Crossing the Valley of Death: The Mechanism Through Which Searching Drives Green Product Development

Abstract

Many green technology innovations developed by academic institutions struggle to cross the “Valley of Death” (VoD), failing to achieve commercialization. Boundary-spanning green technology search (BGTS), as an inter-organizational mechanism facilitating the interaction of green technology knowledge between firms and external stakeholders, can bridge the gap between green scientific research and the commercialization of green products. Drawing on the resource orchestration theory and recombinant search theory, this study empirically analyzed data from 313 Chinese manufacturing enterprises to explore the pathway through which BGTS promotes green product development performance (GPDP) and examines the chain mediating role of knowledge coupling and green technology commercialization capability (GTCC), as well as the moderating role of digital technology adoption and product complexity. The findings reveal that the relationship between BGTS and GPDP is sequentially mediated first by knowledge coupling and then by GTCC. Digital technology adoption positively moderates the BGTS-GPDP relationship. Product complexity moderates the BGTS-GPDP relationship in an inverted U-shape. This study elucidates the micro-mechanism underlying the commercialization of green technologies, providing both theoretical insights and practical guidance for the green and high-quality transformation of manufacturing enterprises.

1. Introduction

As green development becomes a global priority encompassing environmental protection, social welfare, and economic prosperity [1], many countries are encouraging enterprises to undergo green transformation to enhance their competitiveness [2]. The manufacturing industry, as the primary driver of energy conservation and carbon reduction, plays a pivotal role in this transition toward a green, low-carbon and circular economic development system.

China’s manufacturing sector is currently undergoing a strategic transition toward green high-quality development. Green products, a key component of green manufacturing systems, can effectively reduce environmental pollution and promote sustainable development [3]. Enhancing green product development and ensuring effective supply have become critical priorities for manufacturing enterprises’ green transformation [4]. Scientific and technological achievements offer significant opportunities for new product development [5]. However, a substantial gap often exists in translating these achievements into marketable products—a phenomenon known as the “Valley of Death” (VoD).

Compared with traditional innovation, the VoD in green innovation exhibits more complex characteristics. Green innovation typically involves higher technological complexity and longer R&D cycles, encounters greater market uncertainty, and relies more heavily on policy support and financial investment [6]. Consequently, despite the rapid growth of green scientific and technological achievements in China’s manufacturing industry, many fail to be commercialized and translated into tangible productivity, leaving the value of green technology innovation largely unrealized. The VoD disrupts the continuity of the green innovation value chain [7,8].

The VoD refers to the gap between basic scientific research conducted at academic institutions (e.g., universities, research institutes) and the development of commercial products by firms [5]. VoD research has primarily focused on the commercialization of scientific research [9], particularly knowledge transfer between academia and industry [10]. Existing literature recognizes that bridging the gap between scientific research and commercial product development—that is, facilitating the transition from scientific logic to market logic—is crucial for crossing the VoD [11,12,13]. However, few studies have examined the specific mechanisms underlying this transition in depth. Boundary-spanning green technology search (BGTS), as a critical mechanism for firms to acquire external green technology knowledge, facilitates knowledge transfer between academic institutions and enterprises and promotes the commercialization of green scientific research [14]. Unlike traditional organizational search which primarily focuses on knowledge acquisition, BGTS not only acquires technological knowledge but also inherently responds to environmental sustainability imperatives. By systematically integrating scientific and commercial logics, BGTS extends traditional organizational search theory within the context of sustainable development, thereby providing a crucial mechanism for crossing the VoD.

Drawing on resource orchestration theory and recombinant search theory, this study positions BGTS as a critical nexus linking green scientific research and green commercial product development, attempting to unpack the “black box” of BGTS’s impact on green product development performance (GPDP). First, following the “resource structuring–capability building–performance enhancement” framework of resource orchestration theory [15], this study identifies knowledge coupling and green technology commercialization capability (GTCC) as mediating variables that explain how BGTS enhances GPDP. BGTS enables firms to acquire advanced green technology resources, thereby increasing the opportunities for technology knowledge coupling. Knowledge coupling is an interactive process through which a firm’s internal and external green technology knowledge elements achieve complementarity, compatibility, and synergy [16], facilitating the establishment of a firm’s green technology resources system and laying the foundation for green technology commercialization capability. This capability accelerates the market realization of green technologies [17], ultimately enhancing GPDP. Second, organizational behavior research suggests that opportunity factors reflect the extent to which a situation facilitates achieving expected outcomes or creates obstacles, thus influencing firm performance [5]. This study introduces two opportunity factors—digital technology adoption and product complexity—to construct an integrated research framework of “technology–capability–opportunity–performance”, exploring the boundary conditions of BGTS’s impact on GPDP. Digital technology adoption is fundamental to firms’ green and low-carbon development [18], providing greater opportunities for acquiring, accumulating, and deploying green technologies, while accelerating green technology search, recombination and transformation into green products [19]. Product complexity, which represents the quantity of technological components and their interdependencies, serves as a key determinant of search efficiency and outcomes [20], thereby moderating the BGTS–GPDP relationship.

This study makes several key contributions. First, within the context of crossing the VoD in green technology innovation, it identifies BGTS as a unique form of search that integrates environmental, scientific, and market value, and positions it as a crucial channel for bridging the gap between green scientific research and product commercialization. By incorporating inter-organizational technology knowledge transfer into VoD research, this study offers novel theoretical insights for VoD research. Second, this study develops a “technology–capability–opportunity–performance” framework, providing a comprehensive theoretical structure that integrates resource orchestration theory and recombinant search theory into green innovation research. Through this framework, the study elucidates the micro-mechanisms through which BGTS drives GPDP, providing both theoretical and practical guidance for engineering managers seeing to advance the sustainable transformation of the manufacturing industry.

