Appointing practicing scientists to corporate leadership boosts firms’ science-based patents and attracts top inventors; meanwhile, mission-oriented public science programs such as the BRAIN Initiative catalyse venture capital into neurotech by lowering technical uncertainty and supplying talent and complementary technologies.

Main Finding

Yufeng Yao (PhD thesis, UNSW Business School, Jan 2025) shows that embedding active scientists in corporate governance and executive roles and supplying mission-oriented public science both materially improve the commercialization of scientific discovery. Specifically: (1) Scientific Directors (SciDs) on boards and Scientific Executives (SciExecs) inside firms raise the quantity and economic quality of firm innovation—especially in technology areas where those scientists are research-active—and (2) mission-oriented basic-science funding (the BRAIN Initiative) reduced technical uncertainty and spurred follow-on venture-capital investment, higher valuations, and more successful exits for neurotech startups.

Citation: Yao, Y. (2025). Essays on Corporate Innovation and Commercializing Scientific Discovery. UNSW PhD thesis. DOI: https://doi.org/10.26190/unsworks/32688 (CC BY 4.0).

Key Points

• SciDs (outside directors with scientific PhDs/publications) increase firms’ reliance on fundamental science and improve innovation quality (more breakthrough patents, more citations, higher-value patents) — effects concentrated in tech areas matching the SciD’s expertise.
• SciDs operate through two main channels: (i) scientific knowledge transfer (firm patents increasingly cite the SciD’s area/publications) and (ii) professional networks (SciDs help firms access high-caliber inventors within scientific communities).
• Endogeneity/addressing selection: results are supported using director×firm fixed-effects, an instrument based on local SciD supply, and a quasi-exogenous shock (Human Genome Project) that increased SciD demand in pharma.
• Mission-oriented public funding (the U.S. BRAIN Initiative) raised VC activity in neurotech: more deals, larger funding rounds, higher valuations, and better exit outcomes. Mechanisms include greater supply of skilled academic labor, more breakthrough inventions, and improved integration of neurotech with complementary technologies (AI, big data).
• SciExecs (CTO-like executives with publication records) are associated with more science-based and government-backed patents; patents citing SciExec publications are more valuable. Neurotech firms were more likely to appoint SciExecs after the BRAIN Initiative, and those firms produced better innovations.
• Empirical measures of innovation: patent counts, breakthrough patents, forward citations, patent market-value proxies, and whether patents are classified as “fundamental” (science-sourced).
• Robustness checks include alternative neurotech definitions, pre-trend tests, and exclusions (e.g., removing AI/big-data startups where relevant).

Data & Methods

• Data sources: corporate board and executive data; patent datasets (patent texts, citations, classifications); scientific publication data; government grants (NIH/BRAIN Initiative) and award records; VC investment and exit databases; employee/inventor records and name-matching routines.
• Machine learning / NLP:
  • SciBERT / BERT-style models to map and match patents to scientific subject areas and to classify whether patents derive from specific scientific domains.
  • Supervised classifiers for identifying neurotech startups and for assessing the commercial potential of scientific grants/papers.
• Network / community methods:
  • Louvain algorithm and other community-detection techniques to construct scientific/professional communities and inventor networks around SciDs.
• Identification strategies:
  • Within firm–SciD (director × firm) fixed-effects to capture time-varying influence of SciD appointments on firm patents.
  • Instrumental variables: local SciD supply used as an instrument for SciD appointment to address selection.
  • Event-study / difference-in-differences for the BRAIN Initiative (pre/post BI comparisons between neurotech and comparison healthcare sectors), with parallel-trend checks and multiple robustness specifications.
  • Human Genome Project treated as an exogenous shock increasing demand for SciDs in affected industries.
• Outcome metrics:
  • Patent counts, fundamental-patent share, forward citations, breakthrough-patent indicators, patent value measures (market proxies), VC funding size and valuations, exit success (acquisitions/IPOs).

Implications for AI Economics

• Scientific human capital matters for firm adoption and commercialization of frontier technologies. For firms seeking to commercialize AI-enhanced science (e.g., AI-driven drug discovery, neurotech coupled with ML), appointing SciDs and SciExecs can materially improve innovation outcomes and technology integration.
• Mission-oriented public science funding (targeted R&D programs) can catalyze private investment in emerging AI–science hybrids by reducing early technical uncertainty and expanding the pool of scientifically trained labor ready to apply AI tools — a lever for policymakers to steer private capital into socially valuable, deep-technology areas.
• VC markets respond to reductions in technical uncertainty: combining public funding signals and scientific talent increases private financing and valuations. For AI economics this implies that government backing of foundational datasets, benchmarks, and open scientific resources (analogous to the BRAIN Initiative’s data-driven priorities) can accelerate AI commercialization in scientific domains.
• Methodological takeaway: LLMs and domain-specific transformers (e.g., SciBERT) are effective tools for mapping linkages between scientific research and patented commercialization — useful for researchers and policymakers monitoring AI’s diffusion across science-driven industries.
• Corporate governance and executive composition are policy-relevant levers. As AI systems become central to scientific discovery, firms that integrate academic-scientist expertise into governance/executive roles are better positioned to capture value from science–AI complementarities.
• Practical recommendations:
  • Firms commercializing science + AI should consider recruiting active scientists into decision-making roles (board or executive) to access frontier knowledge and networks.
  • Policymakers designing mission-oriented R&D should pair grant funding with programs that foster academia–industry mobility and data/compute infrastructure to maximize private follow-on investment.
  • VCs evaluating AI-for-science startups should weigh signals from government-funded science and academic co-founders/publications as indicators of reduced technical risk.

If you want, I can extract key tables/figures and produce a one-page visual summary (charts or bulletized evidence) for each chapter.