Ambidextrous Alliances, Complementary Assets, and Firms’ Breakthrough Innovations: Evidence from High-Tech Firms in China

Abstract

Breakthrough innovations present both opportunities and challenges for firms navigating unpredictable technological and market dynamics, which is significant to firms’ sustainable development. However, the impact of strategic alliances on breakthrough innovations remains contested, and the underlying mechanisms are yet to be fully clarified. Drawing on ambidexterity theory and profiting from innovation (PFI) theory, this study investigates the relationships among ambidextrous (exploratory vs. exploitative) alliances, complementary assets, and firms’ breakthrough innovations. By analyzing 279 questionnaire responses from Chinese firms, we demonstrate that both exploratory and exploitative alliances, alongside complementary assets, significantly enhance breakthrough innovations. Furthermore, complementary assets play a complete mediating role between exploratory alliances and breakthrough market innovations. In addition, complementary assets also play a partial mediating role between exploratory alliances and breakthrough technological innovations, as well as between exploitative alliances and both breakthrough technological and market innovations. These findings advance ambidextrous theory by delineating the ambidextrous roles of alliances, extend PFI theory through the integration of the mediating roles of complementary assets, and offer actionable insights for managers seeking to leverage ambidextrous alliances for breakthrough innovation success.

1. Introduction

As emerging technologies rapidly evolve and new products and markets continuously emerge, organizations are increasingly seeking new growth opportunities beyond traditional products and technologies to pursue sustainable development [1,2]. Breakthrough innovations, defined as a form of disruptive creation [3,4], transcend existing technological trajectories [5,6], establish new technical standards, and create entirely new markets [7,8]. By providing new directions for technology and market development that are difficult for competitors to imitate, breakthrough innovations enable firms to achieve higher and more stable performance [7,9].

Breakthrough innovations refer to innovations that pursue dual breakthroughs in technology and market domains [6]. Due to their complexity, high risk, and uncertainty, firms are unlikely to achieve breakthrough innovations independently because of limited resource endowments [3,8]. Through forming strategic alliances, organizations can create unique value by partnering with entities that possess complementary assets [10]. However, the number of strategic alliances formed by firms grows at an annual rate of 25%, and 60–70% of these alliances end in failure [11]. To pursue technological and market breakthrough innovations, focal firms should actively engage in strategic alliances with diverse purposes, leveraging the complementarity of innovative resources [12]. On the one hand, alliances enable firms to acquire and integrate heterogeneous resources, such as knowledge, technology, and capabilities, while sharing the risks and reducing the costs of breakthrough innovations [9,13]. On the other hand, alliance cooperation may lead to knowledge leakage [12], aggravation of competition, and conflicts of interest, which can hinder focal firms’ innovation and sustainable development [14,15]. Thus, the impact of strategic alliances on focal firms’ breakthrough innovations is contingent on their purposes to form it [15]. Ambidexterity theory suggests that strategic alliances involve a balance between exploration and exploitation activities [14,16,17]. For instance, Kauppila [14] found that exploratory alliances promote firms’ long-term performance, while exploitative alliances enhance short-term performance. Exploratory alliances primarily foster technological creation, whereas exploitative alliances focus on technological improvement. Building on previous views [16], we categorize strategic alliances into exploratory and exploitative alliances based on their purposes. Exploratory alliances focus on exploration activities, such as basic research, experimentation, and innovation discovery, typically involving upstream partners like universities, research institutes, and research-intensive firms. By contrast, exploitative alliances emphasize activities such as marketing, manufacturing, and customer relationship management, often involving downstream partners like customers, suppliers, and competitors [18,19].

Although previous studies have examined the effects of exploratory and exploitative alliances on firms’ breakthrough innovations, the underlying mechanisms remain unclear [17]. From the perspective of profiting from innovation (PFI) theory, complementary assets—such as manufacturing capabilities, sales channels, and customer relationships—are critical for firms to prevent imitation from competitors and successfully commercialize innovations. These special and unique assets complement incumbent firms’ technological innovation and enhance their value [20]. The purpose of alliances is to acquire or integrate partners’ resources to build complementary assets that firms lack [11]. Therefore, it is essential to consider the role of complementary assets when examining the relationship between ambidextrous alliances and breakthrough innovations. However, little research has considered this issue. This raises our questions: How do ambidextrous alliances influence firms’ breakthrough innovations? How do complementary assets impact this relationship? To address these questions, we conducted a survey study in China. By analyzing 279 valid responses, we found that ambidextrous alliances positively influence firms’ breakthrough innovations, with complementary assets mediating the relationship between them. This study enhances our understanding of how ambidextrous alliances link to firms’ breakthrough innovations in two ways. First, we expand PFI theory in the context of breakthrough innovations by constructing a research framework that illustrates the relationships among ambidextrous (exploratory vs. exploitative) alliances, complementary assets, and breakthrough innovations. This provides a new perspective for understanding the antecedents of firms’ breakthrough innovations. Second, we challenge the one-dimensional paradigm of technological breakthroughs that was dominant in prior research —which overemphasizes technological aspects—and refine the classification of breakthrough innovations by grounding it in the dual dimensions of technology and market, as suggested in the existing literature [5].