2. Theoretical Background and Hypotheses

2.1. Valley of Death

The term VoD originally referred to the stagnation in technological progress caused by funding and support deficiencies, which prevent the transition from the invention to innovation [21]. Subsequent research extended the VoD concept to include the financial challenges experienced by nascent companies, or the gap between scientific research and commercial product development [9,22]. This study defines VoD as the gap between scientific research conducted by academic institutions and commercial product development undertaken by firms [5].

Prior research has extensively documented that VoD significantly hinders the commercialization of basic research outcomes and serves as a major barrier to radical innovations both at the firm and national levels [7]. Researchers have explored various mechanisms for crossing the VoD, including securing additional funding [21], resolving goal conflicts between academic institutions’ scientific and technological inventions and firms’ product development objectives [23], overcoming disconnections between R&D departments and other organizational units, and leveraging key individuals such as entrepreneurs and product champions [8,24].

However, current VoD research has devoted limited attention to green innovation, particularly the role of BGTS in bridging the gap between green scientific research and green product development in the context of green technology innovation. This study seeks to fill this gap by developing a “knowledge–capability–opportunity–performance” conceptual framework.

2.2. Relationship Between Boundary-Spanning Green Technology Search (BGTS) and Green Product Development Performance (GPDP)

Recombinant search theory conceptualizes technological evolution as a process of recombination of prior and/or new technologies [25]. The mechanism through which BGTS drives green product development aligns with this process. BGTS targeting academic institutions enables firms to acquire diverse and cutting-edge green technology knowledge, creates more recombination opportunities between new and existing technologies, enhances the diversity of green technologies [26] and expands the firm’s green technology knowledge base [27]. Such knowledge serves as the foundation for firms’ green value creation, providing more options to generate high-quality product ideas [28] and develop innovative products [4].

Original green scientific research is primarily concentrated in universities, research institutes, and other academic institutions [29]. Moreover, compared to traditional products, the R&D of green products often requires higher technological investments and more complex technical support. For firms, transitioning from green technology R&D to green product deployment typically requires a prolonged accumulation process. BGTS targeting academic institutions allows firms to acquire cutting-edge green technology more rapidly, share green product innovation costs, mitigate green product development risks, shorten green product R&D cycles, and accelerate green product development.

Additionally, BGTS fosters green technology collaboration and information exchange between firms and external stakeholders, expanding their pool of potential partners and facilitating the formation of informal inter-organizational networks. This increases opportunities for discovering novel green technologies and developing innovative green products, enabling firms to transcend existing business and product boundaries while accessing emerging green product markets [30].

Thus, we hypothesize the following:

Hypothesis 1.

BGTS promotes GPDP.

2.3. Serial Mediating Effect of Knowledge Coupling and Green Technology Commercialization Capability (GTCC)

2.3.1. Mediating Effect of Knowledge Coupling

Recombinant search theory posits that knowledge from diverse domains generates value through interaction and feedback mechanisms [31]. BGTS targeting academic institutions provides firms with the technology knowledge necessary for green product development [32]. However, academic institutions and firms operate under distinct institutional logics, resulting in inherent differences in their respective norms and cultural practices [33]. Effectively managing or bridging these logical differences is crucial to prevent them from obstructing the transfer of technological knowledge [34]. Knowledge coupling facilitates the integration of heterogeneous green technology knowledge within corporate relational networks, enriching the technology knowledge base and enhancing its specialization [35], thereby establishing a strong foundation for green product development.

Knowledge coupling enables the interpenetration, interconnection, and recombination of diverse knowledge elements within a specific knowledge domain [16,36]. The emergence of new knowledge helps firms to overcome “cognitive inertia” and develop novel solutions to existing problems [37]. This allows firms to transcend established rules and procedures [38], potentially catalyzing significant inventions, expanding new product domains [39], and ultimately leading to the development of innovative green products. Simultaneously, the generation of new knowledge facilitates the elimination of outdated knowledge and optimizes firms’ green product development activities (e.g., resource allocation, product design), thereby increasing the potential for improvements in existing products to better meet market and customer demands [40], ultimately enhancing market returns from green product development.

Drawing on these insights, we hypothesize the following:

Hypothesis 2.

BGTS promotes GPDP through knowledge coupling.

2.3.2. Mediating Effect of GTCC

From a knowledge management perspective, successful green technology commercialization relies heavily on the availability of green technology resources. All departments require access to technology resources to drive technology commercialization efforts [41]. BGTS targeting academic institutions provides firms with cutting-edge green technology knowledge necessary for green product development through various channels, including personnel mobility, informal interactions, consulting relationships, and joint research projects [11]. Therefore, a strong foundation of green technology knowledge supports the commercialization of green technology R&D achievements.

From a risk management perspective, the commercialization of green technologies faces multiple obstacles, including rapid technological iteration, unclear market acceptance, and complex innovation processes [42,43]. By sharing technological cognition with universities and research institutions, firms can stimulate more innovative ideas, reduce product development risks, and accelerate the market launch of green products. Furthermore, specialized and cutting-edge technological knowledge obtained through search can be directly applied to optimize solutions and improve processes, effectively reducing the costs and risks associated with green technology commercialization while enhancing commercialization capability.

GTCC refers to a firm’s ability to develop and adopt green products and process technologies, create new green products, and accelerate the market launch of these products [41]. GTCC primarily encompasses three dimensions: commercialization speed, market scope, and technology breadth. Regarding commercialization speed, firms that prioritize rapid market entry can avoid competition with partners [44], secure premium pricing, and enhance both market share and profitability. A broader market scope allows firms to distribute green product development costs across various geographic markets and product lines, thereby preserving price advantages and strengthening the competitive edge of green products. Greater technology breadth indicates the incorporation of diverse green technologies, enabling the development of varied features to meet customer needs or the creation of entirely new green products.

Thus, we hypothesize the following:

Hypothesis 3.

BGTS promotes GPDP through GTCC.

