Lead, Own, Share
Sovereign Wealth Funds for Transformative AI
By
Liam Epstein
July 08, 2025
This paper proposes sovereign wealth funds as a foundational policy instrument for states to navigate the economic and strategic challenges posed by transformative artificial intelligence.
Author
Key Advisors
Originally Published
July 08, 2025
Research program
AI Economic Policy Fellowship: Spring 2025
Abstract
This paper proposes sovereign wealth funds (SWFs) as a foundational policy instrument for states to navigate the economic and strategic challenges posed by transformative artificial intelligence (TAI). I explore the rationale for public investment in AI-related assets, drawing on analogies from resource governance and historical state-led development. I evaluate how different types of SWFs—strategic, stabilization, and savings funds—could be adapted to invest in compute infrastructure, upstream supply chains, and equity in frontier AI labs. My analysis identifies key pathways through which public ownership can shape TAI development, mitigate economic volatility, and distribute AIderived wealth more equitably. I assess the design tradeoffs of ownership models, from minority stakes to majority control, and highlight political and institutional constraints on implementation. By comparing SWFs’ potential to that of laissez-faire or purely regulatory approaches, I argue that SWFs offer a flexible, scalable tool for democratic states to retain strategic influence over TAI while preserving fiscal legitimacy. I conclude by outlining scenarios under which SWFs may succeed or fail, providing a structured framework for policymakers to evaluate this institutional lever in managing the AI transition.
Executive Summary
In this paper, I propose that sovereign wealth funds (SWFs) should be considered a foundational governance tool for managing strategic and economic challenges posed by transformative artificial intelligence (TAI). SWFs—state-owned investment vehicles—can be adapted to serve three critical roles in a TAI context:
1
Lead: Supporting national leadership in frontier AI development through strategic investment in upstream infrastructure and key AI firms.
2
Own: Acquiring equity in frontier AI companies to secure public influence over value capture.
3
Share: Distributing the economic returns from TAI in ways that preserve public legitimacy, stabilize the economy, and prevent elite entrenchment.
I evaluate SWFs as a flexible institutional lever, especially for democratic states seeking durable and non-coercive ways to shape AI trajectories.
This summary outlines the key conclusions of my paper. For the reasoning and analysis behind each, see the corresponding sections.
Why SWFs for AI Governance?
AI is increasingly viewed as a likely driver of future economic and geopolitical power. To ensure its development aligns with democratic interests, I propose that:
SWFs can secure long-horizon public influence over upstream AI infrastructure (e.g. chips, energy) and downstream AI companies.
Without public ownership, benefits from TAI are likely to concentrate among a small set of unaccountable actors.
Public ownership via SWFs is more resilient than regulation alone, which may face delay, rollback, or jurisdictional limits.
Proposed Uses
Enabling U.S. AI Leadership
I propose that a U.S. strategic fund could:
Invest in domestic semiconductor manufacturing, especially where the U.S. currently lacks cutting-edge capacity.
Co-invest in allied semiconductor ecosystems to increase redundancy and resilience in upstream chip inputs.
Support energy infrastructure buildout to meet the immense power demands of AI training.
Fund security-critical infrastructure (e.g., secure-by-design hardware) before commercial incentives emerge.
Acquire early minority stakes in frontier AI companies to gain access, governance rights, and strategic veto power without full control.
Ownership Pathways
I evaluate a spectrum of state ownership models:
1
No stake: maximizes private control; risks misalignment, elite capture, or regulatory failure.
2
Minority stake: allows targeted influence, often via side letters, board rights, or golden shares.
3
Majority stake/nationalization: allows full control but poses high political and financial barriers.
I conclude that minority equity positions are the most feasible and strategically attractive lever for influence.
Public Management of AI-Derived Wealth
I propose using savings funds to distribute universal basic dividends, providing citizens a share of AI-derived wealth.
Stabilization funds can smooth tax revenues and labor shocks caused by rapid automation by supporting displaced workers with income supplements and retraining, deploying automatic stabilizers, and maintaining core public services amid economic volatility
Design Principles
SWFs could be explicitly mandated to serve public benefit, with transparent objectives.
Governance risks (e.g., elite capture, opacity, political misuse) could be addressed via legal structures, reporting requirements, and independent oversight.
The erosion of public trust can be mitigated by:
Avoiding full nationalization unless necessary.
Embedding social objectives contractually or in the fund’s mandate.
Limitations and Risks
I identify six key limitations and risks that could undermine the effectiveness of SWFs for TAI:
Institutional fragility: Rapid AI progress may overwhelm or erode state capacity for funds to act as intended.
Speculative uncertainty: If TAI fails to materialize, fund performance may underdeliver, though strategic investments (e.g. energy) may still yield benefits.
Arms race escalation: State investment in AI could accelerate geopolitical rivalry and reduce cooperation incentives.
Governance risks: SWFs may be captured by elite interests or used for political patronage, undermining public benefit
Investment constraints: States may face fiscal, legal, or political barriers to investing in frontier AI companies, limiting their ability to secure public ownership.
Global exclusion: States unable to invest in TAI may be excluded from its benefits, reinforcing global inequalities.
Conclusion
If TAI becomes a dominant source of economic value in the near future, then ownership and governance structures will shape not only who captures that value, but also how it impacts economic stability, inequality, and political authority. SWFs offer a uniquely adaptive tool for helping states lead AI development, exert meaningful public ownership, and share the benefits of TAI more equitably.
Get research updates from Convergence
Leave us your contact info and we’ll share our latest research, partnerships, and projects as they're released.
You may opt out at any time. View our Privacy Policy.