Our study is structured as follows. Section 2 reviews the relevant literature, develops the hypotheses, and constructs the theoretical model. Section 3 describes the research design, including the research setting and procedures. Section 4 presents the empirical results, including hierarchical regression analysis and robustness checks. Section 5 provides a discussion of our findings. Finally, Section 6 concludes with the contributions of this work and illustrates limitations and future research directions.

2. Theoretical Background and Hypotheses

2.1. Ambidextrous Alliances

With the rapid advancement of digital technology and intensifying market competition [21], innovation activities demand increasingly robust capabilities, knowledge, and resources from focal firms. Consequently, firms often choose to form strategic alliances with other organizations to pursue breakthrough innovations. Strategic alliances describe cooperative relationships between at least two independent organizations that aim to share resources and capabilities to achieve mutually beneficial outcomes [1,22].

Previous research has explored the connotations and concepts of ambidextrous alliances to some extent. However, no consensus has been reached [17,18,21,23]. For example, Rothaermel and Deeds [18] noted that exploratory alliances involve strategic cooperation with upstream partners (e.g., suppliers, research institutions) for new product development, technology exploration, and testing innovative methods. Exploitative alliances, by contrast, focus on collaboration with downstream partners (e.g., distributors, retailers) to expand sales networks, franchise operations, or co-branding initiatives. Yang et al. [17] defined alliances targeting upstream value-chain activities (e.g., R&D, new drug development) as exploratory, while those focused on downstream activities (e.g., manufacturing, marketing) were classified as exploitative. Lavie and Rosenkopf [23] differentiated exploratory and exploitative alliances based on the alliance’s purpose, the network structure, and location of partners, as well as the attributes of those partners.

We argue that exploratory alliances aim to transcend an organization’s existing resources by jointly creating new opportunities with partners. Exploitative alliances, meanwhile, prioritize optimizing and managing current knowledge and resources, particularly in activities such as product marketing and manufacturing [24]. Therefore, we define alliances between firms and upstream partners (e.g., universities, research institutes, R&D-intensive firms) that focus on developing new technologies, conducting basic research, or exploring novel knowledge as exploratory alliances. Alliances between firms and downstream partners (e.g., customers, suppliers, competitors) aimed at exchanging exploitable resources, enhancing technology development, or implementing production processes are categorized as exploitative alliances.

2.2. Complementary Assets

Drawing on PFI theory, Teece [20] categorizes complementary assets into three distinct types based on their interdependence with technological innovation: general, specialized, and co-specialized complementary assets. General complementary assets encompass widely available industry-agnostic resources, such as generic manufacturing equipment or standard technical skills [25]. Their value is not contingent on specific innovations, and their absence poses minimal risk to innovation outcomes due to being readily substitutable. Specialized complementary assets are characterized by unilateral dependence on innovation activities; these assets are tailored to specific technologies or processes [26]. Their development involves significant irreversibility and sunk costs, creating high risks for investors. Co-specialized complementary assets exhibit mutual interdependence with innovation, where their value is maximized when combined with specific technologies and vice versa [27]. Unlike specialized assets, co-specialization creates symbiotic value—innovators and asset holders jointly benefit, but their separation diminishes returns for both parties.

Firms can transform resources from these relationships into valuable assets. For example, in the manufacturing industry, widespread vertical alliances between suppliers and buyers play a key role in demonstrating how firms can obtain complementary assets. The R&D capabilities of manufacturers and the channel management and marketing integration abilities of distributors are complementary [28]. Complementary assets are unique and scarce, making them difficult to obtain internally. Therefore, firms often acquire them through market transactions. Scholars have shown that firms cooperate with external organizations to obtain complementary assets that support business model innovation [29]. Since knowledge, resources, and capabilities can be transferred across borders among alliance partners, strategic alliances have become an important strategy for firms to obtain complementary assets.

2.3. Breakthrough Innovations

Breakthrough innovations refer to an innovation type where technology advances by leaps and bounds, significantly transforming the consumer market [1,9]. They are characterized by being cutting-edge, novel, complex, high-risk, and uncertain, requiring knowledge combination, technology synergy, and supportive market and organizational structures [4,8,9].

Previous research highlighted three key aspects that distinguish breakthrough innovations from other types of innovation [1,7]. First, breakthrough technological innovations are a leap forward, causing a discontinuous and disruptive shift in the technological landscape. They play a critical role in technological progress by causing a break in the original technical trajectory and stimulating the development of a new one [1]. Second, the market impact of breakthrough technology is crucial, including whether it creates emerging markets and leads the development of technological and industrial innovations. Capponi et al. [7] argue that breakthrough innovations deviate significantly from established practices, potentially reshaping existing markets or creating entirely new ones. Therefore, when evaluating breakthrough innovations, it is important to consider whether new technologies and products have opened up new markets and had a substantial impact on the industry [30]. Breakthrough innovations can alter the competitive landscape, create new industries, and establish new markets [6,13]. Third, breakthrough innovations not only require firms to have advanced technical skills and capabilities but also require them to acquire resource heterogeneity from external sources to continuously improve their abilities [3]. As a result, firms need to enhance organizational flexibility to facilitate cross-border searches, break down organizational barriers, and avoid falling into the trap of core rigidity [13].