2.3.3. Serial Mediating Effect of Knowledge Coupling and GTCC

Resource orchestration theory posits that the resource management model comprises structuring the resource portfolio, bundling resources to build capabilities, and leveraging capabilities to create value [15]. In the context of green product development, this process is reflected in the pathway where heterogeneous technological knowledge, acquired through BGTS, is effectively integrated through knowledge coupling. This integration is then transformed into GTCC, ultimately enhancing GPDP.

BGTS targeting academic institutions provides firms with access to abundant cutting-edge green technology knowledge. However, this knowledge, originating from diverse domains, is characterized by heterogeneity and fragmentation, which necessitates a knowledge coupling mechanism for effective combination and synergy. Knowledge coupling plays a critical role in ‘bundling’ technological knowledge, enhancing GTCC through the following mechanisms: First, it eliminates cognitive gaps and logical conflicts between different knowledge domains, enabling firms to establish a systematic understanding of technology–market relationships and identify key pathways and potential obstacles in green technology commercialization. Second, the novel knowledge combinations generated during the coupling process enhance the firm’s problem-solving capacity [37], effectively reducing uncertainties inherent in green technology commercialization. Third, the cross-departmental knowledge integration fostered through knowledge coupling promotes synergy among R&D, production, and marketing functions, thereby improving green technology commercialization efficiency.

GTCC plays a vital role in the value realization stage. First, GTCC enables firms to rapidly identify market application scenarios for technologies, shorten technology verification and product development cycles, and accelerate the conversion of knowledge potential into economic momentum, thereby transforming green technologies into commercialized products more quickly. Second, strong GTCC allows firms to better integrate technological knowledge with market demand information, precisely grasp user needs and competitive dynamics during product development, and translate technological concepts into products that meet market expectations [45]. This process realizes the value creation of green technologies and promotes GPDP.

Accordingly, Hypothesis 4 is proposed as follows:

Hypothesis 4.

BGTS promotes GPDP through the serial mediation of knowledge coupling and GTCC.

2.4. Moderating Role of Digital Technology Adoption

The adoption of digital technologies such as Industry 4.0, cloud computing, big data, and the Internet of Things has transformed green innovation ecosystems by revolutionizing corporate mechanisms for value creation, delivery, and capture [46,47]. These technologies create expanded opportunities for the dissemination, diffusion, connection, and reorganization of green technology knowledge within manufacturing enterprises [48].

Digital technology adoption broadens enterprises’ access to green technology knowledge and enhances their efficiency in acquiring, studying, absorbing, and integrating external knowledge [49] by providing enhanced information processing capacity, computing power, communication, and connectivity [50]. At higher levels of digital technology adoption, newly acquired green technology knowledge can be rapidly integrated into green product prototypes for iterative optimization. This significantly enhances technological resource allocation efficiency during product development process [51], accelerates the transformation of green technology knowledge from acquisition to value creation, and ultimately expedites green product development [19].

Furthermore, the adoption of digital technology enhances enterprises’ data-driven insights and analytical capabilities [52], enabling more open transfer, integration, and application of external technology knowledge [47]. Higher levels of digital technology adoption improve enterprises’ ability to capture and identify tacit and dispersed knowledge beyond organizational boundaries. This allows enterprises to acquire higher-quality green technology knowledge through BGTS, facilitates deeper exploration of the value embedded in green technology knowledge, and increases the likelihood of overcoming green technology barriers to develop innovative green products.

Accordingly, Hypothesis 5 is proposed as follows:

Hypothesis 5.

Digital technology adoption positively moderates the relationship between BGTS and GPDP.

2.5. Moderating Role of Product Complexity

The essence of innovation lies in producing and integrating knowledge in novel ways, a knowledge-oriented activity intrinsically shaped by the complexity of knowledge structures [5]. Product complexity is defined by both the number of technological components and their interdependencies [25]. It influences the green innovation process by shaping the opportunity environment for BGTS-driven green product development.

Technological innovation stems from the recombination and synthesis of existing technological components. An increase in product complexity signifies growth in both the number of technological components and their interdependencies. The possibilities and quantity of combinations among technological components increase, thereby creating greater opportunities for technological innovation [20]. Consider an extreme case. A complete lack of interdependence between technological components implies minimal opportunity for creativity [25]. As product complexity increases, the opportunities for recombining internal and external technological components rise, thereby enhancing product innovation potential. Moreover, increased in product complexity elevates the demands on firms’ R&D capabilities [53] and functional diversity [54], prompting firms to broaden their external search scope and intensity. This enables them to acquire higher-quality green technology knowledge from diverse fields. Through the interaction of internal and external technology knowledge, firms achieve optimal technology combination and integration, thereby strengthening R&D capabilities and improving GPDP.

However, when product complexity exceeds a certain threshold, the negative effects of increased complexity may outweigh the benefits [55]. First, from a knowledge attributes perspective, higher product complexity entails greater tacit technology knowledge and stronger interdependencies among technological components, leading firms to develop entrenched technological routines that hinder the adoption and diffusion of new technologies [56]. Second, excessive product complexity can result in “absorptive capacity overload” within firms, meaning that existing technological understanding and knowledge-processing capabilities become inadequate for comprehending and managing the intricate interdependencies among technological components. Third, complex products constrain the creative process of technological component “decomposition and modification” [57]. Inventors and designers must invest considerable effort in understanding the interactions between technological components [25], plunging green product development into a “complexity disaster” that ultimately impedes GPDP improvement.

Thus, Hypothesis 6 is proposed as follows:

Hypothesis 6.

Product complexity exhibits an inverted U-shaped moderating effect on the relationship between BGTS and GPDP.

Figure 1 illustrates the theoretical model of this study.

3. Methods

3.1. Sample

This study targets manufacturing enterprises as research subjects. From January to March 2025, questionnaire surveys were conducted among managers and R&D personnel in manufacturing enterprises through multiple distribution channels: consulting firm intermediaries, collaborating partners, email delivery, and on-site distribution. We distributed 500 questionnaires, and 396 were collected. After excluding questionnaires with missing data or obvious inconsistencies, 313 valid responses were obtained, yielding a valid response rate of 63%. The basic characteristics of the sample enterprises are presented in

Table 1. Basic characteristics of sample enterprises.

