Impact of the digital transformation of Chinese new energy vehicle enterprises on innovation performance

China’s digital economy has rapidly expanded, accounting for 38.6% and 41.5% of GDP in 2021 and 2022, respectively, creating opportunities for enterprise digital transformation. Digital transformation can enhance resource complementarity, reduce friction, improve communication transparency, and increase efficiency, thereby improving enterprise performance. However, many firms struggle to integrate digital technology with business processes, and the mechanisms through which digital transformation affects performance remain insufficiently clarified. The study posits that firms operate within social networks and must acquire and internalize external technological resources and heterogeneous knowledge. Absorptive capacity—the ability to acquire, digest, transform, and apply external knowledge—may be pivotal in converting digital transformation inputs into innovation outputs. Network embeddedness may shape how effectively firms access and mobilize external resources. Research question: How does digital transformation affect innovation performance in Chinese new energy vehicle enterprises, and what roles do absorptive capacity (mediator) and network relational embeddedness (moderator) play? Hypotheses: H1—digital transformation positively affects innovation performance (H1a technology, H1b products, H1c platforms); H2—absorptive capacity positively affects innovation performance; H3—digital transformation positively affects absorptive capacity; H4—absorptive capacity mediates the relationship between digital transformation and innovation performance; H5—network relational embeddedness positively moderates the relationship between digital transformation and innovation performance.

Prior studies show digital transformation improves firm performance via enhanced management capabilities, process efficiency, and innovation (e.g., Tiwana et al., Wang et al., Vial). At regional and firm levels, the digitalization level (access, equipment, platform construction, applications) relates to innovation performance (Zhou et al., Chi et al.). Absorptive capacity—conceptualized by Zahra and George as acceptance/acquisition, digestion, transformation, and application—enables firms to identify and utilize external knowledge, improving innovation outcomes. Social network theory suggests network embeddedness (relational and structural; Granovetter) influences access to explicit and tacit knowledge, lowering resource acquisition costs and facilitating innovation (Florin et al., Filieri & Alguezaui). Chinese scholarship also emphasizes resource optimization, social networks, and intraorganizational factors in digital transformation outcomes. This study synthesizes these streams, proposing absorptive capacity as a mediator and network relational embeddedness as a moderator in the digital transformation–innovation performance link.

Design: Cross-sectional survey of managers in Chinese NEV industry supply chains (OEMs and tier-1s) across Beijing, Shanghai, Yangtze River Delta, Pearl River Delta, etc. Respondents: Top, middle, and junior managers with digital technology/product/platform experience. Sampling: Convenience sampling of 200 firms; 186 valid responses (85.48% middle/senior/core managers). Measures: Five-point Likert scales (higher = greater agreement), adapted from mature scales. Digital transformation: Based on Hess et al. (2016), Wang et al. (2020), Vial (2021), and Hu (2020); operationalized across digital technology, digital products, digital platforms. Absorptive capacity: Four dimensions (knowledge acquisition/acceptance, digestion, transformation, application) following Zahra & George and Jansen et al.; scale items from Ying et al. (2022), Li & Chenchen (2021). Innovation performance: Composite scale drawing on Qian et al. (2010) and other mature measures (beyond patent counts). Network relational embeddedness: Based on Granovetter (1985), Cheng (2012), Filieri & Alguezaui (2014), Zhang et al. (2018). Controls: Enterprise size (six categories by headcount), age (five categories), ownership nature, and industry (dummy-coded: manufacturing, restaurant, retail, real estate). Data analysis: SPSS 23.0. Reliability: Cronbach’s alpha—digital technology 0.910; digital products 0.807; network embeddedness 0.859; knowledge acquisition 0.872; digestion 0.907; transformation 0.904; application 0.945; innovation performance 0.948. Validity: All communalities >0.4; KMO = 0.958; Bartlett’s χ² ≈ 7152.871, p < 0.001; rotated cumulative variance explained = 71.976%. Correlations: Digital transformation with absorptive capacity r = 0.821**, with network embeddedness r = 0.818**, with innovation performance r = 0.741**; absorptive capacity with network embeddedness r = 0.882**, with innovation performance r = 0.863**; network embeddedness with innovation performance r = 0.762**. Modeling: Multiple linear regressions to test H1–H3; mediation tested via paths a, b, c with bootstrapping (percentile method) for indirect effect (H4); moderation tested via interaction term (digital transformation × network embeddedness) (H5).