Studies have argued that alliance leaders can obtain heterogeneous, novel, and complementary resources that facilitate breakthrough innovations [13]. However, there is still a lack of discussion on the intermediate mechanisms through which alliances influence firms’ breakthrough innovations. Additionally, different types of alliance partners bring varying degrees of resource heterogeneity, and the relationship between strategic alliances and breakthrough innovations has not been thoroughly analyzed.

2.4. Ambidextrous Alliances and Breakthrough Innovations

The relationship between alliance partners can be both a source of success and a potential loss, as the success of focal firms often leads to a win–win situation for their partners [31]. Breakthrough innovations based on exploratory alliances allow focal firms and their partners to concentrate on the exploration and research of new knowledge, technologies, and resources.

On the one hand, exploratory alliances, such as R&D consortia and joint ventures, represent strategic alliances structured to combine complementary knowledge bases in developing breakthrough innovations with significant strategic value [16]. These collaborations require dense communication channels and reciprocal knowledge transfer to integrate tacit knowledge, often necessitating relational governance mechanisms to align long-term organizational objectives [18,32]. Focal firms can leverage resources from various fields and industries—such as research and development, design, and technology—by establishing cooperative partnerships with external organizations, which helps them to overcome the technological bottlenecks they face and build new asset barriers. Cross-industry innovation ecosystems enable knowledge recombination and technological convergence through partnerships with organizations beyond traditional value chains [33]. These boundary-spanning alliances create strategic pathways for market diversification while serving as institutional signaling mechanisms in nascent sectors, where collaborative R&D initiatives help to establish technological benchmarks and mitigate uncertainty in emerging product categories [34]. In addition, exploratory alliances resist contractual codification due to their reliance on emergent knowledge recombination and iterative adaption processes [16]. These collaborations involve path creation activities characterized by experimental trial-and-error cycles, where relational trust supersedes formal governance in managing ambiguous technological trajectories and unanticipated co-creation outcomes [32], which helps firms to overcome the “success trap” and “organizational routines”, reducing R&D risks and uncertainties, shortening time-to-market for products, and tapping into new market demand, thus contributing to the emergence of both technological-based and market-based breakthrough innovations. Thus, we put forward the following hypotheses:

H1a. 

An exploratory alliance has a positive effect on a firm’s breakthrough technological innovation.

H1b. 

An exploratory alliance has a positive effect on a firm’s breakthrough market innovation.

On the other hand, exploitative alliances, typically involving limited knowledge-sharing and maintaining contractual boundaries between participating entities, including marketing agreements and licensing arrangements, represent transactional collaborations primarily focused on leveraging pre-existing knowledge and assets to maximize immediate financial gains [16]. First, exploitative alliances help firms to rapidly understand market demand, providing a shortcut for key firms to manage the market effectively and tap into new markets. Exploitative alliances may involve a transfer of tacit knowledge, which requires less knowledge integration between partners, enabling firms to better leverage external experience to improve R&D performance, which is also benefit in guiding firms in the research and development of new technologies and products. For instance, Hoang and Rothaermel [24] found that the interaction between firms and their internal R&D experience positively impacted drug innovation in the biotechnology industry. Second, exploitative alliances operate under low operational interdependence, as contractual governance mechanisms enable firms to fulfill predefined obligations through compartmentalized workflows [18]. These transaction-oriented arrangements leverage modular task architectures, minimizing relational investments while ensuring that predictable resource extraction is aligned with short-term strategic objectives. This allows firms to reduce product costs and improve sales and service [31]. Third, exploitative alliances are also beneficial for absorbing partners’ experiences through trust-based relationships, strengthening a firm’s existing knowledge base and improving the predictability and reliability of its activities [24]. The more intense the cooperation between firms and suppliers, customers, and competitors, the easier it is to understand the product manufacturing process, grasp market demand changes, and keep up with technological advancements in the industry. This, in turn, promotes the transformation of breakthrough technological innovations and market innovations. Thus, we put forward two hypotheses:

H2a. 

An exploitative alliance has a positive effect on a firm’s breakthrough technological innovation.

H2b. 

An exploitative alliance has a positive effect on a firm’s breakthrough market innovation.

2.5. Ambidextrous Alliances and Complementary Assets

The primary goal of strategic alliances is to promote knowledge transfer [31]. Some form of coordination and cooperation is essential due to resource limitations between diverse alliances [27]. When focal firms find it difficult to acquire the necessary resources, they will opt for external learning and coordination form their partners based on their innovation goals, leading to the creation of new complementary assets [27]. Ambidextrous alliances can easily acquire the assets they lack by complementing them with resources from their partners [12]. Thus, exploratory and exploitative alliances help focal firms to internalize new resources and build their own unique complementary assets [26].

