionally, enterprises can boost their green innovation performance by improving internal information transparency, optimizing governance, enhancing technology integration, strengthening dynamic capabilities, alleviating financing constraints, and reinforcing external media scrutiny [7,8,9,10,11,12]. However, digital transformation is not always beneficial. Green innovation driven by digital transformation requires extensive data. Organizational inertia can cause firms to persist in data collection for revenue generation [13]. This can make the costs of managing and utilizing the information surpass the benefits of green innovation. Consequently, Li (2022) argues that a high degree of digital transformation may lead to poorer environmental performance [14]. Additionally, resource dependence theory posits that digital transformation can increase financial pressure on enterprises. It can also drain the energy of executives and IT departments [15]. This shift often diverts resources and time away from green innovation activities. Given the unique characteristics of green innovation compared to traditional innovation [16], the potentially contradictory relationship between digital transformation and green innovation warrants further investigation.

Technological innovation and resource-based perspectives have been the primary focus of prior research on digital transformation and green innovation. However, there has been less emphasis on ambidextrous learning in this context. In the VUCA era, where organizations frequently interact with their environment, ambidextrous learning theory enables enterprises to exploit existing knowledge while exploring new knowledge. This dual capability helps organizations to form new knowledge and adapt to turbulent environments [17]. Exploitative learning serves as the primary method for enterprises to deepen their existing knowledge. Through this process, organizations can efficiently utilize their current knowledge and develop a green knowledge portfolio that meets market demands. In contrast, exploratory learning is essential for acquiring new knowledge; it enables enterprises to obtain the knowledge necessary for green innovation and expand the required knowledge system [18]. Exploitative and exploratory learning together establish the foundational knowledge required to foster green innovation.

To promote green innovation, enterprises must incur significant research and development costs. However, because such innovations are easily imitated and challenging to derive value from [19], companies must construct an innovation appropriability mechanism to achieve favorable returns from their green innovation efforts. The innovation mechanism refers to the measures taken by enterprises to secure their innovation benefits. This includes formal exclusivity measures that provide legal protections, such as patents, trademarks, copyrights, and contracts. It also encompasses informal exclusivity measures, such as trade secrets, complementary assets, and first-mover advantages [20]. In summary, firms can advance their green innovation activities and sustainable development through three key strategies. First, they should implement digital transformation. Second, enhancing ambidextrous learning is crucial. Finally, leveraging innovation appropriability mechanisms can yield significant benefits. Together, these strategies help organizations to advance their sustainability initiatives effectively. This study aims to employ organizational learning theory and innovation appropriability theory to fill the research gaps. There is limited research on how digital transformation connects to green innovation, especially regarding the role of organizational learning. Additionally, scholars have not yet explored how innovation appropriability, which protects traditional innovation gains, might influence green innovation. Specifically, this study seeks to explore how digital transformation fosters green innovation through ambidextrous learning. It also investigates the role of innovation appropriability in this process. The findings aim to provide both theoretical support and practical guidance for enterprises undergoing digital transformation. Utilizing survey data from 155 traditional manufacturing firms in China, the empirical findings reveal that digital transformation directly drives green innovation. Ambidextrous learning, which includes both exploitative and exploratory learning, serves as a partial mediator in this relationship. Additionally, this study finds that innovation appropriability positively moderates the connection between ambidextrous learning and green innovation. It also significantly strengthens the mediating effect of ambidextrous learning in the pathway from digital transformation to green innovation.

This study makes the following contributions: (1) We enrich the research literature concerning digital transformation and green innovation by proposing a comprehensive theoretical framework that integrates digital transformation, ambidextrous learning, green innovation, and innovation appropriability. By doing so, it enhances the existing literature on both digital transformation and green innovation. Additionally, this study is contextualized within the broader imperative of promoting sustainability. (2) We innovate the research perspective of the relationship between digital transformation and corporate green innovation. This study employs ambidextrous learning as a research lens, decomposing it into two essential dimensions: exploitative learning and exploratory learning. We comprehensively investigate the mediating effects of these distinct learning dimensions on the relationship between digital transformation and green innovation. Furthermore, this study introduces innovation appropriability as a moderating variable. This inclusion offers new insights into mitigating the spillover effects of green innovation benefits. (3) The conclusions drawn from this study present crucial recommendations for both businesses and government policymakers. The empirical results offer valuable insights for enterprise management, guiding managers in effectively implementing digital transformation. Additionally, the research results encourage enterprises to strengthen the role of digital transformation in promoting green innovation. This can be achieved by adopting ambidextrous learning strategies and establishing an innovation appropriability mechanism. Such implications are practical for sustaining and enhancing green innovation in manufacturing firms.

The remainder of this paper is structured as follows: (1) Literature Review. This section presents a systematic review of pertinent studies on digital transformation, ambidextrous learning, green innovation, and innovation appropriability. (2) Theoretical Model and Hypotheses. This section outlines the theoretical framework and research hypotheses. (3) Research Methods. This section details the measurement of variables and the questionnaire collection process. (4) Empirical Results. This section discusses the empirical study’s findings in detail. (5) Conclusions and Implications. This section presents the managerial implications and discusses the study’s findings and limitations.

2. Literature Review

2.1. Digital Transformation, Green Innovation, and the Relationship Between Them

Digital transformation acts as a catalyst and presents a major opportunity for traditional manufacturing enterprises. It enables leapfrog development and fosters green innovation. This process involves transforming and improving conventional industries by incorporating emerging technologies. These technologies include artificial intelligence, cloud computing, blockchain, and big data [21]. Central to this process is the extensive utilization of data elements [22]. Companies leverage the advantages of data, such as replicability, seamless linking, simplified simulation, and responsive feedback. This allows them to enhance key areas such as production management, business operations, innovation, research and development, and value creation. Through this integration, enterprises can improve interdepartmental collaboration and provide precise and timely solutions to their customers. Consequently, they cultivate sustainable competitive advantages [23] and facilitate their entry into emerging markets. This study asserts that implementing a digital transformation represents a pivotal shift. It involves not only the deep integration of digital technologies with production efficiency but also marks the emergence of a new framework. This framework supports the transition from traditional manufacturing to the digital economy.

Green innovation, often called “sustainability innovation” or “environmental technology innovation,” is defined by Schiederig (2012). It is characterized as an innovation activity aimed at achieving significant environmental benefits. Additionally, it promotes sustainable development [24]. This study posits that green innovation is not merely a corporate response to environmental changes. It is also crucial to integrate the needs of nature, society, and the economy for long-term sustainable development. Currently, China is navigating a complex landscape, actively pursuing green innovation strategies in response to increasing environmental demands from customers. Upgrading digital technologies, assimilating green knowledge, and standardizing intellectual property protection mechanisms have become core issues for corporate management. These elements are essential for achieving efficient, sustainable, and profitable green innovation. We present an overview of the key drivers of green innovation identified by recent academic research. These factors primarily include the internal corporate governance structure, external scientific and technological developments, and environmental policies [25]. Key internal elements involve executives’ environmental awareness, R&D investment, and employee compensation [26,27]. External factors encompass environmental regulations and governmental regulatory pressures [28]. Furthermore, scholars such as Tipu (2022) emphasize the significance of organizational learning theory and culture for sustainable business growth [29]. Some researchers have noted that fully explaining the complex antecedents of green innovation is challenging when analyzing a single factor. Therefore, it is essential to explore the interactions among technological, organizational, and environmental dimensions. Such exploration will provide a deeper understanding of how green innovation performance is developed [30]. In summary, it is evident that existing research has somewhat overlooked the impact of an enterprise’s capacity for developing and absorbing green knowledge on green innovation.

