This dissertation undertakes to cast a new and discerning eye on the continuing enigma of Canadian federal innovation policy. Towards that end the dissertation employs the Developmental Network State (DNS) framework and pursues the question of whether the DNS can be used to explain why Canada - and specifically federal innovation policy since 2000 - has seemingly been unable to increase national Research and Development investments and positively impact Canada’s rate of innovation. Theoretically the dissertation argues that the DNS framework can be used to mend the gap between National Innovation Systems literature with its proclivity for an undertheorization of the state and historical institutionalism, which while robustly conceptualizing the state, has been less engaged with the innovation problématique. While admittedly including some modest conceptual sharpening of the framework, the argument draws on four key pillars of the DNS: policy durability over time; targeted resourcing that can enable innovation; thickening of triadic networks among business, academe and government and incentives for capitalization. The research methodology encompasses both qualitative and quantitative (particularly in the sense of economic) techniques and includes 54 in-depth interviews conducted with innovation leaders across the nation. With respect to the evidence assembled around the health of the Canadian innovation system, OECD indicators over the period of the 2000s generally show deteriorating trends, although investment in Higher Education R&D (HERD) is very much the stalwart exception. As for the four lines of inquiry investigated regarding federal policy, the issue of policy durability in large part reveals an increasing number of priorities overlaid at times with exceedingly ambitious objectives and the ongoing challenge of fashioning a more enduring federal strategy. As for targeted resourcing, while federal investments have been both sustained and substantive and while there has been particular attention to macroeconomic stability and research infrastructure, there has been a continuing pattern of oscillation between a focus on research and one on commercialization. With respect to triadic thickening of networks, the evidence continues to reveal the relatively shallow nature of collaboration - this despite robust funding, institutional mechanisms for networking and an ongoing priority on this issue by government. Here again however business financing of HERD stands out as the exception. In contrast, on incentives for capitalization there has been significant progress, although access to late-stage capital remains a challenge. Additionally, the study undertakes an examination of two sectors - Artificial Intelligence (AI) and pharmaceutical manufacturing, employing a “quasi-experimental” technique that examines impacts of policy initiatives over a period that dates back to 1987. In the case of AI, the enabling conditions of DNS do coalesce to form a robust innovation ecosystem. In the case of pharma, policy efforts have not been so sustained or holistic and innovation indicators duly tell this tale of deteriorating trends. In sum, what emerges is that federal innovation policy in the Canadian context has developed in rather truncated form - and it has tracked within the confines of more historically adept and enduring strategies such as upstream Research and Development and fiscal policy. Indeed, the pillars of the DNS framework serve to illuminate the dissonance between policy intent and impact, highlighting the unique nature of federal innovation policies as they have endeavoured to establish the agendas, funding, networks and capital that can provide a formula for advancing Canada on this front. This case study also suggests that the DNS itself may need honing - specifically that policy durability is not only a sufficient condition for innovation but in fact an anchoring one enabling other dimensions to follow suit.