On the one hand, first, exploratory alliances expand the cooperation networks of enterprises and enhance the channels of obtaining external resources. Second, by sharing the cost of high-risk projects, exploratory alliances reduce the dependence of enterprises on a single technology path, thus providing a suitable testing ground for the long-term accumulation of complementary assets. Third, exploratory activities of alliances can combine heterogeneous knowledge with resource endowments of focal firms (e.g., cutting-edge technology patents and unverified business models), which can form firms’ new complementary assets.

On the other hand, exploitative alliances focus on deepening the application of existing technologies or markets (e.g., cooperating with mature suppliers to optimize the supply chain), which may improve the utilization efficiency of complementary assets. In addition, exploitative alliances are often generally stable, prompting enterprises to make relationship-specific investments with partners, which are difficult to copy by competitors, thus forming specialized complementary assets. Finally, exploitative alliances often amplify market influence through joint marketing and brand authorization. A partner’s brand reputation and customer base can be transformed into the intangible complementary assets of focal firms.

Therefore, we argue that in an ambidextrous alliance relationship, partners share the complementary assets required for one another’s development and achieve innovation through collaboration, thereby realizing value co-creation among all partners. Based on this, we propose two hypotheses:

H3a. 

An exploratory alliance increases the number of a firm’s complementary assets.

H3b. 

An exploitative alliance increases the number of a firm’s complementary assets.

2.6. Complementary Assets and Breakthrough Innovations

According to PFI theory, complementary assets enable firms to adapt to technological changes and commercialize innovations while maintaining their competitive advantage in the industry [20]. Therefore, complementary assets are considered a set of supporting resources that enable firms to develop new technologies, create new products, and pursue successful commercialization [25].

On the one hand, complementary assets act as a “prism” that reveals the future direction of technological development and helps firms to capture market demand information, thereby enabling more accurate research and development [29]. First, complementary assets facilitate more precise R&D efforts. The use of complementary assets and innovations are mutually reinforcing, requiring dynamic adjustments to resources throughout the innovation process [27]. Second, firms with complementary assets can focus on enhancing their R&D capabilities. Without the necessary R&D capabilities to understand and absorb new technologies, firms cannot effectively utilize their specialized production and marketing assets to commercialize innovations [26]. Third, firms with abundant complementary assets can identify consumers’ latent demands, generate new resources, and develop new capabilities that facilitate exploration.

On the other hand, after the emergence of new technologies, complementary assets serve as a “pipeline mechanism”, helping firms to produce and market new products, thereby seizing opportunities in emerging markets. Firms can leverage complementary assets to control the manufacturing environment, manage sales channels, and establish after-sales service systems for new products, creating a competitive advantage that is difficult for competitors to imitate or replicate [30]. Additionally, complementary assets act as a “buffer mechanism”, enabling firms to identify new business opportunities and mitigate the impact of technological transformation [26]. For example, Roy and Cohen [28] concluded that complementary assets enable firms to understand changes in customer preferences, thereby facilitating the exploration of new markets. Therefore, firms with extensive downstream complementary assets are more likely to become leaders in pursuing breakthrough innovations. Thus, we hypothesize that:

H4a. 

Complementary assets have a positive effect on a firm’s breakthrough technological innovation.

H4b. 

Complementary assets have a positive effect on a firm’s breakthrough market innovation.

2.7. The Mediating Roles of Complementary Assets

Breakthrough innovations demand that firms possess dual capabilities: technical assets for advancing technologies and market assets for commercializing inventions. A deficiency in either can derail innovation; without technical prowess, novel ideas remain unrealized, whereas without market readiness, breakthroughs fail to reach consumers. To overcome resource constraints, firms increasingly engage in ambidextrous alliances (exploratory and exploitative partnerships), which act as conduits for acquiring external knowledge, technologies, and capabilities. These alliances foster bidirectional knowledge exchange [18], enabling firms to access specialized skills critical for breakthrough innovations. Crucially, the complementary assets accumulated through such alliances [26]—spanning manufacturing, marketing, and customer networks—address resource gaps while generating network effects that attract more partners, creating a self-reinforcing innovation loop [20]. Below, we dissect how distinct alliance types drive this process.

On the one hand, first, by collaborating with partners at the innovation frontier, firms assimilate heterogeneous expertise [25], which fuels radical technological creation. Firms can leverage the heterogeneous knowledge and learning experiences of exploratory alliance partners to engage in novel technological creation activities. Second, joint investments in R&D infrastructure (e.g., prototyping labs, patent pools) reduce costs and risks while accelerating commercialization [35]. Third, the breadth of exploratory alliances correlates with the diversity of the complementary assets acquired, directly enhancing the breakthrough potential of focal firms.

On the other hand, exploitative alliances that share production lines and standardized processes have lower operational expenses [16], freeing capital for high-risk R&D. In addition, under exploitative alliances, direct customer feedback (e.g., through co-creation workshops) reveals unmet needs, guiding R&D priorities. Finally, exploitative partnerships deepen market-facing assets like distribution networks or brand equity [16].