This study presents three views on the relationship between digital transformation and green innovation. Firstly, digital transformation may act as a catalyst for green innovation. Vial (2021) notes that recent research on the consequences of digital transformation has predominantly focused on its economic impacts and implications for corporate governance [31]. Digital transformation enhances market competitiveness by significantly reducing operating costs for enterprises [32]. Additionally, it promotes transparency in management processes. This improvement boosts the efficiency of human resource management [33] and capital allocation [34], while also mitigating decision-making risks. However, there is a notable gap in the research regarding the specific environmental impacts of digital transformation and its potential to facilitate green development outcomes. Most of the literature suggests that digital transformation creates favorable conditions for green innovation activities. Digital technologies enable companies to connect the chains of technology, data, and knowledge, thereby reducing information acquisition costs. This allows them to analyze environmental data more effectively and adjust their operational strategies to achieve sustainable development goals [35]. Nonetheless, Li (2022) noted in his study that the operation of digital transformation facilities within enterprises is heavily reliant on substantial energy consumption [14]. This reliance does not consistently lead to the desired results in green innovation [36]. As Chinese enterprises increasingly prioritize “green development” through “digital leadership” as a strategic imperative, the examination of the impact of digital transformation on green innovation is of substantial theoretical and practical significance.

Secondly, digital transformation and green innovation are mutually reinforcing and synergistic. The development of green innovation and the implementation of digital transformation mutually reinforce one another, resulting in a symbiotic cycle. Digital transformation improves the efficiency and quality of green innovation. Simultaneously, the advancement of green innovation offers the necessary impetus and conditions to enhance the digital transformation of enterprises [37]. Enterprises must implement green innovation as a critical strategy to improve their competitive edge [38]. It encompasses aspects of pollution prevention, energy efficiency, waste recycling, green product design, and technological innovations related to environmental management [39]. All of these factors are linked to accelerating digital system innovation. They also enhance the decarbonization capacity of production systems and improve production efficiency. Therefore, if enterprises do not increase their investment in green innovation, they will lack the ability to develop sustainability, even with advanced technologies.

Lastly, the synergy between digital transformation and green innovation is crucial for achieving sustainable development. Promoting digital transformation and fostering green innovation enhance sustainability through effective resource recycling. This synergy generates a virtuous cycle of environmental protection and economic expansion. Additionally, embedding green technologies throughout the digital transformation process can drive the economy toward low-carbon emissions, cleanliness, and circularity. Simultaneously, green innovation offers businesses a future direction for digital transformation, collectively advancing the pursuit of sustainability.

2.2. Ambidextrous Learning

The term “ambidextrous learning” was initially introduced to the field of organizational learning by March (1991). It denotes the simultaneous engagement of organizations in two distinct categories of learning activities: exploitative learning and exploratory learning [17]. Exploitative learning involves exploring and utilizing existing knowledge to enrich the knowledge stock. It emphasizes improving current processes, products, or services for gradual enhancements, such as optimizing supply chains for cost savings. In contrast, exploratory learning entails the continuous search, absorption, and even creation of new knowledge that differs from existing reserves. It seeks new opportunities and innovations, such as creating new technologies or exploring emerging markets. As a dual process, ambidextrous learning helps to enrich the overall knowledge system [40]. Notable differences exist between these two learning types in terms of direction, processes, and objectives. Exploitative learning takes an inward-looking approach. It focuses on learning and research activities that respond to minor changes in organizational processes. This type of learning emphasizes gradual modifications and improvements in existing products or knowledge areas [41]. In contrast, exploratory learning is primarily outward-focused. It involves communication and collaboration with relevant organizations to access knowledge and resources from clients, institutions, universities, and research institutes. This approach often carries uncertainty regarding potential benefits [42].

The knowledge-based view, proposed by Grant (1996), posits that knowledge is crucial for continuous innovation [43]. Moreover, breakthroughs in cross-cutting and cutting-edge technologies rely on the creation of new knowledge [44]. To achieve sustainable green innovation, companies must not only efficiently apply and process existing knowledge but also create or assimilate new knowledge. This dual process is facilitated through both exploitative and exploratory learning. Specifically, exploitative learning builds on existing knowledge and focuses on upgrading current technologies to enhance productivity. In contrast, exploratory learning emphasizes the pursuit of new knowledge to create entirely new products and technologies. In fact, recent studies indicate that these two learning activities collectively enhance a firm’s innovation capability, such as technological innovation [45] and breakthrough innovation [46]. Although the positive influence of ambidextrous learning on corporate innovation has been theoretically verified, it remains unclear whether it significantly influences green innovation at the corporate level.

The theory of organizational ambidextrous learning provides a vital foundation for digital transformation and green innovation. This theory emphasizes the need to balance the use of existing resources and knowledge while exploring new production models and adapting to green technologies. Balancing these aspects helps companies maintain a competitive advantage in a rapidly changing digital environment and provides robust support for green innovation practices. However, implementing ambidextrous learning presents several challenges. First, conflicts in resource allocation may arise, leading to reduced efficiency in exploration and utilization activities. Second, many traditional corporate cultures do not encourage risk-taking, further hindering progress in digital transformation and green innovation. In particular, firms engaged in green innovation must actively incorporate knowledge from external collaborations, which often extends beyond the scope of ambidextrous learning theories. Therefore, future research should explore how to integrate additional theoretical frameworks to address the limitations of ambidextrous learning theory.

2.3. Innovation Appropriability

The theory of innovation appropriability was first proposed by Teece (1986), who emphasized that a firm’s ability to derive economic benefits from innovation depends on two key factors. First, it hinges on effectively discouraging imitation by competitors. Second, it involves enhancing innovation appropriability by raising the cost of imitation [47]. This theory provides a crucial foundation for understanding how firms protect their green innovations in a competitive environment. Innovation appropriability is defined as “maximizing revenue while ensuring that the innovation is not imitated” [48]. For instance, a pharmaceutical company may invest in research and development to create a new drug. It might choose patent protection or trade secrets to prevent other companies from imitation. The dualistic nature of innovation appropriability is widely acknowledged by scholars, who categorize it into formal and informal appropriability [20]. Formal appropriability mechanisms are institutional methods for protecting innovations through legal means. These methods primarily involve legal instruments such as patents, trademarks, and copyrights, granting exclusive rights at the legal level. In contrast, informal appropriability mechanisms comprise a set of strategies employed by firms to protect innovations without relying on legal means. These strategies increase resource barriers and enhance internal capabilities, primarily including tools such as trade secrets, complementary assets, complex technologies, and delivery cycles. They focus on enhancing the imitation proofing of innovations for rapid commercialization [49]. Recent studies highlight the importance of effectively utilizing innovation appropriability tools, which help balance a company’s interests with access to external knowledge [50]. Attaining this balance can significantly enhance the benefits of open innovation [51] and improve the overall performance of the firm [52]. Specifically, when conducting external searches and collaborations, firms must manage their intellectual property rights effectively. This management promotes innovation while safeguarding its competitive advantage. Additionally, firms should employ various types of innovation exclusivity tools during external knowledge searches and collaborations. This approach fosters innovation while preserving their competitive edge.

