Cross-border venture capital in the semiconductor industry: Knowledge spillover or capital deepening?

Cross-border venture capital (CB VC) is widely assumed to accelerate technological diffusion, yet whether it generates knowledge spillovers or primarily reallocates financial capital remains empirically unresolved. Drawing on 13,362 semiconductor VC investment rounds worldwide (1981–2024), from which we identify 3,363 cross-border rounds across roughly 470 investor→target corridors and analyse the 47 corridors with complete covariates, merged with USPTO patent data, this study contrasts two mechanisms: knowledge spillover through technological transfer and capital deepening through market-driven investment dynamics. Investment patterns are shaped by target-country location advantages and, with comparable weight, by investor-country technological capability; that ownership advantage, however, does not translate into recipient-side innovation gains. The raw positive association between CB VC and patenting is not detectable under within-corridor fixed effects and is not restored when innovation is measured across all major patent offices (CNIPA, EPO/WIPO, and others) rather than the USPTO alone. This null holds across patent counts, technological breadth, and citation impact: the estimates rule out average effects larger than roughly 0.10 log points while remaining compatible with smaller ones, a pattern consistent with absorptive-capacity constraints in knowledge assimilation. Firm-level patterns are consistent with investor selection, although cross-sectional firm data cannot by themselves exclude localized micro-level spillovers, and moderation analyses find no robust absorptive-capacity threshold above which spillovers emerge—though the small number of highly capable treated recipients leaves that test underpowered. Industrial-policy shocks are followed by gains in domestic semiconductor patenting—most robustly in the number of patenting firms—with no detectable moderation by CB VC, while cross-border deal activity is strongly persistent, driven primarily by returning and previously connected investors rather than first-time entrants: a pattern consistent with financial-market network dynamics, not technological diffusion. These findings reframe the function of CB VC from knowledge transfer to capital allocation. The study contributes to research on technological change in global innovation systems and offers policy implications, confined to the technology-diffusion channel, for the governance of frontier technologies.