Moreover, when alliance partners are willing to share and exchange knowledge, technology, and capabilities, the complementary assets firms acquire can be applied to achieve a balance between disruptive and stable activities [36], further promoting breakthrough innovations. We therefore hypothesize that:

H5a. 

Complementary assets play a mediating role between an exploratory alliance and a firm’s breakthrough technological innovation.

H5b. 

Complementary assets play a mediating role between an exploratory alliance and a firm’s breakthrough market innovation.

H6a. 

Complementary assets play a mediating role between an exploitative alliance and a firm’s breakthrough technological innovation.

H6b. 

Complementary assets play a mediating role between an exploitative alliance and a firm’s breakthrough market innovation.

Figure 1 depicts our research framework.

3. Research Methodology

3.1. Sample and Data Background

We collected data from Chinese firms in high-tech industries, including ICT, aerospace, advanced materials, biomedicine, advanced manufacturing, and so on. There were two reasons for selecting these industries. First, firms in these industries tend to have more alliance partners. Second, these industries typically require higher investment in innovation, making them more likely to pursue breakthrough innovations.

We collected samples from Chengdu, Chongqing, Nanjing, and Wuhan, which are representative cities in China, from June to November 2024. Drawing on the study by Gerbing and Anderson [37], we followed a three-step procedure to conduct our research. First, we designed an English language questionnaire based on a maturity scale and then translated it into Chinese using a double and reverse translation procedure [38]. Second, we conducted a pretest at a university in Chengdu, where 30 eligible MBA students participated. We reviewed and revised the questionnaire based on the pretest results and invited three professors to check the final version. Third, we invited 450 qualified firms to participate in our study via telephone, internet, and email. We received 295 responses, and after excluding responses with obvious errors or missing data, 279 valid questionnaires remained, yielding an effective response rate of 62%.

Table 1. Sample characteristics.

Variables	Categories	Samples	%	Variables	Categories	Samples	%
Industry	ICT	94	33.7	Firm year (year)	<5	67	24
	Biomedicine	30	10.8		5~10	54	19.4
	Aerospace	38	13.6		10~20	74	26.5
	Advanced materials	41	14.7		>20	84	30.1
	Advanced manufacturing	48	17.2	Firm assets (million yuan)	<3	7	2.5
	Others	28	10		3~20	36	12.9
R&D	<1%	18	6.5		20~400	106	38
	1%~3%	28	10		>400	130	46.6
	3%~5%	42	15.1	Ownership	State-owned	68	24.4
	5%~8%	37	13.3		Private-owned	147	52.7
	>8%	154	55.2		Joint venture	44	15.8
					Others	20	7.2

presents the sample characteristics. The sample covers a diverse range of firm ages, sizes, and industries, with a significant number of firms reporting R&D investment between 3% and 8%, making the sample representative.

3.2. Variables and Measures

Dependent variables: Breakthrough innovation. Based on the research by Zhou et al. [5], what needs to be pointed out is that a reverse item of breakthrough technological innovation (BTI) showed poor reliability and validity in the pretest, and we finally adapted the scale to three items focused on technological breakthroughs to align with the Chinese high-tech context after discussion with three professors. Thus, we included three items to measure breakthrough technological innovations and four items to measure breakthrough market innovations.

Independent variables: Ambidextrous alliances. Drawing on the work by Kauppila et al. [14], we designed four items to measure exploratory alliances and four items to measure exploitative alliances.

Mediating variable: Complementary assets. We used a five-item scale adapted from Christmann’s study [39] to measure complementary assets.

Control variables: Previous research suggested that several factors may influence a firm’s breakthrough innovation. Building on the study by Jin et al. [30], we selected firm age, R&D expenditure, number of firm assets, and number of alliances as our control variables.

3.3. Reliability, Validity, and Common Method Bias (CMB)

The ages and scales of the research samples were diverse, and the types of industries were rich. The number of samples with R&D investment of 3%~8% was large and could be considered representative.

Table 2. Reliability and convergent validity.

Items	Loadings
Exploratory alliance (ERA) Cronbach’s α: 0.838; CR: 0.834; AVE: 0.561
We will seek cooperation with universities, research institutes, and other relevant departments in conducting innovation activities.	0.849
The purpose for our firms to form alliances is to carry out pioneering work.	0.850
The main motivation for our firms to form alliances is to find new opportunities.	0.657
Through inter-organizational cooperation, we have realized innovative activities that we cannot carry out on our own.	0.609
Exploitative alliance (EIA) Cronbach’s α: 0.800; CR: 0.800; AVE: 0.501
We simplify or optimize business operations through subcontracting cooperation with alliance partners.	0.720
Suppliers play an important role in developing new products or providing new services.	0.745
By cooperating with the alliance, we have improved the cooperation efficiency among organizations.	0.696
The main motivation for our firms to form alliances is to obtain complementary resources.	0.667
Complementary assets (CA) Cronbach’s α: 0.891; CR: 0.878; AVE: 0.596
Compared with competitors, we are more concerned about taking the lead in trying new methods and technologies in the industry.	0.846
Compared with our competitors, we pay more attention to using the latest production technology.	0.890
Compared with our competitors, we are more concerned about investing in new equipment and mechanical equipment.	0.816
In the past three years, we have been a leader in product innovation.	0.589
In the past three years, we have been a leader in process innovation.	0.679
Breakthrough technological innovation (BTI) Cronbach’s α: 0.796; CR: 0.806; AVE: 0.583
Our products are highly innovative and have replaced inferior substitutes.	0.798
Our products incorporate brand new technical knowledge.	0.835
Our technology/products are completely different from those of our main competitors.	0.645
Breakthrough market innovation (BMI) Cronbach’s α: 0.859; CR: 0.861; AVE: 0.608
Our product/technology concept is difficult for mainstream customers to evaluate or understand.	0.704
Our products involve higher switching costs for mainstream customers.	0.806
Our products need mainstream customers to learn how to use them.	0.816
It takes a long time for mainstream customers to understand the full benefits of our products.	0.787