However, analyzing the innovation environment from an external perspective reveals that China’s current property rights protection mechanism is still immature. Formal appropriability requires a robust legal system and efficient enforcement, both of which are currently lacking. Informal appropriability depends on the difficulty of imitation and the non-replicability of core technology. Yet, it faces challenges such as high management costs and irregular operational cycles [53]. Gaps in appropriability allows competitors the opportunity to obtain, copy, and utilize information about innovative products at a lower cost. This ultimately hinders the firms’ ability to recoup their investments in innovation. The theory of innovation appropriability effectively assists firms in gaining competitive advantage through unique technologies during digital transformation, especially in high-tech and sustainable sectors. This theory emphasizes that firms can gain market-based competitive advantage through intellectual property protection and unique innovations, thereby motivating companies to enhance their research and development efforts in digital and green innovation. However, in a rapidly changing digital environment, technological iterations and market demand fluctuations may undermine innovation appropriability’s sustainability. Additionally, an overemphasis on appropriability may cause organizations to overlook collaboration with external partners, consequently limiting the breadth and depth of innovation.

2.4. Literature Gaps

Despite scholars having examined the potential relationships among digital transformation, ambidextrous learning, green innovation, and innovation appropriability, significant research gaps persist in the existing literature. (1) There is a dearth of research on the precise mechanisms that connect digital transformation and green innovation. Most existing studies primarily examine the direct relationship between digital transformation and technology innovation [54]. However, they often overlook the influence of organizational ambidextrous learning theory as a potential mediating variable. This oversight can result in a degree of bias in the findings. (2) The research on the factors that motivate green innovation is still insufficient. The realization of green innovation often necessitates the synergistic effect of multiple factors. Existing literature primarily emphasizes the internal governance structure and external environmental policies that drive green innovation [55]. It does not adequately address the interactive roles of digital transformation and ambidextrous learning in promoting green innovation. While scholars have explored how ambidextrous learning facilitates technological innovation, the specific mechanisms and pathways of ambidextrous learning in the context of green innovation are still unclear. (3) The mechanism of innovation appropriability is insufficiently researched. Previous studies have confirmed the impact of innovation appropriability mechanisms on traditional innovation capabilities, such as open innovation. Therefore, the author suggests that these mechanisms may play a significant moderating role in the relationship between digital transformation, ambidextrous learning, and green innovation. This possibility has not been systematically explored in current research.

3. Theoretical Model and Hypothesis

3.1. Impact of Digital Transformation on Green Innovation

Enterprises undergoing digital transformation can improve productivity, conserve resources, reduce pollution, and enhance their capacity for green innovation [56]. This study posits that the influence of digital transformation on green innovation is mainly due to two fundamental aspects.

First, highly digitized companies are more inclined to adopt stringent environmental standards for both product and process improvements, which stimulates green innovation. As these companies improve their products and processes, they also develop robust environmental management capabilities. This development leads to the creation of innovative technologies that are difficult for competitors to replicate, fostering a competitive advantage in sustainability. Notably, green technology innovation must reach a specific threshold to demonstrate significant effects [57]. The integration of digital technology and green technology increases the complexity of industrial technology and creates technical barriers. It also expands and deepens the knowledge required for enterprises’ green innovation activities. Consequently, this dynamic also increases the learning costs and imitation barriers for competitors, slowing the spillover of proprietary knowledge. As a result, it enhances the firm’s core competitiveness and yields long-term benefits.

Second, big data technology—a core innovation from digital transformation—excels at integrating and analyzing large datasets [58]. This capability supports data-driven decision-making for green innovation. This capability not only accelerates the marketization and diffusion of green technologies but also enables enterprises to leverage data elements effectively. Moreover, digital transformation facilitates the integration of traditional and digital production factors. This approach attracts more intellectual capital and improves the efficiency of applying advanced technologies in product manufacturing. As a result, firms can shorten their R&D cycle and enhance their ability to integrate data to drive innovation, enabling them to connect with global innovation networks swiftly [58]. It is self-evident that mature big data technology can significantly benefit enterprises. By maximizing the use of data elements, it accelerates the implementation and marketing of green technologies, thereby enhancing their overall competitive advantage in green innovation. Consequently, based on the above analyses of the two ways in which digital transformation promotes green innovation, the following hypothesis is proposed:

H1. 

Digital transformation positively impacts green innovation.

3.2. Impact of Digital Transformation on Ambidextrous Learning

In the digital economy, digital technologies and tools help enterprises acquire and utilize tacit and heterogeneous knowledge more effectively, thereby enhancing ambidextrous learning within organizations. This study posits that the influence of digital transformation on ambidextrous learning is mainly due to two fundamental aspects.

First, digital transformation enhances exploitative learning by increasing the efficiency of knowledge integration, which in turn unlocks greater value from the organization’s existing knowledge and technology. Knowledge integration within an organization relies on a digital platform that effectively combines employees’ dispersed technologies and knowledge with the organization’s existing resources. This integration generates a new portfolio of knowledge that can be applied to internal business activities, creating value-added knowledge. Both organizational and external knowledge integration focus on accelerating the reconfiguration and leveraging of existing internal knowledge and learning [48]. Digital tools, such as videoconferencing and online learning platforms, help mitigate barriers to collaboration among companies across regions. They also minimize the loss of tacit knowledge during extraction and expedite the integration and reengineering of inter-firm knowledge. This process continually reconfigures existing technologies and enhances product manufacturing. As a result, digital transformation continuously reconfigures existing technologies and processes, leading to enhanced product manufacturing and improved operational efficiencies. This aligns the organization’s strategic objectives with its learning capabilities, ultimately driving sustained competitive advantage through exploitative learning.

Second, digital transformation fundamentally boosts exploratory learning. It does this by improving companies’ efficiency in exploring diverse knowledge. This improvement comes from enhanced information retrieval and analysis capabilities. As a result, organizations can better navigate and utilize a variety of knowledge sources. The knowledge-sharing platform developed through digital technology overcomes the spatial and temporal constraints of information exchange. It also addresses the challenges of knowledge transfer inherent in traditional organizational learning theory. This advancement enhances information openness and deepens the embedding of knowledge networks [59]. Thereupon, digital platforms connect companies with collaborative groups that possess diverse backgrounds and knowledge, broadening the scope of knowledge retrieval. This allows companies to mine and capture innovative insights through multiple searches. Additionally, digital technology is deeply integrated into business processes, enabling enterprises to leverage big data analytics to identify market foresight information. This integration helps companies explore and develop new technologies and products that align with market needs and fosters knowledge re-creation. Consequently, based on the above analysis of the impact of digital transformation on both aspects of ambidextrous learning, the following hypotheses are proposed:

H2a. 

Digital transformation positively impacts exploitative learning.

H2b. 

Digital transformation positively impacts exploratory learning.

3.3. Mediating the Effects of Ambidextrous Learning

Green innovation is a systematic and resource-intensive innovation activity undertaken by enterprises, grounded in knowledge integration. This framework involves the management and flow of knowledge in production, pollution reduction, and energy control. Its primary aim is to develop environmentally sustainable products and processes that foster sustainable development [60]. Ambidextrous learning theory internalizes key technological elements and diverse knowledge resources enabled by digital transformation. This process enables organizations to transform these elements into a competitive advantage by broadening existing pathways and exploring innovative routes for green innovation. This study posits that the mediating effects of ambidextrous learning are mainly due to two fundamental aspects.

First, exploitative learning plays a crucial role in promoting green innovation by enabling enterprises to identify and leverage new opportunities. Exploitative learning emphasizes the importance of enhancing the reliability of existing experiences and making continuous advancements on established foundations. Transformation enterprises identify and address their shortcomings by continuously refining and applying the tacit knowledge within their existing domains. They strive for excellence in their current practices. Moreover, optimized and upgraded green knowledge and technologies are continuously integrated into subsequent exploitative learning, enhancing enterprises’ sensitivity to green innovation resources. This enables them to identify potential external threats and effectively utilize their resources and capabilities to address these challenges. As a result, they can mitigate risks and facilitate the incremental development of green innovation.

