RegTech Acceleration: FINMA Leads Smart Surveillance for Crypto Markets
Marlene Amstad, Chair of the Swiss Financial Market Supervisory Authority (FINMA) and also Chair of the International Organization of Securities Commissions (IOSCO), has revealed that regulators are accelerating the adoption of artificial intelligence and regulatory technology (RegTech) tools, specifically targeting the crypto market. FINMA, as Switzerland's core financial regulator, has embedded AI into daily supervision processes to enhance the efficiency and accuracy of identifying risks associated with crypto assets.
Amstad highlighted that the high volatility, cross-chain operational complexity, and potential for contagion risk among institutions make traditional supervision methods insufficient. In response, FINMA has developed a real-time crypto monitoring dashboard that combines quarterly reported crypto asset holdings from institutions with daily market price data, generating a continuous risk view.
Real-Time Crypto Monitoring Dashboard: Pinpointing Concentration and Operational Risks
FINMA's real-time dashboard moves beyond static reports by analyzing two core risk dimensions through live data streams: first, excessive concentration of a single institution's crypto exposure—where one entity holds an overly large proportion of a particular crypto asset, exposing it to severe price shock risks; second, operational risks arising from tokens being concentrated on a single blockchain, where network congestion, attacks, or governance failures can simultaneously affect all related assets.
The dashboard automatically correlates the crypto asset quantities reported by institutions (including Bitcoin, Ethereum, and various tokens) with real-time market prices, generating concentration indicators. When an institution's single-asset exposure exceeds preset thresholds, or when token holdings on one chain become abnormally concentrated, the system triggers alerts for regulatory staff to investigate. This data-driven proactive approach significantly shortens the lag between risk accumulation and detection.
Generative AI Tool: Dual-Verification Mechanism Boosts Inspection Efficiency
Beyond the dashboard, FINMA is developing a generative AI tool specifically for document review prior to on-site inspections. The tool scans regulatory filings submitted by institutions, automatically flagging anomalies such as suspicious transaction patterns, unusual position change explanations, or contradictions in compliance reports. Flagged items are then passed to a second, independent AI model for verification, forming a dual-verification process. Only findings that pass both AI rounds are escalated to human regulators for final review.
Amstad explained that this design reduces human resource consumption in routine document screening while leveraging the error-tolerance of AI to minimize false positives. The two AI tools act in a 'red-team, blue-team' fashion, the first for broad screening and the second for validation, ensuring reliability. The project is still under development but has already demonstrated potential to greatly expand regulatory coverage and speed.
FINMA's AI-driven approach sets a benchmark for global crypto regulation. As generative AI matures, more regulators are likely to adopt similar dual-verification models, embedding AI across the entire chain—from data collection and risk identification to on-site inspections—pushing crypto supervision from experience-based to data-intelligence-driven.