Characteristics	Type	Number	Percentage
Firm age	1–5 years	15	4.8
	6–10 years	81	25.9
	11–15 years	106	33.9
	16–20 years	54	17.3
	More than 20 years	57	18.2
Firm size	Less than 100 people	32	10.2
	100–300 people	95	30.4
	300–500 people	69	22.0
	500–1000 people	58	18.5
	1000–2000 people	20	6.4
	More than 2000 people	39	12.5
Ownership type	State-owned enterprises	77	24.6
	Foreign-invested enterprises	8	2.6
	Joint venture enterprise	53	16.9
	Private enterprise	175	55.9
Industry	Chemical materials and chemical products manufacturing	27	8.6
	Pharmaceutical industry	32	10.2
	Electronic and telecommunication equipment manufacturing	52	16.6
	Electrical machinery and equipment manufacturing	61	19.5
	Automobile and parts manufacturing	30	9.6
	Food manufacturing	103	32.9
	Other manufacturing	8	2.6

. Overall, the sample encompasses manufacturing firms with diverse establishment periods, organizational scales, ownership types, and industry sectors, demonstrating strong representativeness.

3.2. Measures

BGTS. Drawing on green technology characteristics and the boundary-spanning search scale developed by Xiao and Zhu [58]. BGTS was measured using four items.

Knowledge coupling. Following Chen et al. [36], knowledge coupling was measured from two aspects: substitutive knowledge coupling (Subs KC) and complementary knowledge coupling (Com KC), comprising ten items in total.

GTCC. Building upon Chen’s [41] research on technology commercialization capability and adapting it to the characteristics of green technology, GTCC was measured through three aspects: commercialization speed (Comm Spd), market scope (Mkt Scp), and technology breadth (Tech Brd), comprising eleven items in total.

Product complexity. Following Dean et al. [5], product complexity was measured using four items.

Digital technology adoption. Referring to Eller et al. [59], the measurement consisted of five items.

GPDP. Drawing on new product development performance studies by Wang and Gao [60] and Ma et al. [61], and incorporating Chen et al.’s [62] research on GPDP. We conceptualized GPDP as a formative construct composed of three aspects: market performance, creativity, and environmental performance, with a total of nine items.

Controls. Following prior research [58], we controlled for firm’s age, size, ownership type, and industry sector.

Table A1. Scale design.

Variables		Measures	Factor Loading
BGTS		Our firm regularly collaborates with universities and research institutions on green technology talent training programs.	0.795
		Our firm frequently exchanges information on green technology development, industrial green policies, and emerging trends with government research departments.	0.815
		Our firm actively tracks the green technology research advances of leading universities, research institutes, and national key laboratories.	0.742
		Our firm consistently monitors information on green technology standards, green patent, and related developments.	0.673
Knowledge coupling	Subs KC	Our firm frequently acquires relevant experience and skills from our partners for comparable products or services.	0.747
		Our firm has market development and customer retention strategies similar to those of our partners.	0.738
		Our firm’s target customers are highly similar to those of our partners in terms of needs, preferences, and behaviors.	0.737
		Our firm’s management structure and organizational norms are highly consistent with those of our partners.	0.650
		Our firm’s management processes, systems, and culture are highly similar to those of our partners.	0.686
	Com KC	Our firm can utilize the experience and skills of our partners in product production or service delivery.	0.793
		In producing products or delivering services, our firm often applies our partners’ proprietary technologies and methods to solve practical problems.	0.819
		Our firm often learns from the experiences and know-how of our partners in customer service.	0.725
		Our firm often learns from our partners’ experiences and methods in organizational management.	0.755
		Some of our partners’ organizational management practices can compensate for deficiencies in our firm’s management.	0.742
GTCC	Comm Spd	Our firm has the ability to initiate green product concepts in a timely manner.	0.726
		Our firm is capable of developing green products in a timely manner.	0.741
		Our firm has the ability to launch green products to market in a timely manner.	0.762
	Mkt Scp	Our firm excels at improving existing green products for different demographic markets.	0.731
		Our firm excels at improving existing green products for different geographic markets.	0.720
		Our firm excels at developing new green products for different demographic markets.	0.695
		Our firm excels at developing new green products for different geographic markets.	0.756
	Tech Brd 1	Our firm has the ability to acquire the technologies for improving existing green products.	0.753
		Our firm has the ability to acquire the technologies for the development of new green products.	0.737
		Our firm has the ability to integrate technologies to improve existing green products.	0.713
		Our firm has the ability to integrate technologies to develop new green products.	0.749
Digital technology adoption		Our firm has adopted the Internet of Things.	0.803
		Our firm has adopted 5G network.	0.814
		Our firm has adopted big data analytics.	0.813
		Our firm has adopted cloud computing.	0.831
		Our firm has adopted digital twin technology.	0.622
Product complexity		The product design is complex.	0.818
		The product structure is sophisticated.	0.778
		The product is functionally complex.	0.794
		The product structure is intricate (i.e., there are many interdependencies among components).	0.752
GPDP	Market performance	Our firm’s green product development projects have achieved the expected profit margins.	0.700
		Our firm’s green product development projects have increased the firm’s market share.	0.823
		Our firm launches green products faster than its major competitors.	0.778
	Creativity	Our firm’s green product development projects are more innovative than those of our competitors.	0.809
		The technologies used in our firm’s green product development projects are substantially different from those of other companies in the industry.	0.790
		Compared to competitors, our firm’s green products offer unique advantages.	0.771
	Environmental performance	Our firm’s green product development projects achieve resource utilization optimization goal.	0.751
		Our firm’s green product development projects achieve environmental protection goals.	0.738
		The green products developed by our firm are safe, healthy, and non-toxic.	0.793

¹ Technology breadth refers to a firm’s ability to integrate multiple technologies into products, determined by the firm’s capability to acquire and integrate external heterogeneous technologies. The first two items capture technology acquisition, representing the ability to expand the technological scope, which forms the foundation of technology breadth. While the latter two capture technology integration for product development, representing its value realization.

presents the scale design.