Second, exploratory learning is essential for driving green innovation by transcending the limitations of existing experiences to explore new avenues. Exploratory learning emphasizes the necessity of transcending the limitations of existing experience to pursue innovation in new areas. Transformation firms integrate advanced green knowledge and achieve technological self-innovation and iteration by swiftly identifying the potential green needs of stakeholders and market trends. In doing so, they disrupt the path dependency of green innovation activities by actively seeking and generating new knowledge. Additionally, transforming enterprises leverage the technical advantages of digital simulation and digital twins. They implement simulation algorithms for the pre-coupling and restructuring of heterogeneous technological knowledge essential for green innovation. This approach enables them to accurately depict the processes of exploratory learning and green innovation in the digital realm. As a result, they can accelerate the innovation and application of core technologies, reduce trial-and-error costs, and mitigate the risk of uncertainties associated with green innovation. Ultimately, this commitment to exploratory learning leads to significant breakthroughs in green innovation development. Consequently, the following hypotheses are proposed:

H3a. 

Exploitative learning partially mediates the relationship between digital transformation and green innovation.

H3b. 

Exploratory learning partially mediates the relationship between digital transformation and green innovation.

3.4. Moderating Effects of Innovation Appropriability

Innovation appropriability influences the relationship between ambidextrous learning and green innovation in mainly two fundamental aspects. First, formal innovation appropriability mechanisms play a crucial role in enhancing the relationship between ambidextrous learning and green innovation. Formal appropriability mechanisms safeguard firms’ innovations through legal means, thereby reducing the market risks associated with these innovations. As a result, firms are encouraged to invest more resources and effort into the ambidextrous learning process. With robust intellectual property protection, companies can engage in both exploitative and exploratory learning. They do this without fear of their green innovations being easily imitated or stolen. Establishing formal ambidextrous appropriability mechanisms builds trust between firms. It also fosters collaborative innovation, providing a richer source of knowledge and practical opportunities for ambidextrous learning. This environment encourages firms to invest in and pursue green innovation. Ultimately, strong innovation appropriability establishes a stable and predictable environment, enabling firms to fully harness the positive effects of ambidextrous learning on green innovation.

Second, the informal appropriability mechanism acts as a vital strategy that enhances the positive relationship between ambidextrous learning and green innovation. It complicates imitation and enhances the marketability of these innovations, motivating firms to engage in deeper exploitative and exploratory learning. By relying on informal innovation appropriability, firms can effectively safeguard their patented technologies and know-how. This approach reduces the threat of imitation from competitors. The unique complementary assets and first-mover advantages associated with informal appropriability can rapidly transform green innovation concepts into products that resonate with consumers. This accelerates the implementation and market penetration of green innovations. It also helps companies stand out in a competitive market. Ultimately, informal innovation appropriability significantly strengthens the connection between ambidextrous learning and green innovation, fostering a climate where sustainable innovations can thrive and grow. Accordingly, the following hypotheses are proposed:

H4a. 

Innovation appropriability positively moderates the relationship between exploitative learning and green innovation.

H4b. 

Innovation appropriability positively moderates the relationship between exploratory learning and green innovation.

Based on the above hypotheses, we posit that innovation appropriability may moderate the mediating effect of ambidextrous learning on the relationship between digital transformation and green innovation. First, formal innovation appropriability mechanisms, such as patents, copyrights, and formal contracts, create strong barriers to knowledge protection. They also effectively curb the free-riding behavior among certain participants on knowledge-sharing platforms. This is particularly relevant for those who take unfair advantage of the openness of other members to exploit new knowledge [61]. Formal innovation appropriability serves as an indicator of a firm’s innovation strength. A high level of formal innovation appropriability encourages firms to proactively establish collaborative relationships with complementary partners that match their innovation capabilities. This collaboration facilitates the rapid clustering and widespread sharing of inter-organizational information and diverse data resources. As a result, enterprises can better understand user needs and business opportunities. By boosting the efficiency of ambidextrous learning, this understanding accelerates the advancement of green innovation.

Second, informal appropriability mechanisms, such as trade secrets and first-mover advantages, increase the exclusivity and complexity of replicating technologies. They effectively balance the threat of external imitation with the need for internal knowledge protection. Additionally, these mechanisms reduce the dependence of green innovation on data resources and digital technologies. High levels of informal appropriability are typically associated with a stronger brand image and robust marketing capabilities. These advantages are deeply rooted in the knowledge of experienced employees. This connection effectively reduces the risk of organizational knowledge leakage due to digital technological interconnectivity. As a result, informal appropriability broadens the scope of application and the profitability potential of a firm’s green innovation [62]. Conversely, a low level of appropriability makes it challenging for firms to capitalize on green innovation, hindering communication and collaboration with external partners. Additionally, it also reduces the frequency of exploitative and exploratory learning, ultimately impeding the green innovation process. Therefore, innovation appropriability facilitates knowledge aggregation and utilization. It enhances knowledge protection in digital transformation firms, thereby further promoting green innovation. Accordingly, the following hypotheses are proposed:

H5a. 

Innovation appropriability influences the relationship between digital transformation and green innovation by positively moderating the partial mediating effect of exploitative learning.

H5b. 

Innovation appropriability influences the relationship between digital transformation and green innovation by positively moderating the partial mediating effect of exploratory learning.

Building on the theoretical analyses and research hypotheses presented above, a theoretical model is constructed in this study, as illustrated in Figure 1. The model illustrates the intricate relationship between digital transformation, exploitative learning, exploratory learning, green innovation, and innovation appropriability. In addition, the specific path of the model is labeled with the nine research hypotheses proposed in this paper. The validity of these nine hypotheses will be further explored in subsequent research.

4. Research Methods

4.1. Sample Selection, and Data Collection

Traditional manufacturing enterprises are the primary contributors to carbon emissions. Thus, it is urgent for them to implement green innovation to achieve sustainable development. Therefore, this study focuses on Chinese manufacturing enterprises as the research subject. Data were collected through a questionnaire survey, utilizing validated questionnaires that were translated into Chinese following Brislin’s (1980) procedure [63]. Ethical considerations are essential in the questionnaire collection process. We provided a detailed explanation of the study’s purpose in the questionnaire, clarifying that participation was voluntary. Additionally, we assured participants that the collected questionnaire data would be securely stored and used exclusively for academic research purposes. All participants provided informed consent after reviewing the instructions. Since the managers are familiar with the basics of digital transformation, ambidextrous learning, green innovation, and innovation appropriability, middle and senior managers in manufacturing companies were chosen to complete the questionnaire. Two methods were used for distribution and collection. The surveyed enterprises encompass traditional manufacturing sectors, including chemicals, electronic components, building materials, energy, machinery, metals, paper, textiles, transportation, and others. Before formally distributing the questionnaire, starting in June 2024, we visited 20 traditional manufacturing firms from Beijing, Hebei, and Shanxi offline for pre-testing. We invited the middle and senior managers of these enterprises to fill in the questionnaire items. In addition, they were asked to suggest improvements in the setting of the question items and language presentation. Based on the feedback, the questionnaire items were revised and improved. Then, in October 2024, 80 paper questionnaires were distributed to MBA and MEM students at the author’s university who were employed in manufacturing enterprises, and all were collected on-site. Finally, 150 questionnaires were distributed to middle and senior managers of relevant manufacturing enterprises using online methods, including email and WeChat push notifications. After a week of follow-up with those who had not responded, 163 completed questionnaires were collected. By December 2024, a total of 250 questionnaires had been distributed, with 193 collected, resulting in a recovery rate of 77.2%. After eliminating non-standard questionnaires due to incomplete responses, missing information, and irregular submissions, 155 valid questionnaires remained. This resulted in an effective recovery rate of 62%, which satisfies the requirements for empirical data analysis. Figure 2 clearly expresses the research process of this study.

