Capital, Sovereignty, and the AI Infrastructure Arms Race: An Economic and Geopolitical Analysis of Trillion-Dollar Multinational Compute Investments

This dissertation investigates the economic and geopolitical dimensions of the unprecedented capital deployment into hyperscale artificial intelligence (AI) infrastructure between 2024 and 2026. During this period the four largest United States hyperscalers advanced toward aggregate annual capital expenditure that secondary estimates place between $720 billion and $745 billion for 2026, while headline commitments such as the up-to-$500 billion Stargate platform, a reported $300 billion Oracle-OpenAI compute arrangement, and a staged up-to-$100 billion NVIDIA-OpenAI framework signaled a structural shift in how compute is financed, sited, and governed. The study frames this build-out as a simultaneously economic and geopolitical phenomenon in which sovereign wealth capital, industrial policy, export controls, and supply-chain concentration intersect with private capital allocation. Employing a quantitative-dominant, sequential mixed-methods design, the research combines panel econometric analysis, input-output modeling, event studies, and scenario analysis of firm- and region-level data with structured comparative case studies of six flagship projects: the Microsoft-OpenAI partnership, the Oracle-OpenAI compute deal, the Stargate consortium, Meta's hyperscale expansion, the aggregate hyperscaler capital-expenditure surge, and the NVIDIA-OpenAI graphics-processing-unit investment. All data are drawn from publicly available sources retrievable without registration or fees, including securities filings, earnings transcripts, national economic accounts, energy statistics, cross-national indicators, construction data, executive orders, and patent records. Four research questions organize the inquiry, addressing macroeconomic and regional effects, the geopolitics of sovereign capital, the national-security implications of accelerator concentration, and the relationship between capital expenditure and equity valuation. An integrative theoretical framework unites geopolitical economy, hegemonic stability theory, Schumpeterian creative destruction, and platform economics to treat compute as strategic infrastructure. A central methodological commitment is evidence discipline: every quantitative observation is classified by evidence status and never aggregated across categories without adjustment. The study contributes a rigorous, replicable, evidence-status-aware account of the largest sustained wave of private infrastructure investment in modern economic history.