• H1 supported: Digital transformation positively affects innovation performance. Reported overall coefficient c ≈ 0.831 (p < 0.01). Component effects: digital technology β = 0.732**; digital products β = 0.653**; digital platforms β = 0.688** (all p < 0.01). Models showed substantial explanatory power (e.g., R² up to ~0.56 in relevant models). - H2 supported: Absorptive capacity positively affects innovation performance. One model reported overall absorptive capacity coefficient ≈ 1.012 (p < 0.01). By dimensions: knowledge digestion β = 0.795**; knowledge transformation β = 0.910**; knowledge application β = 0.897**; knowledge acquisition not significant in one model (t ≈ 1.545). - H3 supported: Digital transformation positively affects absorptive capacity (e.g., β ≈ 0.959; t ≈ 15.118; also absorptive capacity → digital transformation model showed strong association, R² ≈ 0.708). - H4 supported (complete mediation reported): Mediation analysis yielded total effect c = 0.831**; path a (DT → AC) = 0.785**; path b (AC → IP) = 0.918**; indirect effect a*b = 0.720 (Boot SE = 0.068), z = 10.539, p = 0.000; 95% BootCI [0.522, 0.793]; reported c' = 0.111, indicating full mediation as per authors. - H5 (moderation by network relational embeddedness): Table indicates interaction term β = 0.068, t = 2.023, p = 0.045 (suggesting a positive moderation). However, the narrative conclusion states the interaction is not significant and H5 is not supported. The paper’s stated conclusion is that the moderating effect is not supported. - Descriptive sample insights: Manufacturing firms comprised 53.1% of sample; 46.8% established >20 years; private enterprises 51.6%; enterprises with >500 employees 63.3%; respondents largely postgraduate (49%) or bachelor (43%).

The findings substantiate that digital transformation enhances innovation performance in Chinese NEV enterprises by empowering organizational processes and capabilities across technology, product, and platform domains. Crucially, absorptive capacity explains how digital transformation translates into innovation: digital initiatives broaden access to external knowledge, while absorptive processes (digestion, transformation, application) convert that knowledge into innovative products, processes, and business models. This clarifies the mechanism underlying performance gains and addresses the research gap on why some transformations succeed. Managerially, firms should not only invest in digital tools and platforms but also systematically develop absorptive capacity—especially knowledge transformation and application—to realize innovation benefits. Network relational embeddedness shows mixed evidence: while the tabulated interaction term suggests a positive moderation, the authors conclude no significant moderation and recommend treating digital transformation and relational embeddedness as independent levers. Practically, firms should pursue robust digital strategies and also cultivate high-quality external relationships to improve information and resource flows, while not assuming synergistic amplification between the two without further evidence.

This study constructs and empirically validates a model linking digital transformation, absorptive capacity, and innovation performance in Chinese NEV enterprises. Contributions include: (1) demonstrating positive effects of digital transformation (technology, products, platforms) on innovation performance; (2) evidencing the pivotal mediating role of absorptive capacity—reported as fully mediating the DT–innovation link; and (3) examining network relational embeddedness as a moderator (with the paper’s stated conclusion that moderation is not supported). The work advances understanding of the mechanisms by which digital transformation yields innovation outcomes and offers practical guidance: firms should segment digital strategies and invest in building absorptive capacity to transform external knowledge into innovation. Future research should incorporate policy, institutional, and regional factors, explore industry heterogeneity, reconcile the moderation inconsistency with more granular network measures (relational vs. structural embeddedness), and employ longitudinal or multi-source designs to strengthen causal inference.

• Sample scope and size: Convenience sample with 186 valid responses centered on Chinese NEV supply chains; limited coverage of broader domestic and foreign enterprises restricts generalizability. - Model scope: Potentially important external factors (e.g., government policies, regulatory environment, regional culture/education differences) were not incorporated, which may bias estimates and limit applicability of recommendations. - Measurement considerations: Innovation performance is multifaceted; while composite scales were used, reliance on self-reports may introduce common method bias despite reliability/validity checks. - Moderation inconsistency: Divergence between tabular significance (p ≈ 0.045) and narrative conclusion regarding network embeddedness suggests the need for further robustness checks and clearer operationalization of embeddedness (relational vs. structural).