Table 1. Features of sample enterprises.

Variate	Characteristic	No.	%
Enterprise age	Within 3 years	56	36.1%
	4 to 5 years	43	27.7%
	6 to 10 years	25	16.1%
	11 to 20 years	20	12.9%
	More than 20 years	11	7.1%
Enterprise scale	Less than 50 people	37	23.9%
	51 to 100 people	48	31.0%
	101 to 300 people	37	23.9%
	301 to 500 people	8	5.2%
	More than 501 people	25	16.1%
Ownership of enterprise	State-owned business	69	44.5%
	Private enterprise	86	55.5%

illustrates the descriptive statistical findings for the samples. Notably, the majority of the responding enterprises were aged between 4 and 10 years, representing 43.8% of the total. These enterprises, at their developmental peak, were strategically positioning themselves for long-term success. They sought to maintain core competitiveness through two key strategies: digital transformation and green initiatives. These approaches are critical for sustainability. Regarding enterprise size, the majority were small and medium-sized enterprises (SMEs) with fewer than 300 employees, comprising 78.7% of the total. These SMEs exhibited high flexibility and could swiftly adjust their business strategies in response to external changes. The balanced representation of state-owned and private enterprises suggests that the selected samples have a degree of representativeness and applicability.

4.2. Variables Measurement

First, an initial questionnaire was created using well-established scales that scholars have repeatedly utilized domestically and internationally. Each variable in this study—digital transformation, ambidextrous learning, green innovation, and innovation appropriability—was measured using a five-point Likert scale. In this scale, the numbers 1 to 5 represent a gradual increase from “strongly disagree” to “strongly agree”. The appropriate adjustment of certain content guaranteed the scientific validity of the questionnaire results. This adjustment was based on existing maturity scales from both domestic and international sources. The changes considered the development status of traditional manufacturing enterprises and the objectives of the study.

Table 2. Variable measurement items.

Variable	Code	Subfactor	Reference
DT	DT1	Your company utilizes digital technology to transform and enhance existing goods, services, and processes.	Li F. [64] Zhou Y. et al. [65] Fachrunnisa et al. [66]
	DT2	Your company actively promotes digital design, manufacturing, and management.
	DT3	Your company develops digital products and services.
	DT4	Your company is dedicated to the dissemination and promotion of management knowledge and digital skills.
	DT5	There is a consensus within your company that adopting digital technology and management benefits its development.
EIL	EIL1	Your company aims to enhance the quality of existing products.	Atuahene-Gima et al. [67] Wang et al. [68]
	EIL2	Your company seeks greater flexibility in production.
	EIL3	Your company aims to reduce production costs.
	EIL4	Your company seeks to expand its presence in existing markets.
	EIL5	Your company aims to adapt its services to maintain customer satisfaction.
EAL	EAL1	Your company is working to introduce a new generation of products and services.	Atuahene-Gima et al. [67] Zhao et al. [69]
	EAL2	Your company is focused on expanding its range of products and services.
	EAL3	Your company is working to open new markets.
	EAL4	Your company is aiming to enter a new technological field.
	EAL5	Your company is seeking creative ways to meet customer needs.
GI	GI1	Your company’s energy-saving and emission-reduction rates are among the best in the business.	Chan et al. [70] Abro et al. [71]
	GI2	Green and innovative product sales are increasing as a proportion of total sales.
	GI3	Your company cultivates a positive social image through green innovation.
	GI4	Your company continuously develops green technologies and innovative products.
	GI5	Your company is continually enhancing its manufacturing processes to meet higher green production standards.
IA	IA1	Patents and copyrights can help your company grasp the rewards of innovation.	Gimenez-Fernandez et al. [51] Zobel et al. [72] Junior et al. [73]
	IA2	Obtaining trademarks can help your company grasp the rewards of innovation.
	IA3	Non-disclosure agreements can help your company grasp the rewards of innovation.
	IA4	Retaining key technical personnel by increasing their salaries can help your company grasp the rewards of innovation.
	IA5	Encrypting documents containing sensitive company information can help your company grasp the rewards of innovation.
	IA6	Encrypting documents containing sensitive company information can help your company grasp the rewards of innovation.

Notes: DT, digital transformation; EIL, exploitative learning; EAL, exploratory learning; GI, green innovation; IA, innovation appropriability.

presents the measurement items for each variable.

Additionally, firm age, firm size, and nature of ownership were established as control variables. Specifically, the age of the firm was categorized into five classes based on the year of establishment; the nature of ownership included two categories: state-owned enterprises and non-state-owned enterprises; and firm size was classified into five categories based on the number of employees.

5. Empirical Analysis Results

5.1. Reliability and Validity Test

SPSS 27 was employed to conduct a reliability test of the scale. The results are presented in

Table 3. Variable reliability and validity analysis results.

Variate	Item	Load	Cronbach’s α	CR	AVE
Digital transformation	DT1	0.853	0.839	0.896	0.638
	DT2	0.806
	DT3	0.827
	DT4	0.679
	DT5	0.807
Exploitative learning	EIL1	0.842	0.804	0.890	0.617
	EIL2	0.766
	EIL3	0.737
	EIL4	0.785
	EIL5	0.795
Exploratory learning	EAL1	0.817	0.853	0.895	0.630
	EAL2	0.792
	EAL3	0.826
	EAL4	0.786
	EAL5	0.746
Green innovation	GI1	0.747	0.802	0.867	0.567
	GI2	0.698
	GI3	0.753
	GI4	0.742
	GI5	0.819
Innovation appropriability	IA1	0.856	0.853	0.921	0.660
	IA2	0.795
	IA3	0.752
	IA4	0.855
	IA5	0.841
	IA6	0.769

. The Cronbach’s α values for the five variables, including digital transformation, exceeded 0.8, while the overall Cronbach’s α for the questionnaire was 0.9. This implies that the questionnaire was reliable and appropriate for subsequent analysis.

For the validity test, both convergent validity and discriminant validity tests were conducted. The scales utilized in this study are well-established both domestically and internationally. Amos28 was implemented for confirmatory factor analysis. Table 3 illustrates the results. The composite reliability (CR) surpassed 0.8, and the average variance extracted (AVE) exceeded 0.5. Additionally, the factor loadings for all measurement elements surpassed 0.6. These results suggest that the measurement scale exhibits good convergent validity. Furthermore, the results in

Table 4. Model fitting test results.

Model	Combination	Χ2	df	Χ2/df	TLI	CFI	RMSEA
one-factor model	DT + EIL + EAL + GI + IA	1005.364	300	3.351	0.534	0.570	0.124
two-factor model	DT, EIL + EAL + GI + IA	869.090	301	2.887	0.626	0.653	0.111
three-factor model	DT, EIL, EAL + GI + IA	846.496	300	2.822	0.639	0.667	0.109
four-factor model	DT, EIL, EAL, GI + IA	782.060	298	2.624	0.678	0.705	0.103
five-factor model	DT, EIL, EAL, GI, IA	433.077	293	1.478	0.905	0.915	0.056

indicate that the five-factor model exhibited the best fit (χ²/df = 1.478, TLI = 0.905, CFI = 0.915, RMSEA = 0.056). This model significantly outperforms other alternative factor models, demonstrating the good discriminant validity of the measurement model.