3.3. Reliability and Validity Analysis

Reliability and validity were assessed through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA), with results presented in

Table 2. Scale analysis of reliability and validity.

Variables	Dimensions	Cronbach’α	CR	AVE	KMO
BGTS		0.843	0.843	0.575	0.816
Knowledge coupling	Subs KC	0.836	0.837	0.508	0.903
	Com KC	0.876	0.877	0.589
GTCC	Comm Spd	0.787	0.787	0.552	0.882
	Mkt Scp	0.815	0.816	0.527
	Tech Brd	0.827	0.827	0.545
Digital technology adoption		0.883	0.885	0.609	0.874
Product complexity		0.865	0.866	0.618	0.807
GPDP	Market performance	0.809	0.812	0.591	0.840
	Creativity	0.831	0.833	0.624
	Environmental performance	0.804	0.805	0.579

Note: AVE = Average variance extraction.

. The Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy exceeded 0.7 for all constructs, indicating the appropriateness of factor analysis. Most of the factor loadings ranged from 0.7 to 0.9 (see Table A1), demonstrating strong validity. All Cronbach’s alpha coefficients exceeded 0.7, and all composite reliability (CR) values exceeded 0.8, indicating high internal consistency reliability.

Model fit indices are presented in

Table 3. Model fitting analysis results.

Model	NC (X2/df)	RMSEA	RMR	SRMR	IFI	TLI	CFI
Overall model	1.612	0.044	0.031	0.047	0.924	0.918	0.924

. The normed chi-square (NC) was less than 3, the root mean square error of approximation (RMSEA), the root mean square residual (RMR), and the standardized root mean square residual (SRMR) were all below 0.05; and the incremental fit index (IFI), Tucker–Lewis index (TLI), and comparative fit index (CFI) were all exceeded 0.9, indicating good convergence validity.

Discriminant validity test results are presented in

Table 4. Discriminant validity test results.

Variables	Product Complexity	Digital Technology Adoption	BGTS	Environmental Performance	Creativity	Market Performance	Tech Brd	Mkt Scp	Comm Spd	Subs KC	Com KC
Product complexity	0.618
Digital technology adoption	0.135	0.609
BGTS	0.122	0.147	0.575
Environmental performance	0.109	0.074	0.449	0.579
Creativity	0.104	0.071	0.430	0.541	0.624
Market performance	0.107	0.073	0.441	0.555	0.532	0.591
Tech Brd	0.198	0.088	0.478	0.552	0.529	0.542	0.545
Mkt Scp	0.198	0.088	0.476	0.549	0.527	0.540	0.619	0.527
Comm Spd	0.193	0.086	0.465	0.537	0.515	0.527	0.605	0.602	0.552
Subs KC	0.219	0.071	0.476	0.575	0.551	0.565	0.587	0.584	0.571	0.589
Com KC	0.205	0.067	0.446	0.538	0.516	0.529	0.549	0.547	0.534	0.583	0.508
$\sqrt{AVE}$	0.786	0.780	0.758	0.761	0.790	0.769	0.738	0.726	0.743	0.767	0.713

Note: The diagonal line is AVE.

. The AVE for each construct exceeded all inter-construct correlations, confirming adequate discriminant validity.

3.4. Common Method Bias Assessment

First, Harman’s single-factor test extracted eleven factors from the unrotated data, with the first factor accounting for 27.68% of the total variance, below the 40% threshold [63]. Second, two-factor method test showed that incorporating a common method factor into the original measurement model. Comparison of fit indices between the two models revealed negligible changes (ΔNC = −0.003, ΔRMSEA = 0, ΔRMR = 0.001, ΔSRMR = −0.001, ΔIFI = 0.003, ΔTLI = 0, ΔCFI = 0.002). These results suggest that common method bias was not a significant concern in this study.

4. Results

4.1. Descriptive Statistics and Correlation Analysis

Table 5. Descriptive statistics and correlation matrix.

Variables	Mean	Std.	1	2	3	4	5
BGTS	3.798	0.719
Knowledge coupling	4.020	0.551	0.449 **
GTCC	4.030	0.554	0.475 **	0.666 **
Digital technology adoption	3.983	0.696	0.130 *	0.070	0.093
Product complexity	3.944	0.652	0.101	0.215 **	0.195 **	0.119 *
GPDP	3.986	0.557	0.454 **	0.668 **	0.676 **	0.072	0.112 *

Note: **, * denote significance at 1%, and 5% levels.

presents the descriptive statistics and correlation matrix of the variables. BGTS exhibited significant positive correlations with knowledge coupling (β = 0.449, p < 0.01), GTCC (β = 0.475, p < 0.01), and GPDP (β = 0.454, p < 0.01). Similarly, significant positive correlations were found between knowledge coupling and GPDP (β = 0.668, p < 0.01), GTCC and GPDP (β = 0.676, p < 0.01), and knowledge coupling and GTCC (β = 0.666, p < 0.01). These findings provided preliminary support for Hypotheses H1, H2, H3, and H4. Collinearity diagnostics revealed that all variance inflation factors (VIFs) were below 2, and the correlation coefficients among the variables were all below 0.7, indicating the absence of multicollinearity concerns.

4.2. Verification of Direct Effect and Mediating Effect

The direct and mediating effects were examined using the sequential testing procedure, with results presented in

Table 6. Hierarchical regression test results of direct and mediating effects.