shows the reliability and convergent validity. We chose the Harman one factor analysis method to examine the common method bias (CMB). Observing that the explanatory variance of the first component factor after rotation in this study was 17.507%, which was lower than the threshold of 30%, it was considered that there was no serious CMB in our samples [40].

4. Results

4.1. Correlation Analyses

Table 3. Descriptive statistics and correlations.

Variables	1	2	3	4	5	6	7	8	9
1. Firm age	-
2. Firm assets	0.422 **	-
3. R&D	−0.044	0.164 **	-
4. Alliance number	0.119 *	0.275 **	0.084	-
5. ERA	−0.072	0.082	0.174 **	0.151 *	0.749
6. EIA	−0.046	0.087	0.120 *	−0.022	0.286 **	0.708
7. CA	−0.03	0.075	0.273 **	0.161 *	0.656 **	0.370 **	0.772
8. BTI	−0.118 *	0.036	0.169 **	0.079	0.497 **	0.451 **	0.485 **	0.764
9. BMI	−0.047	0.083	−0.018	0.036	0.276 **	0.324 **	0.308 **	0.301 **	0.78
Mean	2.630	3.290	4.010	2.920	5.772	4.933	5.493	5.012	4.229
S.D.	1.149	0.784	1.300	1.357	1.054	1.048	1.081	1.072	1.388
Min	1	1	1	2	1	1	1	1	1
Max	4	4	5	6	7	7	7	7	7

Notes: * p < 0.05, ** p < 0.01; Numbers on the diagonal indicate the square root of AVE.

depicts the descriptive statistics and correlations of all variables in our research. The results show that there is no excessive correlation exists between variables, and the main variables indicate positive significant correlations.

4.2. Hypotheses Test

Hypotheses 1a~1b and Hypotheses 2a~2b posit that ERA and EIA have positive effects on BTI and BMI, respectively. Models 2 and 5 in

Table 4. Regression results.

Variables	BTI			BMI			CA
	Model1	Model2	Model3	Model4	Model5	Model6	Model7	Model8
Firm age	−0.142 *	−0.071	−0.078	−0.106	−0.061	−0.068	−0.041	0.039
Firm assets	0.052	−0.014	−0.005	0.131	0.084	0.094	0.012	−0.051
R&D	0.149 *	0.059	0.032	−0.045	−0.102	−0.131	0.257 ***	0.154 **
Alliance Number	0.069	0.036	0.023	0.016	0.003	−0.011	0.141 *	0.076
ERA		0.382 ***	0.282 ***		0.205 **	0.100		0.568 ***
EIA		0.333 ***	0.298 ***		0.267 ***	0.231 ***		0.197 ***
CA			0.175 *			0.185 *
R2	0.049	0.360	0.376	0.017	0.155	0.172	0.095	0.495
adj-R2	0.035	0.346	0.36	0.003	0.136	0.151	0.082	0.484
F	3.550 *	25.536 ***	23.310 ***	1.204	8.306 ***	8.053 ***	7.189 ***	44.442 ***

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

show that ERA is significantly and positively related to BTI (β = 0.382, p < 0.001) and BMI (β = 0.205, p < 0.001). EIA is significantly and positively related to BTI (β = 0.333, p < 0.001) and BMI (β = 0.267, p < 0.001). Hence, Hypotheses 1a~1b and Hypotheses 2a~2b are supported.

Hypotheses 3a~3b posit that ERA and EIA have positive effects on CA. Model 8 in Table 4 shows that ERA and EIA are significantly and positively related to CA (β = 0.568, p < 0.001; β = 0.197, p < 0.001), which means Hypotheses 3a~3b are supported.

Hypotheses 4a~4b posit that CA has a positive effect on BTI and BMI, respectively. Models 3 and 6 in Table 4 show that CA is significantly positively related to BTI (β = 0.175, p < 0.05) and BMI (β = 0.185, p < 0.05). Thus, Hypotheses 4a~4b are supported.