5.2. Common Method Bias

The paper may encounter a common method bias (CMB) issue if it utilizes the same data source. To address this concern, this paper implements strict design controls both in advance and during post-testing to minimize the potential CMB. During the questionnaire collection, this paper guarantees anonymity for all respondents, emphasizing that there are no right or wrong answers to encourage honest responses. Additionally, independent and dependent variables are strategically placed in separate sections to achieve appropriate psychological separation and ensure data objectivity. In the post hoc analysis, the Harman single-factor test is employed. The results indicate that the cumulative contribution rate of the first factor is 30.04%, which is below the critical threshold of 40%. This finding suggests that common method bias is unlikely to significantly affect the findings of this study.

5.3. Correlation Analysis

A correlation analysis of digital transformation, exploitative learning, exploratory learning, green innovation, and innovation exclusivity is conducted using SPSS 27.0. The results are presented in

Table 5. Correlation analysis of variables.

Variate	MV	SD	1	2	3	4	5
DT	3.857	0.779	1
EIL	3.679	0.623	0.348 ***	1
EAL	3.690	0.760	0.296 ***	0.620 ***	1
GI	3.514	0.683	0.500 ***	0.538 ***	0.607 ***	1
IA	4.265	0.623	0.222 **	0.124	0.110	0.168 *	1

Notes: * p < 0.05, ** p < 0.01, *** p < 0.001.

. At a significance level of 0.001, digital transformation and green innovation are significantly positively correlated (r = 0.500). Digital transformation exhibits a strong positive relationship with exploitative learning (r = 0.348) and exploratory learning (r = 0.296). Exploitative learning (r = 0.538) and exploratory learning (r = 0.607) also demonstrate strong positive relationships with corporate green innovation. Additionally, the correlation coefficients among the main variables are less than the critical value of 0.7. Clearly, the relationships among the variables align with the proposed hypotheses, providing preliminary support for subsequent hypothesis testing.

5.4. Hypothesis Testing

5.4.1. Main and Mediation Effects Tests

In this study, digital transformation is posited to facilitate corporate green innovation (H1). This direct effect was rigorously tested using the statistical software SPSS 27.0 with a stratified linear regression approach. First, we accounted for the control variables that may impact corporate green innovation, including firm age, firm size, and form of ownership. In this analysis, digital transformation serves as the independent variable, while green innovation is the dependent variable.

Table 6. Regression results for the direct effect and mediation effects.

Variate	EIL	EAL	GI
	M1	M2	M3	M4
age	0.013	0.078	−0.079	−0.113
size	0.031	0.015	0.098	0.087
ownership	0.021	0.041	0.124	0.104
DT	0.333 ***	0.271 ***	0.470 ***	0.302 ***
EIL				0.173 **
EAL				0.408 ***
R2	0.123	0.096	0.270	0.519
ΔR2	0.099	0.072	0.251	0.500
F	5.251 ***	3.966 ***	13.879 ***	26.657 ***
Max VIF	1.473	1.473	1.473	1.709

Notes: ** p < 0.01, *** p < 0.001.

provides a comprehensive presentation of the regression results. The analysis of Model 3 indicates that digital transformation has a positive and strong relationship with corporate green innovation (β = 0.470, p < 0.001). This finding verifies hypothesis H1, confirming the essential role of digital transformation in promoting sustainable practices.

To test the mediating effect of ambidextrous organizational learning theory, we first conducted an initial test using the hierarchical regression method proposed by Baron and Kenny (1986) [74]. Subsequently, the findings were further validated using the Bootstrap method. The regression outcomes are presented in Table 6. Initially, the regression analysis results from Models 1 and 2 indicated that digital transformation has a significant positive effect on both exploitative learning (β = 0.333, p < 0.001) and exploratory learning (β = 0.271, p < 0.001). This supports hypotheses H2a and H2b. The regression analysis results from Model 3 demonstrate that digital transformation has a significant positive impact on corporate green innovation (β = 0.470, p < 0.001). Model 4 incorporates ambidextrous learning as a mediating variable based on Model 3. The analysis reveals that both exploitative learning (β = 0.173, p < 0.01) and exploratory learning (β = 0.408, p < 0.001) have a significant positive impact on corporate green innovation. Additionally, digital transformation maintains a substantial direct effect on corporate green innovation (β = 0.302, p < 0.001). Therefore, exploitative learning and exploratory learning positively mediate the relationship between digital transformation and green innovation, serving as partial mediators. This provides tentative support for hypotheses H3a and H3b.

To further validate the significance of the mediated paths, the Bootstrap method was employed. The sample size was set at 5000 with a confidence level of 95%. Additionally, bias-corrected nonparametric percentiles were used as the sampling method. The results are summarized in

Table 7. Results of bootstrap analysis of mediation effect.

Path Relationship	Effect Type	Effect Size	SE	LLCI	ULCI
DT → EIL → GI	Total effect	0.413	0.066	0.282	0.543
	Direct effect	0.292	0.062	0.169	0.415
	Indirect effect	0.120	0.048	0.041	0.228
DT → EAL → GI	Total effect	0.413	0.066	0.282	0.543
	Direct effect	0.292	0.057	0.181	0.404
	Indirect effect	0.120	0.045	0.041	0.217

. With exploitative learning as a mediating variable, the indirect effect of digital transformation on firms’ green innovation was significant, with a confidence interval of [0.041, 0.228] that excludes zero. Its direct effect was also significant, with a confidence interval of [0.169, 0.415] that excluded zero. This suggests that exploitative learning plays a partially positive mediating role. Similarly, after adding exploratory learning as a mediating variable, the indirect effect of digital transformation on corporate green innovation remained significant. The confidence interval for this effect was [0.041, 0.217], which excluded zero. Its direct effect was also significant, with a confidence interval of [0.181, 0.404] that did not include zero. This indicates that exploratory learning also plays a partially positive mediating role. In summary, hypotheses H3a and H3b are further supported.

5.4.2. Moderating Effects Test

Cascade regression analysis was utilized to assess the moderating influence of innovation appropriability in the research models. Firstly, we tested the effect of ambidextrous learning on green innovation. In this analysis, green innovation was treated as the dependent variable. We included exploitative learning and exploratory learning as the independent variables in Models M5 and M8, respectively. Next, innovation appropriability was introduced as the moderating variable in Models M6 and M9. Finally, we added the interaction terms of exploitative learning and innovation appropriability into Model M7, and the interaction terms of exploratory learning and innovation appropriability into Model M10 for testing. Regression analyses were then conducted, with the results presented in

Table 8. Regression results for moderation effects.

Variate	EIL			EAL
	M5	M6	M7	M8	M9	M10
age	−0.068	−0.068	−0.062	−0.107	−0.106	−0.107
size	0.182 *	0.181 **	0.143	0.193 **	0.192 **	0.159
ownership	0.164 *	0.156 *	0.111	0.153 *	0.144 *	0.100
EIL	0.510 ***	0.499 ***	0.492 ***
EAL				0.586 ***	0.577 ***	0.536 ***
IA		0.090	0.008		0.090	−0.034
EIL × IA			0.444 ***
EAL × IA						0.369 ***
R2	0.333	0.341	0.527	0.413	0.421	0.534
ΔR2	0.315	0.318	0.508	0.397	0.401	0.516
F	18.701 ***	15.394 ***	27.501 ***	26.356 ***	21.637 ***	28.320 ***
Max VIF	1.366	1.366	1.374	1.356	1.356	1.366

Notes: * p < 0.05, ** p < 0.01, *** p < 0.001.