Variables	Knowledge Coupling	GTCC	GPDP
	Model 1	Model 2	Model 3	Model 4	Model 5
Constant	2.599 ***	2.520 ***	2.602 ***	1.074 ***	1.078 ***
Firm age	0.012	0.028	0.003	−0.004	−0.014
Firm size	0.008	0.033	0.016	0.012	−0.004
Ownership type	0.020	0.012	0.002	−0.010	−0.005
Industry	0.000	−0.012	0.001	0.001	0.008
BGTS	0.342 ***	0.350 ***	0.346 ***	0.145 ***	0.134 ***
Knowledge coupling				0.588 ***
GTCC					0.605 ***
R2	0.204	0.244	0.208	0.477	0.482
ΔR2	0.192 ***	0.199 ***	0.193 ***	0.269 ***	0.274 ***
F	15.767 ***	19.818 ***	16.143 ***	46.553 ***	47.454 ***
f2	0.241	0.264	0.243

Note: *** denotes significance at 0.1% level. f² effect size: 0.02 (small), 0.15 (medium), and 0.35 (large).

. Additionally, the Bootstrap method was employed to estimate the mediating effect values, as shown in

Table 7. Tests of mediating effects based on Bootstrap.

Mediating Pathway	Effect Value	SD	95%CI
BGTS → Knowledge coupling → GPDP	0.162	0.051	[0.073, 0.267]
BGTS → GTCC → GPDP	0.078	0.033	[0.027, 0.156]
BGTS → Knowledge coupling → GTCC → GPDP	0.097	0.035	[0.038, 0.175]

.

First, we tested the main effect. As shown in Model 3 of Table 6, BGTS positively affected GPDP (β = 0.346, p < 0.001, f² = 0.241), supporting Hypothesis H1.

Second, the mediating effect of knowledge coupling was tested. Model 1 demonstrates that BGTS positively influenced knowledge coupling (β = 0.342, p < 0.001, f² = 0.264). Model 4 reveals that knowledge coupling positively affected GPDP (β = 0.588, p < 0.001). Furthermore, the Bootstrap results in Table 7 indicate that the mediating effect of the path “BGTS → knowledge coupling → GPDP” was 0.162, with a 95% confidence interval of [0.073, 0.267], excluding zero. These results confirm that the mediating effect of knowledge coupling is significant, thereby validating Hypothesis H2.

Third, we examined the mediating effect of GTCC. Model 2 demonstrates that BGTS significantly and positively affected GTCC (β = 0.350, p < 0.001, f² = 0.243). Model 5 reveal that GTCC positively influenced GPDP (β = 0.605, p < 0.001). The Bootstrap results in Table 7 indicate that the mediating effect value of the path “BGTS → GTCC → GPDP” was 0.078, with a 95% confidence interval of [0.027, 0.156], excluding zero, indicating that the mediating effect of GTCC is significant. Therefore, Hypothesis H3 is supported.

Fourth, we examined the serial mediating effect. The results from Table 6 and previous analyses demonstrate that BGTS positively affected knowledge coupling, knowledge coupling positively influenced GTCC, and GTCC positively promoted GPDP. Moreover, Table 7 shows that the serial mediating effect of the path “BGTS → knowledge coupling → GTCC → GPDP” was 0.097, with a 95% confidence interval of [0.038, 0.175], excluding zero, indicating that knowledge coupling and GTCC function as serial mediators between BGTS and GPDP. Hence, Hypothesis H4 is verified.

4.3. Verification of the Moderating Effect

We constructed Equations (1) and (2) to test the moderating effects of digital technology adoption and product complexity, where X represents BGTS, W₁ denotes digital technology adoption, W₂ denotes product complexity, and Y represents GPDP.

Y = β₀ + β₁X + β₂W₁ + β₃X × W₁ + ε

(1)

Y = β₀ + β₁X + β₂W₂ + β₃W₂² + β₄X × W₂ + β₅X × W₂² + ε

(2)

The regression results are presented in

Table 8. Tests of moderating effects.

Variables	GPDP
	Model 6	Model 7	Model 8	Model 9
Constant	3.915 ***	3.912 ***	3.922 ***	3.844 ***
Firm age	0.004	0.003	0.002	−0.003
Firm size	0.016	0.019	0.018	0.036
Ownership type	0.002	0.004	0.002	0.013
Industry	0.001	−0.005	0.001	0.002
BGTS	0.345 ***	0.350 ***	0.338 ***	0.395 ***
Digital technology adoption	0.010	−0.002
BGTS × Digital technology adoption		0.186 ***
Product complexity			0.043	0.078
Product complexity2			−0.015	−0.067
BGTS × Product complexity				0.113
BGTS × Product complexity2				−0.124 **
R2	0.208	0.236	0.213	0.263
ΔR2	0.193 ***	0.028 **	0.005	0.050 ***
F	13.420 ***	13.468 ***	11.795 ***	12.005 ***
f2		0.037		0.068

Note: ***, ** denote significance at 0.1%, and 1% levels. f² effect size: 0.02 (small), 0.15 (medium), and 0.35 (large).

. Model 7 reveals that the interaction term between BGTS and GTCC was significantly and positively associated with GPDP (β = 0.186, p < 0.001, f² = 0.037), indicating that digital technology adoption positively moderates the relationship between BGTS and GPDP. The moderating effect of digital technology is depicted in Figure 2. The slope for the high digital technology adoption group was steeper than that for the low digital technology adoption group, thereby supporting Hypothesis H5.

In Model 9, the interaction term between BGTS and the quadratic term of product complexity was significantly and negatively associated with GPDP (β = −0.124, p < 0.01, f² = 0.068), indicating that product complexity exerts an inverted U-shaped moderating effect on the relationship between BGTS and GPDP. These results support Hypothesis H6.

Figure 3 was generated using Stata 17 to visualize the moderating effect of product complexity, which exhibited a threshold effect. Below this threshold, an increase in product complexity enhanced the positive impact of BGTS on GPDP. However, beyond the threshold, further increases in product complexity weakened this positive impact. This finding indicates an inverted U-shaped moderating effect of product complexity, thereby supporting Hypothesis 6.