Following the suggestion of Baron and Kenny [41], the mediation effect test was divided into three steps. First, test whether the total effect of independent variable X on dependent variable Y is significant. Secondly, test whether the independent variable X has a significant influence on the intermediate variable M. Finally, after controlling the influence of independent variable X, test whether the effect of intermediate variable M on dependent variable Y is significant. If the direct effect of X on Y is smaller (but still significant) than the total effect, it means that there is a partial mediation effect. If the direct effect of X on Y is not significant, it represents that there is a complete mediation effect.

Comparing models 2 and 3, we can find that after adding CA, the positive influences of ERA and EIA on BTI are still significant, but the effects decrease (β_ERA decreases from 0.382 to 0.282; β_EIA decreases from 0.333 to 0.298). Combined with the previous discussions, we can conclude that CA partially mediates the relationships between ERA, EIA, and BTI, respectively, which proves that Hypotheses 5a~5b are supported. Comparing models 4 and 5, we can find that after adding CA, the positive effect of ERA on BMI is not significant (β_ERA decreases from 0.205 to 0.100). The positive influence of EIA on BMI is still significant, but the effect decreases (β_EIA decreases from 0.267 to 0.231). Combined with the previous discussion, we can conclude that CA completely mediates the relationships between ERA and BMI and partially mediates the relationship between EIA and BMI. Thus, hypotheses 6a~6b are verified.

4.3. Robustness Test

Following the suggestion of Hayes [42], all of the effects were re-tested using the bootstrapping method in the “PROCESS” plug-in of SPSS 24.0. We selected “Model 4”, and set the confidence interval as 95% and the sample size as 5000. The

Table 5. Bootstrapping results.

Paths	Type	Effects	S.E.	Bootstrap (95% CI)
				Lower	Upper
Path1:	Total effect	0.482	0.055	0.375	0.59
ERA→CA→BTI	Indirect effect	0.173	0.058	0.059	0.288
	Direct effect	0.309	0.069	0.173	0.445
Path2:	Total effect	0.368	0.078	0.213	0.522
ERA→CA→BMI	Indirect effect	0.214	0.060	0.099	0.333
	Direct effect	0.153	0.100	−0.043	0.349
Path3:	Total effect	0.446	0.055	0.337	0.554
EIA→CA→BTI	Indirect effect	0.126	0.042	0.048	0.211
	Direct effect	0.319	0.055	0.211	0.428
Path4:	Total effect	0.427	0.076	0.277	0.577
EIA→CA→BMI	Indirect effect	0.115	0.035	0.051	0.191
	Direct effect	0.312	0.080	0.155	0.469

shows the results, indicating that all hypotheses are supported again, which represents our results are robust.

5. Discussion

First, our findings revealed that both exploratory and exploitative alliances exert significant positive effects on firms’ breakthrough innovations, albeit through distinct reasons. Exploratory alliances enable firms to transcend internal resource constraints by engaging in high-risk, high-reward collaborations with universities, research institutes, and R&D-intensive partners. For instance, joint R&D activities accelerate the development of novel technologies. In addition, firms gain access to untapped customer segments by pooling resources with partners possessing complementary market insights. These influences promote the development of breakthrough technological and market innovations. Meanwhile, exploitative alliances, often forged with suppliers, distributors, or established industry players, focus on refining existing capabilities to amplify breakthrough innovation returns. Firms’ close collaboration with customers (e.g., co-design workshops) provides real-time feedback, allowing firms to align R&D with latent market needs. Furthermore, shared production infrastructure and standardized processes lower operational costs, freeing capital for reinvestment in high-impact R&D. These impacts also accelerate the development of breakthrough technological and market innovations. These results align with ambidexterity theory, affirming that firms balancing exploratory and exploitative activities achieve superior innovation outcomes by simultaneously pursuing novelty and efficiency [43].

Second, the results show that ambidextrous alliances serve as conduits for acquiring and integrating complementary assets—resources critical for bridging the gap between invention and commercialization [28]. Collaborations with ambidextrous alliances yield complementary assets. From a technological aspect, focal firms can access cutting-edge scientific knowledge to fill internal R&D gaps and pursue breakthrough technological innovations through exploratory alliances. Furthermore, jointly developed advanced equipment reduces capital expenditure while accelerating prototyping through exploitative alliances. From a market aspect, partnerships with upstream firms or manufacturers (exploratory alliances) provide brand power to enhance consumer trust. Meanwhile, partnerships with downstream firms or retailers (exploitative alliances) expand geographic reach to obtain access with instant market changes. This dual pathway explains why even firms with leading technologies still fail without complementary assets: breakthroughs like Nokia’s early smartphone prototypes faltered due to inadequate manufacturing ecosystems and distribution channels.

Third, complementary assets play a complete mediating role between exploratory alliances and breakthrough market innovations. And they also play a partial mediating role between exploratory alliances and breakthrough technological innovations, as well as between exploitative alliances and both breakthrough technological and market innovations. These findings provide new insights into why firms with leading technologies may still fail to realize innovation benefits. On the one hand, the key to achieving breakthrough innovations lies in building competitive barriers through complementary assets. Complementary assets not only help firms to develop new technologies but also facilitate the commercialization of those technologies and mitigate the disruptive impacts of innovation. Through ambidextrous alliances, firms acquire new technologies, products, sales channels, supply chains, manufacturing management, and other essential resources, thereby securing complementary assets. On the other hand, our findings show that the positive impact of exploratory alliances on breakthrough market innovations is fully mediated by complementary assets. This supports the view that complementary assets also play a crucial role in a firms’ technological innovation [28]. Our research demonstrates that even if a firm possesses advanced technology, it must rely on complementary assets to drive breakthrough market innovations and successfully commercialize its technology.

