The Federal Reserve is shifting from a "more is better" approach to financial supervision toward targeted, risk-based oversight that prioritizes material threats to stability. This modernization strategy, outlined by Vice Chair for Supervision Bowman at a recent Bank Policy Institute conference, reflects a broader rethinking of how regulators should govern complex, rapidly evolving technologies. The principles guiding this financial reform may offer a roadmap for AI governance as governments worldwide grapple with controlling AI development and deployment.

What Went Wrong With Traditional Financial Oversight?

The Federal Reserve's push for modernization stems directly from lessons learned during the Silicon Valley Bank collapse. Supervisors had dozens of outstanding findings against the bank, yet failed to identify and act on the material risks that ultimately led to its failure. The key insight: more regulatory findings do not equal better supervision. This realization has prompted the Fed to rethink how it allocates supervisory resources and focuses attention on what actually threatens institutional safety and soundness.

The Fed's modernization effort rests on four core principles that regulators worldwide are now adopting:

• Material Risk Focus: Supervisors must concentrate resources on risks that can impair an institution's viability and broader financial stability, rather than pursuing every possible compliance issue.
• Tailored Regulation: Rules should match each institution's actual risk profile; a community bank offering traditional lending should not face the same standards as a multi-trillion-dollar global institution with complex trading operations.
• Transparency and Accountability: Regulatory expectations should be publicly disclosed so firms understand what supervisors expect and how supervision is conducted, enabling better risk management on both sides.
• Forward-Looking Innovation: Frameworks must support responsible innovation while addressing emerging risks, using principles rather than overly prescriptive rules that cannot adapt to new circumstances.

How Could These Principles Apply to AI Governance?

The Fed's framework mirrors challenges now facing AI regulators globally. Just as financial supervisors struggled to prioritize among dozens of compliance matters, AI governance bodies are wrestling with how to focus limited resources on the most consequential risks. The principle of tailored regulation is particularly relevant: a small startup building a specialized language model for healthcare faces fundamentally different risks than a major technology company deploying large-scale AI systems across critical infrastructure. One-size-fits-all AI regulation could create unnecessary burden without improving safety outcomes.

The transparency principle also resonates with emerging AI governance debates. Regulators and companies alike are recognizing that AI governance expectations should be clear and publicly understood, not imposed as surprise audits or enforcement actions. This aligns with growing calls for explainability in AI systems and clearer regulatory standards for AI deployment in sensitive sectors.

What Is Sovereign AI, and Why Does It Matter for Global Governance?

While the Fed focuses on modernizing financial regulation, governments worldwide are simultaneously grappling with a different AI governance challenge: who controls the systems that power critical decisions. The concept of "sovereign AI" has emerged as a central concern in international technology policy. Sovereign AI refers to a country or organization's ability to develop, deploy, and govern AI systems using infrastructure, data, and models that remain subject to its own rules and priorities.

Sovereign AI extends beyond data sovereignty, which focuses on where information is stored and which legal frameworks govern its use. A government could store citizen data domestically while relying entirely on a foreign proprietary AI model to analyze documents, assess risks, or support administrative decisions. In such a case, the data remains under national jurisdiction, but the analytical process depends on external systems. This distinction between controlling data and controlling intelligence has become crucial as AI systems make increasingly consequential decisions in healthcare, defense, public administration, and critical infrastructure.

The growing interest in sovereign AI is closely connected to geopolitical competition and concentrated technological capabilities. AI development depends on semiconductors, cloud infrastructure, computing capacity, specialized talent, and advanced models. Many of these capabilities are concentrated among a relatively small number of countries and companies. Export controls on advanced chips and restrictions on access to certain technologies have prompted governments to consider whether reliance on external providers could create vulnerabilities during geopolitical conflicts.

How Are Different Regions Approaching Sovereign AI?

Countries and regions are pursuing sovereign AI for different reasons, reflecting their distinct strategic priorities and technological capabilities:

• China's Approach: Emphasizes technological self-reliance and control across the entire AI system, including infrastructure, models, and industrial capacity, reflecting broader national strategies aimed at reducing dependence on external technologies deemed strategically important.
• European Union Strategy: Combines regulation, industrial policy, and infrastructure investment through programs like InvestAI, which allocates substantial funding toward AI "gigafactories" and large data centers with significant computing capacity, while remaining connected to global innovation networks.
• Gulf Region Focus: Links sovereign AI with economic diversification and development of new technology sectors, investing in domestic data centers and Arabic-language AI models to gain control over critical digital infrastructure while adapting capabilities to regional languages and markets.

These approaches demonstrate that sovereign AI does not represent a single model. For some countries, the priority is security and resilience. For others, it is regulatory control, cultural representation, or economic development. However, all face a common challenge: balancing greater autonomy with continued participation in global technology networks. Open source models, international research, and cross-border expertise remain important drivers of AI progress, even as countries pursue greater control over critical capabilities.

What Does This Mean for Organizations Operating Across Borders?

The rise of sovereign AI creates new challenges for multinational organizations. As AI becomes embedded in business operations, companies may need to adapt not only their compliance strategies but also their technical architectures. A single AI workflow may require different implementations across markets depending on data governance requirements, infrastructure availability, and regulatory expectations. Organizations may need to manage multiple AI environments rather than rely on a single universal approach.

This fragmentation mirrors the regulatory tailoring principle the Fed is now emphasizing in financial supervision. Just as banks must adapt their operations to different jurisdictional requirements, technology companies are discovering that a one-size-fits-all AI governance approach is neither feasible nor appropriate. The intersection of financial regulation modernization and sovereign AI governance suggests a broader shift toward risk-based, context-specific oversight frameworks that acknowledge the complexity and diversity of how technologies are actually deployed and governed.