4.4. Robustness Test

Following Ma et al. [64], we divided GPDP into three dimensions for robustness testing: market performance, creativity, and environmental performance. The mediating effects were examined using the Bootstrap method. As shown in

Table 9. Robustness tests of mediating effects.

Mediating Pathway	Effect Value	SD	95%CI
BGTS → Market performance	0.333	0.053	[0.229, 0.437]
BGTS → Knowledge coupling → Market performance	0.103	0.052	[0.001, 0.207]
BGTS → GTCC → Market performance	0.071	0.031	[0.021, 0.142]
BGTS → Knowledge coupling → GTCC → Market performance	0.087	0.035	[0.030, 0.167]
BGTS → Creativity	0.408	0.052	[0.307, 0.510]
BGTS → Knowledge coupling → Creativity	0.128	0.044	[0.047, 0.218]
BGTS → GTCC → Creativity	0.050	0.024	[0.014, 0.105]
BGTS → Knowledge coupling → GTCC → Creativity	0.062	0.028	[0.018, 0.128]
BGTS → Environmental performance	0.298	0.053	[0.193, 0.403]
BGTS → Knowledge coupling → Environmental performance	0.155	0.051	[0.063, 0.261]
BGTS → GTCC → Environmental performance	0.066	0.029	[0.020, 0.133]
BGTS → Knowledge coupling → GTCC → Environmental performance	0.081	0.031	[0.029, 0.149]

, the results using dimension-specific GPDP indicators were consistent with those presented in Table 7. The mediating effects of knowledge coupling and GTCC, as well as the serial mediating effect, were statistically significant across all dimensions.

Using multilevel linear regression analysis, we examined the moderating effects of digital technology adoption and product complexity, with results presented in

Table 10. Robustness tests of moderating effects.

Variables	Market Performance		Creativity		Environmental Performance
	Model 10	Model 11	Model 12	Model 13	Model 14	Model 15
Constant	3.632 *	3.575 ***	4.241 ***	4.164 ***	3.860 ***	3.791 ***
Firm age	−0.005	−0.017	−0.036	−0.039	0.052	0.047
Firm size	0.061	0.081	0.001	0.015	−0.004	0.011
Ownership type	0.025	0.032	−0.039	−0.029	0.027	0.036
Industry	0.005	0.011	−0.009	−0.003	−0.011	−0.003
BGTS	0.332 ***	0.348 ***	0.415 ***	0.471 ***	0.303 ***	0.364 ***
Digital technology adoption	0.030		−0.030		−0.004
BGTS × Digital technology adoption	0.153 *		0.159 *		0.247 ***
Product complexity		0.078		0.072		0.088
Product complexity2		−0.053		−0.062		−0.085
BGTS × Product complexity		0.222 **		0.074		0.046
BGTS × Product complexity2		−0.056		−0.146 *		−0.167 **
R2	0.150	0.175	0.190	0.209	0.138	0.148
ΔR2	0.012 *	0.036 **	0.013 *	0.031 **	0.031 ***	0.035 **
F	7.714 ***	7.139 ***	10.191 ***	8.922 ***	6.958 ***	5.841 ***
f2		0.030		0.024		0.012

Note: ***, **, * denote significance at 0.1%, 1%, and 5% levels. f² effect size: 0.02 (small), 0.15 (medium), and 0.35 (large).

. The moderating effect of digital technology adoption was consistent with the findings in Table 8. The interaction term between BGTS and digital technology adoption was significantly and positively associated with market performance (β = 0.153, p < 0.05), creativity (β = 0.159, p < 0.05), and environmental performance (β = 0.247, p < 0.001). Regarding product complexity’s moderating effect, the interaction term between BGTS and the quadratic term of product complexity showed a non-significant effect on market performance (β = −0.056, p > 0.05), while demonstrating significant negative effects on creativity (β = −0.146, p < 0.05) and environmental performance (β = −0.167, p < 0.01). Since market performance, creativity, and environmental performance constitute three dimensions of GPDP, we conclude that the inverted U-shaped moderating effect of product complexity remains statistically significant.

5. Discussion

5.1. Research Findings

Drawing on resource orchestration theory and recombinant search theory, this study investigates the mechanisms through which BGTS influences GPDP. The main findings are as follows:

First, BGTS significantly promotes GPDP. As an important channel for the commercialization of green scientific research outcomes, BGTS facilitates interactions between enterprises and academic institutions on green technology. It provides enterprises with access to cutting-edge green technology knowledge, thereby enabling the development of unique product functions, performance and quality, which ultimately enhances GPDP. For instance, Tesla has collaborated with Dalhousie University on green technology, focusing on enhancing the energy density, reducing costs, and extending the lifespan of lithium batteries. This collaboration has successfully translated cutting-edge laboratory battery technology into commercial products, thereby significantly strengthening Tesla’s competitiveness in the global market.

Second, knowledge coupling and GTCC exert a serial mediating effect on the relationship between BGTS and GPDP. BGTS enables enterprises to acquire differentiated green technology knowledge, thereby increasing the likelihood of knowledge coupling. Through knowledge coupling, enterprises systematically screen, select, and integrate green technology knowledge across various domains, which continuously enriches their knowledge base and provides critical technological support for the development of green technology commercialization capabilities. Moreover, GTCC facilitates the transformation of green technologies into market-ready green products, ultimately yielding more competitive green products.

Third, digital technology adoption positively moderates the relationship between BGTS and GPDP. Digital technology adoption enables enterprises to establish comprehensive systems for acquiring and analyzing technological resources. By expanding the scope of external green technology resource search, it promotes the deep integration of internal and external green technologies, thereby fully unleashing the value of green technology resources and effectively advancing the green product development process.