6. Conclusions

6.1. Theoretical Contributions

First, we expand the research on PFI theory in the context of breakthrough innovations, which demand concurrent advancements in technology and markets. Prior PFI research emphasized value capture through complementary assets in a specific technological context (e.g., [44], the digital environment), often overlooking the interplay between radical technological shifts and novel market strategies. By integrating PFI with breakthrough innovations’ dual requirements, we address this gap. Unlike studies focused mainly on breakthrough technological innovations (e.g., [36]), we argue that firms struggling with internal capabilities must leverage both exploratory and exploitative alliances to secure heterogeneous resources—a valuable adjustment to the PFI framework’s focus on internal asset control [20].

Second, we reconcile exploration–exploitation tensions in ambidextrous alliances. While the ambidexterity literature highlights the risks of balancing exploratory and exploitative activities [16], our study demonstrates that strategic alliances mitigate these challenges through structured resource partitioning. Our results contribute to understanding how firms achieve breakthrough innovations, complementing Capponi et al.’s [7] work on the pathways to pursuing breakthrough innovations by firms. Furthermore, existing research related to ambidexterity suggests that exploration and exploitation have distinct effects on strategic alliances [23]. When pursuing breakthrough technological and market innovations simultaneously, firms may risk “resource depletion” [45]. Our work thus complements Rothaermel and Deeds’ [18] alliance typology, clarifying how distinct alliance types differentially channel exploration and exploitation to pursue disruptive changes.

Third, this work investigates the factors that influence the acquisition of complementary assets, thereby deepening the understanding of how these assets impact firms’ breakthrough innovations [25]. PFI theory focuses on how firms use complementary assets to profit from innovation, but few studies have explored how firms acquire these assets. We reveal that ambidextrous alliances serve as primary conduits for acquiring technical assets (e.g., patents via exploratory partnerships) and market assets (e.g., distribution networks via exploitative alliances). Furthermore, previous research has confirmed the pivotal role of complementary assets in incumbent advantages [25]. We advance PFI theory into a dynamic framework to illustrate how different alliance types enable technological and market breakthroughs through complementary assets.

6.2. Implementation Suggestions

First, managers should leverage the differentiated strengths of ambidextrous alliances to target specific innovation outcomes. Focal firms may partner with universities, R&D labs, or tech startups to access cutting-edge knowledge while engaging suppliers, distributors, or customers to refine commercialization. Our results provide empirical evidence for focal firms to choose different types of alliance partners based on their innovation needs. For example, managers should conduct a partner audit to map alliance candidates against innovation priorities (technological vs. market), and implement stage–gate processes to align exploratory alliances with early R&D phases and exploitative alliances with scaling stages.

Additionally, managers should enhance the unified leadership of the alliance to drive breakthrough innovations and overcome weaknesses in resource allocation and distribution. By establishing well-organized, division-of-labor structures, firms can form a cooperative mechanism based on complementary resources and value creation and build an alliance network centered on breakthrough innovations. Thus, during the breakthrough innovation process, managers should appoint a cross-functional alliance manager to oversee coordination, resolve conflicts, align goals, and prioritize partners offering non-overlapping resources.

Finally, complementary assets are central to enabling firms to realize technological R&D innovation, transform new technologies into innovative outputs, and achieve commercialization success. Managers should recognize the key role of complementary assets in building barriers to breakthrough technologies and markets while maintaining the firm’s sustainable competitive advantage. Therefore, focal firms should not only focus on acquiring complementary assets through different alliances during the commercialization phase but should also invest in them during the R&D and design stages to lay the foundation for long-term and sustainable breakthrough technological and market innovation success.

6.3. Limitations and Future Prospects

Our study has several limitations. First, existing research has started to explore the balance and interaction of firms’ ambidextrous alliances from both balanced and combined ambidextrous perspectives. These two viewpoints are valuable for a comprehensive understanding of how firms allocate resources and differentiate between the two types of alliances when simultaneously engaging in exploratory and exploitative alliances. Second, this study relied solely on a cross-sectional questionnaire survey, without considering a phased follow-up survey. Since innovation activities unfold over time, the need for selecting alliance partners and acquiring complementary assets may vary at different stages. Future research could consider using time-series analysis to gain deeper insights into the innovation process. Third, the boundary conditions of this mechanism were not considered completely. Considering China’s unique institutional environment as the boundary conditions (e.g., state subsidies, guanxi networks) may be an interesting topic in future research. Fourth, although the robustness test using the “PROCESS” plug-in has been studied [42], this test method may have potential limitations. Future studies can use second-hand data to carry out more objective and diversified robustness tests and obtain more reliable research results.