. To reduce multicollinearity between the variables, the independent and moderating variables were centralized. The results revealed that innovation appropriability significantly moderates the relationship between exploitative learning and green innovation (β = 0.444, p < 0.001). It also plays a moderating role in the relationship between exploratory learning and green innovation (β = 0.369, p < 0.001). These findings preliminarily support hypotheses H4a and H4b. To elucidate the moderating effect of innovation appropriability more clearly, a simple slope analysis was conducted. Regression analyses of ambidextrous learning and green innovation were performed under different grouping conditions. Corresponding moderating effect diagrams were also generated. Figure 3 illustrates that the slope for high levels of innovation appropriability is steeper than the slope for low levels of innovation appropriability. This suggests that as the degree of innovation appropriability increases, the contribution of exploitative learning to green innovation becomes more pronounced. Similarly, Figure 4 demonstrates that the slope for high levels of innovation appropriability is steeper than that for low levels. This indicates that as the degree of innovation appropriability increases, the facilitating effect of exploratory learning on green innovation becomes more significant. These findings further support hypotheses H4a and H4b.

5.4.3. Moderated Mediation Effect Test

This study examined the moderated mediating role of innovation appropriability in the relationship among digital transformation, ambidextrous learning, and corporate green innovation. The analysis was conducted using Bootstrap testing in the PROCESS framework. First, 5000 sampling repetitions were conducted to ensure robust results, and 95% confidence intervals were constructed. Additionally, bias-corrected nonparametric percentiles were used as the sampling method. The model included the dependent variable (green innovation), the independent variable (digital transformation), the mediator variables (exploitative learning and exploratory learning), and the moderator variable (innovation appropriability). The final test results are presented in

Table 9. Results of bootstrap analysis with moderated mediating effect.

	Indirect Effect	SE	LLCI	ULCI
Index of moderated mediation (via exploitative learning)	0.231	0.075	0.088	0.382
Conditional indirect effect at innovation appropriability = M ± 1 SD
M + 1 SD	0.263	0.085	0.102	0.438
M − 1 SD	−0.024	0.022	−0.071	0.015
Index of moderated mediation (via exploratory learning)	0.164	0.060	0.057	0.289
Conditional indirect effect at innovation appropriability = M ± 1 SD
M + 1 SD	0.210	0.073	0.078	0.357
M − 1 SD	0.006	0.019	−0.030	0.049

. The findings revealed that the indirect effect of “digital transformation → exploitative learning → green innovation” is insignificant when the level of innovation appropriability is low. This is shown by confidence intervals of [−0.071, 0.015], which include zero. Conversely, when the level of innovation appropriability is high, the indirect effect of “digital transformation → exploitative learning → green innovation” is significant. The confidence interval for this effect is [0.102, 0.438], which does not include zero. Additionally, the moderated mediation effect is significant, with a determination index of 0.231 and a confidence interval of [0.088, 0.382] that excludes zero. These results indicate that innovation appropriability significantly moderated the mediation effect of exploitative learning, thus supporting hypothesis H5a.

At low levels of innovation appropriability, the indirect effect of digital transformation → exploratory learning → green innovation is not significant due to the mediating effect of exploratory learning. This is indicated by confidence intervals of [−0.030, 0.049], which include zero. Conversely, at high levels of innovation appropriability, the indirect effect of digital transformation → exploitative learning → green innovation is significant. The confidence intervals for this effect are [0.078, 0.357], which do not include zero. Additionally, the moderated mediation effect is significant, with an indicator of 0.164 and a confidence interval of [0.057, 0.289] that excludes zero. These results indicate that innovation appropriability significantly moderates the mediating effect of exploratory learning, thus supporting hypothesis H5b.

6. Conclusions, Discussion, Implications, and Research Limitations

6.1. Conclusions

This study examined traditional Chinese manufacturing firms to determine how digital transformation influences green innovation. The model of this study possesses good reliability and validity. In addition, this model fits significantly better than the alternative model (χ²/df = 1.478, TLI = 0.905, CFI = 0.915, RMESA = 0.056). Scientific methodologies were used to evaluate the mediating role of ambidextrous learning and the moderating role of innovation appropriability. The findings indicate that digital transformation significantly contributes not only to green innovation but also to exploitative and exploratory learning. Digital transformation and green innovation are partially mediated by exploratory and exploitative learning. Furthermore, innovation appropriability boosts the contributions of exploitative and exploratory learning to green innovation. It also enhances their mediating roles between digital transformation and green innovation. An in-depth discussion follows in response to the conclusions drawn in the study. The final part of this paper discusses the limitations and potential consequences in detail. It highlights the dynamic relationship between digital transformation and green innovation. Additionally, it addresses other potential moderating variables and pathways. These aspects are identified as important areas for future research.

6.2. Discussion

(1) The main effect of the research model was significant, suggesting that digital transformation serves an essential function in fostering green innovation. This aligns with the findings of Vial (2021) and Feng (2025), who discuss the relationship between digital transformation and green innovation [6,31]. Specifically, digital transformation prompts firms to participate in green innovation, thereby reinforcing the validity of our hypothesis. Its energy-saving and sustainable features align closely with the goals of green innovation. Therefore, enterprises must integrate digital transformation into their green innovation processes. This integration helps reduce the costs associated with searching for and utilizing the necessary knowledge for green initiatives. It also maximizes the innovative potential of digital transformation. Importantly, this process is not driven by a single factor [55]; thus, other factors that may affect green innovation must also be taken into account. In reality, through the release of the “Overall Layout Plan for the Construction of Digital China”, the Chinese government aims to promote green innovation by improving facilities for digital transformation. This initiative aims to create a greener economy and promote sustainable development. Our research contributes to existing theories of digital transformation and green innovation, highlighting how these areas are interconnected. Additionally, it offers practical guidance for companies to effectively use digital transformation in their sustainability efforts.

(2) The mediating effect of ambidextrous learning in the research model was confirmed. Both exploitative and exploratory learning serve as positive partial mediators in the relationship between digital transformation and green innovation. As shown in Table 6, the mediation effect of exploitative and exploratory learning each accounts for 70.7%. This indicates that both types of learning are critical in understanding how digital transformation influences green innovation, underscoring the importance of balancing exploitative and exploratory learning. In environments where organizations frequently interact with external entities, firms that effectively leverage existing knowledge while actively seeking new insights are more likely to pursue green innovation. Additionally, companies should utilize advanced digital technologies to support these ambidextrous learning activities. These findings build upon the research conducted by Shao (2022) and Chen (2024) [45,50], who demonstrated the critical role of ambidextrous learning in organizational adaptability and innovation within dynamic business environments. Correspondingly, the Chinese government’s “Outline of the National Medium- and Long-Term Scientific and Technological Development Plan” actively promotes firms’ engagement in both exploitative and exploratory learning strategies to foster green innovation.

(3) The moderating effect of innovation appropriability in the research model was confirmed. Innovation appropriability significantly moderates the relationship between exploitative and exploratory learning and green innovation. It enhances the positive effects of both types of learning on green innovation. As shown in Table 9, when the level of innovation appropriability is high, it strengthens the mediating role of exploitative and exploratory learning in the relationship between digital transformation and green innovation. In contrast, when the level of innovation appropriability is low, the moderated mediation effect is not significant. These results further extend the findings of Gimenez-Fernandez (2023) and Stefan (2017) [51,52]. Therefore, establishing a corporate innovation appropriability mechanism is crucial. It not only protects the gains from green innovation but also fosters its development through ambidextrous learning. Indeed, China is enhancing innovation appropriability through various policies, including strengthening intellectual property protection and promoting technology transfer. These measures foster an environment that encourages firms to engage in both exploitative and exploratory learning, thereby enhancing their green innovation capabilities.