Fourth, product complexity exhibits an inverted U-shaped moderating effect on the relationship between BGTS and GPDP, yet this effect varies significantly across different GPDP dimensions. Specifically, the moderating effect is significant for the creativity and environmental performance dimensions but not significant for the market performance dimension. This discrepancy may arise from the varying sensitivity of each performance dimension to technological factors. Creativity assesses technological breakthroughs, novelty, and inventiveness, while environmental performance focuses on green technology outcomes such as resource utilization efficiency and pollution reduction. The relationship between BGTS and these two dimensions is more strongly influenced by product complexity, a technology-oriented factor. In contrast, market performance, measured by product profitability and market share, is predominantly influenced by market-oriented factors, such as brand reputation, channel coverage, pricing strategy, and marketing promotion. Therefore, market performance is less sensitive to changes in product complexity.

5.2. Theoretical Contributions

This study makes significant contributions to the management literature.

First, this study introduces the concept of BGTS and positions it as a key strategy for enterprises to bridge the gap between green scientific research and green product development, thereby facilitating crossing the VoD. Existing research predominantly adopts a holistic innovation chain perspective, exploring how to overcome the VoD by bridging financing gaps, enhancing university–industry collaborations, and establishing intermediary organizations. Few studies have investigated the VoD from a firm-level perspective, particularly within the context of green innovation. This study reframes the role of enterprises from “co-participant” in the innovation ecosystem to “proactive agents” actively crossing the VoD. Beginning with BGTS—a technology search behavior initiated by enterprises—this study uncovers the micro-mechanisms and boundary conditions underlying it enhances GPDP, thereby offering a novel perspective on the VoD in green innovation.

Second, this study constructs an integrated “knowledge–capability–opportunity–performance” framework of green technology commercialization and introduces resource orchestration theory and recombinant search theory—both widely applied and highly explanatory frameworks in general innovation research—into the context of manufacturing enterprises’ green product development. Resource orchestration theory delineates the comprehensive pathway from the integration of green technology resources to the development of capabilities and, ultimately, to performance enhancement. Recombinant search theory clarifies a critical conversion mechanism within this pathway: the transformation of external green technology knowledge into a firm’s capacity for green technology commercialization through search and recombination. This integration advances the application of established theoretical achievements into the green innovation research within manufacturing enterprises, providing novel insights for this field.

Third, this study identifies two critical “opportunity” factors—digital technology application and product complexity—as key mechanisms influencing the relationship between BGTS and GPDP. The findings demonstrate that digital technology adoption accelerates the commercialization of green technology, thereby advancing research on the role of digital technology in green product development. Furthermore, by examining the nonlinear moderating effect of product complexity, this study provides empirical support for recombinant search theory, which posits that the interdependencies among components are key determinants of both the process and outcomes of recombinant search [20,25].

5.3. Managerial Implications

The findings of this study offer several insights for managers in manufacturing enterprises.

First, managers should recognize the strategic significance of BGTS and strengthen their boundary-spanning capabilities. Specifically, they should enhance engagement with key stakeholders such as universities and research institutions and proactively explore heterogeneous green technology knowledge across organizational boundaries thereby obtaining novel ideas and experiences, facilitating the reconstruction and upgrading of their technological systems, and ultimately driving green product development initiatives.

Second, managers should establish a comprehensive pathway from green technology search to green product development to promote the commercialization of green technology. Specifically, they should enhance knowledge coupling by continuously integrating green technology knowledge across multiple technological domains to support green technology innovation and R&D. This approach strengthens internal green technology knowledge reserves, develops GTCC, and ultimately enhances the value of the enterprises’ green innovation chain.

Third, managers should promote the adoption of digital technologies—including artificial intelligence, big data, blockchain, 5G, and virtual reality—across enterprise functions such as technology R&D, technology management, production processes, and marketing. By continuously enhancing digital capabilities, enterprises can more accurately and effectively identify, integrate, and allocate internal and external technological resources.

Fourth, managers should strategically manage product complexity by tracking indicators such as R&D cycles and cross-departmental coordination costs to assess the level of complexity. In product design, adopting a modular architecture can help control complexity by reducing coupling between technological components, thereby maintaining product complexity within the optimal range that drives innovation. This approach helps prevent excessive complexity from causing organizational lock-in and innovation bottlenecks.

5.4. Limitations and Future Research

Despite its contributions, this study has several limitations that suggest directions for future research.

First, given the increasingly blurred boundaries between science and the market, many large firms have established internal R&D departments to engage in cutting-edge scientific research. From an intra-organizational perspective, exploring how internal R&D departments collaborate with other functional units to bridge the VoD represents an important research direction. Future research could employ qualitative methods, such as case studies, to reveal the underlying mechanisms of “how” and “why” enterprises successfully bridge the VoD.

Second, numerous factors may influence the relationship between BGTS and GPDP. Due to time and space constraints, this study examined only digital technology adoption and product complexity as moderating factors. Future research could explore additional contextual factors such as the industry environment and government policies. Moreover, the focus on China’s manufacturing sector limits the generalizability of our findings, particularly regarding the optimal threshold of product complexity, which likely varies across countries and industries. Future research could conduct cross-cultural or cross-industry comparative studies to validate and refine the boundary conditions identified in this study.

Third, this study primarily controls for basic organizational characteristics, including firm age, size, ownership type, and industry, without accounting for other enterprise-level factors such as R&D intensity or top management’s environmental commitment. Future research could integrate survey data with objective financial metrics or develop more comprehensive measurement scales to enhance the robustness of findings.

Fourth, the cross-sectional, self-reported survey data has several limitations. Cross-sectional data limits causal inferences and cannot rule out reverse causality. Additionally, although the survey targeted executives and R&D personnel with green-product development experience, individual responses may not fully capture organizational-level phenomena. Furthermore, self-reported data may introduce endogeneity concerns. Future research could adopt longitudinal designs or utilize secondary data to strengthen causal evidence. Employing instrumental variable methods or triangulation techniques could also help address potential endogeneity issues.