(4) Given the challenges of global environmental degradation and resource depletion, promoting green innovation is essential. Sustainable development has become a central concern for all countries. The United Nations Sustainable Development Goals (SDGs) emphasize the importance of green innovation and achieving long-term sustainable development relies heavily on it. This study introduces a theoretical model designed to promote green innovation, based on research conducted with traditional manufacturing enterprises in China. The findings demonstrate significant potential for replication in other countries and industries.

On the one hand, digital transformation offers advanced tools and effective data support to enterprises worldwide. However, it also presents many challenges. For instance, Germany has made notable progress in cleaner production and resource efficiency since adopting Industry 4.0 and smart grid policies. Yet, there is still pressure for continuous innovation and technology integration. Thus, manufacturing and energy companies have the capacity and demand for utilizing ambidextrous learning. This approach helps them to balance the improvement of existing processes with the development of new technologies that enhance green innovation. In the service sector, especially in finance and consulting, digital transformation improves service efficiency and customer experience. Firms can use ambidextrous learning to develop and enhance green financial products. This practice promotes sustainable development while allowing them to expand their markets.

On the other hand, the moderating role of the innovation appropriability mechanism in the model deserves further exploration. Innovation appropriability is particularly vital for firms with limited resources, as an effective means of ensuring reasonable innovation returns. It not only helps these firms maximize their return on investment, but also attracts external resources by showcasing their technological capabilities and profit potential. As a result, this mechanism enhances their sustainability effectively. Furthermore, firms in different countries or industries can adapt their exclusivity strategies. They need to align these strategies with local institutional environments and market characteristics. For example, in countries with strong institutions (such as the United States, Germany, and China), firms can use formal exclusivity, such as patents to protect their green innovations. In contrast, in countries with more open market environments (such as India and Brazil), firms can rely on informal exclusivity. This can help them to sustain market dominance through industry alliances or informal cooperation. In other words, the innovation appropriability mechanism can only be maximized when it is customized to fit the local context.

6.3. Implications

Combining the research findings provides a summary of this study’s implications. We propose implications for government policymakers and enterprises managers based on three main conclusions.

(1) For government policymakers, strategic intervention is critical in enterprise digital transformation. Targeted policy initiatives can effectively accelerate green innovation within manufacturing sectors. Businesses must integrate fully with green technology innovation as they progress in their digital transformation in order to attain sustainability.

Firstly, the government should establish a framework that aligns enterprise digital strategies with green innovation. This framework should include targeted subsidies and preferential policies in finance, taxation, and talent.

Secondly, the government should promote investments in green technology and innovation by creating a targeted subsidy policy for digital transformation. This policy should link subsidy amounts to the enterprise’s green performance, ensuring effective fund utilization. Additionally, the government could offer extra loan incentives for digital transformation projects that comply with green standards, thereby steering enterprises toward sustainable development during their transformation.

Thirdly, the government should adopt various strategies to assist enterprises in establishing mechanisms for innovation appropriability. To enhance formal appropriability, the government should strengthen patent regulations to ensure adequate protection of enterprises’ intellectual property rights and prevent the illegal acquisition of their innovations. Additionally, to foster informal exclusivity, the government should encourage the establishment of industry_university_research collaboration platforms to facilitate technology accumulation and enhancement in enterprises. The government should support industry associations and networking platforms to establish innovation collaboration networks, facilitating information sharing and knowledge dissemination among enterprises. Through these measures, enterprises can sustain their competitive edge in green innovation, thereby achieving their sustainable development goals.

Finally, these measures will promote the smart transformation and digitalization of enterprises. By implementing these policy measures, the government can effectively motivate enterprises to integrate green innovation concepts into their digital transformation efforts. This approach not only promotes sustainability but also encourages a more environmentally responsible business model.

(2) For enterprises managers, they should implement strategic measures focusing on digital transformation, ambidextrous learning, and robust innovation appropriability mechanisms to promote green innovation. Firstly, it is important to integrate digital strategies throughout the entire process of green technology innovation. Enterprises should promote the close integration of digital technology with green innovation. They must continue to utilize digital technology’s role in this process. Enhancing their technological integration capabilities is crucial, as it will help unlock the potential of data elements and improve digital productivity. Ultimately, these efforts will advance green technological innovation in both quantity and quality.

Secondly, enterprises must strengthen ambidextrous learning to continuously enhance their capacity for green innovation. They should adopt a “two-pronged” strategy that emphasizes both exploitative and exploratory learning. Enterprises should regularly organize internal sharing sessions and knowledge competitions to reward employees for their achievements in exploitative learning. Simultaneously, enterprises should establish a comprehensive cloud-based digital learning platform, covering online course learning, digital training equipment, virtual simulation experiments and other digital tools. These platforms will facilitate targeted and systematic learning related to green innovation for employees outside of working hours. Maintaining close contact with suppliers and customers is also essential, as it helps organizations stay informed about customer needs and engage in exploitative learning. Next, enterprises should focus on enhancing exploratory learning. For instance, they can form alliances with partners to gain insights into market dynamics and competitive information. This collaboration will allow them to integrate knowledge and resources to develop new green products. Moreover, companies should encourage employees to engage in learning exchanges at universities or research institutes. This will help them acquire cutting-edge green scientific knowledge.

Thirdly, enterprises should effectively utilize innovation appropriability tools to foster green innovation. On the one hand, when firms encounter challenges in pursuing green innovation independently, they can leverage their patents and copyrights. These intellectual properties serve as signals to showcase their potential for green innovation. By doing this, firms can attract external partners to engage in green innovation research. Additionally, through digital platforms and big data analytics, firms can identify potential partners with complementary knowledge. This collaboration can help develop new business opportunities. By leveraging formal appropriability, firms can strengthen their capacity for green innovation. On the other hand, enterprises can selectively disclose and share core technologies that have a first-mover advantage and are hard to replicate in the short term. They can be achieved through value networks and ecosystems established on digital platforms. Such a strategy not only expands the firm’s market share but also enhances its application in knowledge areas such as standards, norms, and experience. This approach allows firms to capture monopoly profits from their core technology while securing a favorable position in the development of industrial technology standards. Consequently, this enhances the market position of green innovations based on informal innovation appropriability.

6.4. Research Limitations and Future Research Directions

Grounded in organizational learning theory and profit from innovation theory, this study examined the connections among digital transformation, ambidextrous learning, green innovation, and related innovation appropriability variables. This examination was conducted through both theoretical analysis and empirical validation. However, several limitations are acknowledged that present opportunities for future investigation. Primarily, this study employed cross-sectional data, which limits its ability to capture the temporal effects of ambidextrous learning on green innovation. Future research could utilize time series survey data, conducting questionnaires across different time periods and expanding the sample size. This approach aims to enhance the generalizability and objectivity of the results and will help reveal the dynamic relationship between digital transformation and green innovation. In addition, robustness testing was not included in the design of the questionnaire for this study. In future studies, items for substitute and control variables could be added to the questionnaire. Additionally, distributing questionnaires across different regions would help differentiate samples, enabling robustness testing to be carried out to further enhance the reliability of the results. Moreover, this study developed a mechanism of “digital transformation → ambidextrous learning → green innovation.” Only the moderating effect of innovation appropriability on the relationship between ambidextrous learning and corporate green innovation was examined. This limitation may hinder the model from accurately reflecting real-world situations and could affect the validity of the findings for informing management practices. To overcome this issue, future research should incorporate additional moderating variables. For example, factors such as environmental regulation, leadership style, and corporate cultural orientation should be considered. Researchers might also explore other potential pathways. By doing so, the theoretical model will be enhanced. The study could also be extended to include service, financial, and energy firms to examine the applicability and variability of the findings in different contexts. This will enable a more comprehensive exploration of how digital transformation drives corporate green innovation through its influence on ambidextrous learning.